Copyrig ht © 2006

By the American Association of Perroleum Cc~>logists (AAPG) and SEPM (Society fo r Sedimentary Geology)
All Rights Reserved

!SllN: 978-0-89181-704-8
0-89181-704-2

AAPC a nd SEPM grant permission for a single pho tocopy of an item fro m this publicatio n fo r personal use .
Authorization for add itio nal copies o f items from this publication for personal or internal use is g ran ted
by AAJ>G and SEPM provided that the base fee of $3.50 per copy and $.50 per page is paid d irectly to the
Co pyrig ht Cleara nce Center, 222 Ros<>wood Drive, Danvers, Massachllsetts 01923 (phone: 978/750-8400).
Fees are subject to change. Any form of electron.ic or d igital sca n.niJ>g or ot her d igital transformatio n of
portions ot' t·his publication i.nto computer-readable M>d/or t·ra nsm ittable form for persoMI or corporate use
re<.1 uires specia l permission f rom, and is subject to fee charges by, the 1\APG ru\d SEPM.

AAJ'>G Editor: Ernest A. Mancini SEJ>M Special P ublicatio n Ed itors:

Geoscience Director: )ames B. Blankenship Laura Crossey and Donald Me~'leill
SEPM Publications Coordinator: Robert Clarke

Tf\is publicat io t\ is available fro m:

The AAPG Bookstore SEPM

P.O. Box 9·79 6128 E.1St 381h St«->et, Suite 308
T,~so. OK U.S.A. 74101-0979 Tulsa, OK U.S.A. 74135-5814
Phone: l -918-584-2555 Phone: l-&J0-865-9765 (North America)
or '1-1:!00-364-AAPG (U.S.A. only) or ·t-9'1 8-610-3361
Fax: 1-918-260-2652 Fax: 1-918-62H685
o r 1-800-898-2274 (U.S.A. only) E ~ 1na iJ: eeUisW'sepm.org
E-mail: bookt>tore®aapg.org www.sepm.org
w-..\1"\v,aapg.org

Ceolosic.ll Society Publi•hiog 1-toU>e

Unit 7, Brnssmill Enterpl'ise Centre
BrassmiU Lane, Bath BA13JN
U.K.
Phone: +44-1225-445046
Fax: +44-'1 225-442836
E~roai J : $ale$@geolwc.org.uk
W\\11\'.geolsoc.org. tr.k

The American Associatio n of Petroleum Geologists (AAPG) does no t endorse or recommend products or
services that may be cited_, used, o r discussed in AAPG publications o r in presentations a t e vents associated
with AAJ'>C.
 Table of Con tents
Acknowledgments ······································ ~~-~~--..!<.

About the Edjtors vj

Introductj on

Giant Hydrocarbon Reservoirs of the World: From Rocks to

Reservoir Characterization and Modeling
P. M. Harris and L. [. Weber

Chapterl ................................. ··----------------------------L----

Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup, Precaspian
Basin, Kazakhstan: Depositional Evolution and Reservoir Development
[ . A . M. Kenter, P. M. Harris, [ . F. Collins, L. [. Webe1; G. Kuanyslwva, and D . [. Fischer

Chapter 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Facies and Reservoir-quality Variations in the Late Visean to Bashkirian Outer Platform,
Rim, and Flank of the Tengiz Buildup, Precaspian Basin, Kazakhstan
[.F. Collins,[. A.M. Kenter, P.M. Harris, G. Kuanysl1eva, D . J. Fischer, and K. L. Stef(en

Chapter 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Ghawar Arab-D Reservoir: Widespread Porosity In Shoaling-upward
Carbonate Cycles, Saudi Arabia
Robert P. Lindsay, Dave L. Cantrell, Geraint 111. Hugl1es, Thomas H. Keith,
Harry W. Mueller Ill, and S. Duffy f<usse/1

Chapter 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
High-resolution Sequence Stratigraphy and Reservoir Characterization of Upper Tha mama
(Lower Cretaceous) Reservoirs of a Giant Abu Dhabi Oil Field, Un ited Arab Emi rates
Christian[. Stmhmeuge1; L.jim Weber, Ahmed Ghani, Klwlil AI-Mehsin, Omar Al-jeelani,
Abdulla AI-Mansoori, l'aha AI-Dayyani, Lee Vaughan, Sameer A . Khan, and
john C. Mitchell

Chapter 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir,
Lower Cretaceous, Abu Dhabi (Un ited Arab Em irates)
Lyndon A . Yose, Amy S. Ru(, Christian f. Strohmenger, jim S. Schuelke, Andy Combos,
Ismail AI-Hosani, Slwmsa AI-Maskary, Gerald Bloch, Yousuf 111-Mehairi, and
Imelda G. jolmson
Ch apter 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Sequence Stratigraphy and Reservoir Architecture of the Burgan and Mauddud Formations
(lower Cretaceous). Kuwait
Christian f. Strolmrenger; Pemrv E. Patterson, Glraida AI-Salrlan, John C. Mitchell,
Howard R. Feldman, Timothy M. Demko, E<obert W. Wellner, Patrick f. Lelrmann,
G. Glen McCrimmon, E<obett H( Broom/rail, and Neama ;1/.J\jmi

Ch apter 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
The Sequence Stratigraphy of the Maastrichtian (Upper Cretaceous) Reservoir at Wafra Field,
Partitioned Neutral Zone, Saudi Arabia and Kuwait: Key to Reservoir Modeling and Assessment
Dennis 111. Dull, /laynumd A. Garber, and 111. Scott Meddaugh

Ch apter 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Stratigraphic Organization and Predictability of Mixed Coarse- and Fine-grained Lithofacies
Successions in a Lower Miocene Deep-water Slope-channel System, Angola Block IS
M. L. Porter, II. II. G. Sprague, M. D . Sul/ivmr, D. C. Jennette, II. T. Beaubouet T. R. Gar{ield,
C. Rossen, D. K. Sicktt(oose, G. N./ensen, S. /. Friedmann, and D. C. Mohrig

Ch apter 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Multiscale Geologic an d Petrophysical Modeling of the Giant Hugoton Gas field (Permian),
Kansas and Oklahoma. U.S.A.
Martin K. Dubois, 111/ttn P. BJ•rrres. Geoffrey C. Bo/1/irrg, and john H. Dovet011

Ch apter 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355

Key Role of Outcrops and Core.~ In Carbonate Reservoir Characterization and Modeling, Lower
Permian Ful.lerton Field. Perm ian Basi.n. United States
Step/ten C. Ruppel cmd Rebecca H. Jones

Chapter 11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395

Cart>onate Sequence Stratigrap hy and Petroleum Geology of the j urassic Deep Panuke Field,
Offshore Nova Scotia. Canada
John A. W Weissettberger, Richard 11. Wierzbicki, and Nancy f . Harland

Chapter 12. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433

Sedi mento logy, Sequence Stratigraphy, and Reservoir Ao:chltecture o f the Eocene 1-Hrador
Formation, Cupiagua Field, Llanos Foothills, Colombia
Juan Carlos Ramon cmd Anclres Fajardo

tv
 Acknowledgments
AAPG and SEPM
are grateful and appreciative
for the support of the AAPG Founda t io n .

The follow ing are gratefully

acknowledged fo r their financial support:

Chevron
EJf{onMobil

Contributions are applied against the production

costs o f publicatio n, thus directly reducing the hoo k's
purchase price and ma king the volu me
available to a greater a udience.

v
 About the Editors
Paul M. (Mitch) _f:l.a rris is n. Carbonate Reservoir Consultan t with Chevron Energy
T('chnolog:y Compa ny in San Ramon, Califo rnia. He performs <·a rbona te research,
1echnical support projects, consulting, and train ing for the various operating u nils
of Chevron. His \\10tk during t he last~ years has centered o n facies-re lated, strati-
graphic. and diagenetic probl(;'ms th at p€'rtain lO carbom:ne reservoirs and expiO·
ra tion play,o; in most carbo nate basins ,vorldwide.
Mitch rt-ceived h is B.S. and M.S. degre,o.s from West Virginia University a nd Ph.D.
from the University o f l\ttiami, Florida. He has published nu merous pa pers, edited sev-
eral books, a nd is active In AAI'G a nd SEI'I\1. lie has been a Dlstlng11ished Lecturer and
Inte rnational Distinguished Lecturer fo r AAPG. was awarded 11onorary lvlembershlp
from SEPM, and has received the Wallace E. Pratt Memorial Award and Robert H. Dott
Sr. Memorial Award fro m AAPG. Mitch is also ad junct faculty at Rlce University, the
University o f Miami, and the University o f Southern California .

L. J. Qim) W eber works for Ex.• onMobil ExploratJon Company as a Catbonate

Stratigraphy Expert specializing in seque nce/seismic stratigraphy and reservoir chmac·
tedzalion. liis curfent wo rk assignment invoJves techn ica l leadership an d pro ject
toordination as a stratigraphic advisor. J im h<as been employed in th(' oil industry fo r
more than 17 years, working for both heritage Exxon and ~·fobil in production, d eveJ.
o prnent, exploration, and research o rganizations. Pas t· work assignmen ts include SE
Asia, North a nd South Am(' rica, th (' Middle E.:1st, a nd th~ North Caspian rl'gion.
Jim received a bachelo r 1s degree in geology fro m DePau w Un iversity, a master's
d egree from the New Mexico Institute or Mini ng a nd Technology, aud his Ph.D.
from th e University of Ten nessee. He is ac tive in various geological socie ties includ-
in g SEPM and AAPG. He has published vario us papers and abstrac ts ranging from
t he Cambro.Ordovicia n to th e Holocene.

vi
 _,. M., and 1.. J. Wehc:r, 2006, Giant hydrocarbon re..'icrvoirs of
Harris, C
the world: from rods to reservoir e.hara<.'1CrizatiOH and modeling,. iu
P. M. Harris and 1.. J, Weber. eds., Giant hydrocarbon reservoirs of

Introduction the world: From rocl-:s to reservoir <:haracterlz.atlon and mo<lellng;

AAPCi Memoir 88/ Sti1M Special PublicatiOn, p. 1- 6.

Giant Hydrocarbon Reservoirs of

the World: From Rocks to Reservoir
Characterization and Modeling
P.M. Harris
Chevron E11ergy Tectmology Company, Sm1 Ramon, California, U.S.A.

L. J. Weber
fuonMo/Jil /Jxploration Company, Houston, Texas, U.S.A.

INTRODUCTION It is our goal that the technical content o f the chapters

The SEPM/AAPG core workshop "Giant Hydroca r·
bon Reservo irs o f the World: From Rocks to Reservoir
Characterization and Modeling" and th is co mpan ion
publication are an attempt to assemble information
on giant (>500 MOEB recoverable reserves) reservoirs
that is of value to a wide audience. Various examples
and methods of reservoir characterization, develop·
ment, and modeling practices are documented in this
volume. Although far from exhaustive, this compila·
lion includes a wide range of reservoirs when exam·
ined from any perspective, i.e., location, geology, pro·
duction history, and characterization. Figure 1 shows
the geographic distribution of the reservoir examples
that are Included In this volume.
Agood understanding of geologic variability in time
and space is a prereq uisite for any successful reservoir
description. Geologic and engineering data obtlli ned
from core are fundamental buildi.ng blocks in reservoir
characterization. A com mo n goal is to describe the
reservoir in sufficient detail to identify remain ing hy·
drocarbons and then to produce these reserve_~ effi.
ciently. To this end, a focus in this volume centers on
aspects of geologic model ing as they relate to het·
erogeneity In facies, which typic.a lly controls vari·
ability In porosiry, permeability, and fluid saturations.
presented here result in discussion d irected toward
• fundamental concepts and methods of reservo ir
characterization, which include accurate represen·
tation of internal and external reservoir geology,
precise and quantitative description, and level of
detail that satisfies simulation capabilities
• an understanding of the scales of rock hetero·
geneity and its effect o n petrophysical and engi·
neering properties and their relationships to fluid
flow and hydrocarbon recovery
• improved methods o f reservoir description an.d de-
velopment through application of high-resolution
sequence and seismic stratigraphy and seismic vi-
sualization techniques
• identifying a lternative approache.~ to mo re e ffec-
tive reservoi r management practices
The reservoir examples described in this volume
(!) explore historical and alternative approaches to
reservoir description, characterization, and manage·
ment and (2) examine appropriate levels and timing
of data gathering, technology applications, evalua·
tion techniques_, risk assessment, and nt.a nagement
practices In various stages In the life of Individua l
development projects.

COp)'rigllt 0 2006 by 1'hc American Associat:iou of PNrolcum CicoJogim.

001:10. 1306/1 21S872M8824?

1
2 I liarris and Weber
The giant reservoirs of th is volume account for
approximately 0.5 trillion bbl recoverable hydrocar-
bons. Reservo ir examples a re both ca rbonate and si-
liciclastic, and collectively, they accoun t for a wide
range of variability in reservoir parameters (e.g., gross
rock volume, net-to-gross, porosity, and permeabi i·
ity). Most of the reservoirs in this volu me occur o ut·
side the United States and Canada, a nd until now,
core from many of these reservoirs have not been
widely observed .
Enhanced recovery of hydrocarbons requires a criti·
cal understanding of reservoir heterogeneity by both
geoscientists and engineers. Spatial heterogeneity that
affects fluid now occurs over a ll scales of investiga-
tion from the intr<!well, in terwell, reservoir, and ba-
sin scales of exam ination. We feel that the giant fields
discussed herein address issues important to reservo ir
description, characterizatlon, and manage ment from
both geologic and engineering perspeclives.
SUMMARY OF RESERVOIR EXAMPLES
Caspian Basin
Kenter, Harris, Collins, Weber, Kuanysheva, and
fischer discuss the t:entral pla tfo rm part oftheTengiz.
field in "L<!te Visean to Bashkirian Platform Cyclicity
in the Centra l Tengiz Buildup (Precaspian Basin):
Depositional Evolution and Reservoir Development."
The Tengiz buildup, an inten sely cored and studied
iso lated ca rbonate platform in the Precaspian Basin,
contains a succession of shallow-water deposits ra ng-
ing fro m Famenn ian to Bashkirian in age. The upper
Visean, Serpukhovian, a nd Bashkirian form the main
h ydrocarbon-bearing interval in the platform. Depo-
sitio nal q•cles (high-frequency sequences) in this in·
terval arc several to tens of meters thick fo r the Visean
and Serpukhovian, and decimeter to meter scale for
FIGURE 1 . Reservoirs included
in Giant Hydrocarbon Reser-
voirs of the World. Caspian
Basin: (1) Tengiz platform and
(2) Tenglz rim; Middle East:
(3) Ghawar Arab-D, (4) Abu
Dhabi Thamama, (5) Abu Dhabi
Shualba, (6) Kuwait Burgan -
Mauddud, and (7} Wafra
Maastrichtian; West Africa:
(8) Angola Block 15; North
America: ( 9) Hugoton,
(1 0) Fullerton, and (11) Deep
Panuke; South America:
(12) Cuplagua.
the B<lShkirian . Visean and Serp ukhovian cycles are
genera lly easy to correlate fro m well to we.ll over sev-
eral kilometers distance, whereas Bash kirian cycles
are incomplete and are more difficu lt to correlate.
The distribution of reservoir rock types in the central
platform is determined by buria l diagenetic modifl·
cation of an earlier reservoir system that includes
meteoric alteration and po rosity enhancement be-
low major sequence boundaries and reduced disso-
lu tion along higher order sequence boundaries asso-
ciated with the presence of volcanic ash. The lateral
continuity of tight layers at sequence boundaries prob-
ably grea tly affected later fluid flow as well as t he
ultimate distribution o f cementS, dissolution, and bi·
lumen in the central pl<ltfo rm reservoir. The buri<Jl
diagenetic overprint includes two majo r phases of
reservo ir modification: (1) a corrosion and cementa-
tion phase signillcantly enhanced existing matrix po-
rosity In the interior central platform while reduci ng
porosity in the exterior central and outer platform by
pore-filling equant calcite cement, and (2) bitumen
emplacement and associated corrosion.
Authors Collins, Kenter, Harris, Kuanysheva, Fischer,
and Steffen examine the outer platform, rim and fla nk
of the Tengi z field in " Facies and Reservoir Qual-
ity Vari<ltions in th e L;ote Visean to Bashkiri;m Out-
er Platform, Rim, <lnd Fl<mk of the Tengiz Buildup,
Precaspian Basin, Kazakhstan." Platform backstep-
ping fro m Tournaisian through late Visean resulted
in approximately 800 m (2624 ft) o f bathymetric
relief. This topography was enve loped by as much as
2 km (1.2 mi) of Serpukhovian progradation, which
formed a depositiona l wedge a round the o lder plat-
forms referred to as the Serpukhovian rim a nd flank.
Lower-slope facies include mudstone, volcanic ash,
and platform-derived skeletal packstone to grainstone
interbedded with boundstone breccia; m iddle-slope

facies are poorly bedded to massive boundstone brec-
cia with subtypes based on clast co mposition, size,
and packing; and upper-slope facies con sist of in-situ
microbial bo undstone. Facies of the outer platform
are shallow-platform skeletal, coated-grain, and ooid
packsto ne to grai n.stone. Periodic large-sca le failure
of the rim during both the Serpukhovian and Bash·
kiria n resulted in a high degree of lateral facies dis-
continuity. Solution-en larged fractures, large vugs,
and lost circlllation zones produced mainly during late
diagenesis form a high-permeability, well-connected
reservo ir in the rim and flank.
Middle East
Lindsay, Cantrell, Hughes, Keith, Mueller, and Rus-
sell describe Ghawar field, which is the world's largest
and most prolific o il field, in "Ghawar Arab-0 Res-
ervoir: Widespread Porosity in Shoaling-upwa rd Car-
bonate Cycles, Saudi Arabia." Production occurs fro m
the j urassic Arab-D carbonates. The upper half of the
reservoir is dominated by exceptiona lly high reser-
voir quality, and t he lower half conta ins interbeds
of less porosity. Th e reservoir is com posed o f two
composite sequences. The upper composite sequence
boundary is the top o f Arab-D carbonate, which is
locally characterized by colla pse breccia. The lower
composite sequence boundary between tbe Arab-D
and the underlying j ubaila is m arked by deeper water
cycles occu rring over gra in-dom inated cycles. Several
h igh-frequency sequences are each composed o f cy-
cle sets that each contai n approximately five individ-
ual carbonate cycles. These carbonates formed upon
a broad, arid storm -domina ted ram p with a variety
of rock types deposited . Diagenesis tha t is common
within the Arab-D reservoir includes severa l dissolu-
tion even ts, recrystallization, and physical compac-
tion. The resultant limesto ne porosity is a m ixture of
interparticle (dom inant), moldic (c.ornmon), in tra-
particle (common), and microporosity (common) pore
types. Less common dolostone porosity is a mixture o f
moldic (less common), intercrystalline (less common).
and intracrystalline (least common) pore types. The
vertical seal for the reservoi r is the overlying Arab C-D
an hydrite.
In "High -resolution Seq uence Stratigraph y an d
Reservoir Characterization o f Upper Thama ma (Low-
er C retaceous) Reservoirs o f a Giant Abu Dhabi Oil
Field, United Arab Emirates" authors Strohmenger,
Weber, Ghan i, AJ-Mehsln, Al-Jeelani, AJ-Mansoori,
Al-Dan•ani, Vaughan, Khan, and Mitchell describe
the Lower Cretaceous Kh araib (Barrem ian and early
Aptian) and Shuaiba (Aptian) formations (upperTha-
Introduct ion I 3
mama Group) o f Abu Dhabi in wb icb important hy-
droca.rbon accumu lations occur in p latform carbon-
ates. The Kharaib and Lower Shuaiba form ations
contain th ree reservoir units separated by low-porosity
and permeability dense zones. Core and well-Jog data
fro m a gia nt oil field in Abu Dhabi and outcrop data
fro m Wadi Rah abah in Ras Al-Khaimah were used to
establish a seq uence-s tratigraphic framework and a
litl1ofacies scheme. The Lower a nd Upper Kharaib
Reservoir Units, as well as the upper dense zone, are
part of a late transgressive sequence set of a scc.ond-
order supersequence, built by two third-ord er com-
posite sequences. The overlying Lower Shuaiba Res-
ervoir Unit belongs to the la te transgressive sequence
set and the early highstand sequence set o f this second-
order supersequence a nd comprises one third-order
composite sequence. The th ree third-o rder compos-
ite sequences a re composed o f fourth-ord er parase-
quence sets that show p redo minantly aggradational
and progradational stacking patterns, typical of green-
house cycle.~. Reservoir lithofacies range from lower-
ramp to shoal crest to near backshoal open-platform
deposits, whereas nonreservoir (dense) lithofacies rep-
resent an inner-ramp, restricted shallow lagoonal set-
ting. Integration of subsurface and o utcrop data leads
to more insightful and realistic geological models of
the subsurface stratigraphy, and the geological model
realizations based on core, outcrop, well-log, and seis-
mic data co nstrain now-simulation models.
Yose, Ruf, Strohmenger, Sch uelke, Gombos, AI-
Hosani, AI Maskary, Bloch, Al-Mehairi, a nd Johnson
integrate high-resolution three-dimensional (3-D) seis-
mic with geologic and production data to describe
the Lower Cretaceous (Aptian) reservoir in Abu Dhabi in
"Volume-based Characterization of a Heterogeneous
Carbonate Reservoir, Lower Cretaceous, Abu Dhabi
(United Arab Emirates)." The. reservoir is positioned
over a platform-to-basin transition and rec.ords a di-
verse range of depositional facies and stratal geom-
etries. A second-order sequence set is divided into
five depositiona l sequen ces. Sequences 1 and 2 are a
transgressional phase showing the initial formation
of buildup ma rgins and do minated by a lgaJ.prone
f•cies. The subsequent h i.ghstand phase of Sequence 3
is mainly aggradational and records the proliferation
of rudists across t he platfo rm !OJ>. A late highsta nd
phase of sequences 4 and 5 is progradational showing
th e progressive downstepping of the platform margin
onto a low-angle slope. Three-dimen sional seismic
data in the southern field area show a com plex mo-
saic of tidal channels, high-energy rudist shoals, and
intershoal ponds. The geometry and reservoir-quality
4 1 Harris and \Veber
variations of these geologic features have a strong
impact on reservo ir S\oveep and co nformance in the
platform interior. ln the northern field area, seismic
images of prograding slope clinoforms reveal system -
atic variations in architecture and reservoir qualitythat
reflect multiple scales of stratigraphic cyclicity. A pat·
tern gas llood has been Implemented in the clinoforms
to add pressure support and improve recovery. Busi·
ness applications of the reservoi r framework include
(1) 3-0 seismic visua lization as a tool for optimizing
well placement, identifying bypassed reservoirs and
evaluating reservoir connectivity; (2) integra tion of
quantitative, volume-based seismic informatio n into
reservoir models; (3) m<Jximjzing recovery through full
integration of a ll subsurface data; and (4) enh<~nced
communicatio n a mong geoscientists and e ngmeers,
leading to improved reservoir management practices.
In "Sequence Strati.graphy and Reservo ir Arch itec-
ture of the Burgan and Mauddud Formations (Lower
Cretaceous), Kuwait" authors Strohmenger, Patter·
son, AI-Sahlan, Mitchell, Feldman. Demko, Wellner,
Lehmann, McCrimmon, Broomhall, and AI·Ajml pro·
pose a new seq uence-stratlgraphic framework for the
Burgan and Mauddud formations (Albian) of Kuwait,
which resu lts in a predictable distribution of reset·
voir and seal fades. The Burgan and Mauddud for·
mations form two second-order composite sequences,
the o ldest of which constitutes the lowstaod, trans·
gressive, and h ighstand sequence sets of the Burgan
Formation. This composite sequence is subdivided into
high-frequen cy sequences that a re characterized by
tidal-inlluenced, marginal-marine deposits in norUl ·
east Kuwait grading into more fluvial-dominated, con·
tinental deposits to the southwest. The younger com·
posite sequen ce consists of the lowstand sequence set
of the uppermost Burgan formation and the trans·
gressive. and highstand sequence sets of the overly·
jng Maud dud l'ormation. This composite sequence is
sand and mud prone in southern and southwestern
Kuwait and is carbonate prone in northern and north·
eastern Kuwait. The lowstand se<Juence set of the Bur·
gan is subdivided into five high-frequency sequences.
and the Mauddud transgress ive and highstand se-
q ue nce sets a re subdivided into eight h igh-frequency
sequences. n 1e traditional lithostratigraphic Burgan -
Mauddud contact is time transgressive. The upper
Mauddud h ighstand sequence set is carbonate prone
and thins southward because of depositional thinning.
AuU10rs Dull, Garber, and Meddaugh describe the
Maastrichtian (Upper Cretaceous) reservoir in the giant
Wafra o il field in "The Sequence Stratigraphy of the
Maastrichtian Reservoir at Wafra field, Partitioned
Neutral Zone, Saudi Arabia and KuwaH: Key to Res·
ervoir Modeli ng and Assessme nt." Oi l production is
largely from subtidal dolomite formed on a very gently
dipping, shallow, arid, and restricted ramp setting
that transitioned between no rma l ~marine conditions
to restricted lagoonal environments. The key to mod·
ellng the reservoir was the construction of an appro·
priately d etailed sequence -stratigraphic fram ework
for use in building the geostatistical reservoir model.
Within the sequence-stratigraphic framework, 10 ltigh·
frequency sequences are correlated, a lbeit with some
dllficulty, across the entue field. Ultimately, diagen·
esis is a major factor in the distribution of po rosity
and permeability. Dolomitjzation is pervasive, but fa·
cies exerts some control on the distribution of pO·
rosity. Average porosity o f the reservoir interval is
15%, with values as much as 45%, a nd permeability
averages30 md wiUl core-plug measurements as much
as 1200 md. The geostatistical model of the Maas-
trichtian reservoir demonstrates the layered a nd com ·
partmentalized nature of the reservoir a nd clearly
shows that the location of the reservoir facies is con·
trolled by the o riginal depositional fabric and sub·
sequent dolomitization, both of which have been
influenced by the paleotopography. Th is study w<Js
undertaken to determine reservoir volumetrics, un·
derstand the distributio n of inte rvals likely to yield
h igher volumes of better quality oil, and provide a
reservo ir property model fo r use in fluid-flow simu-
lation. Such an understanding is critica l to efficiently
develop the 1.5 billion bbl oil Maastricthian resource
at Wafra field .
West Africa
In "Stratigraphic Organization and Predictability
of Mixed Coarse· and fine-grained Lithofacies Succes·
sions io a Lower Miocene Deep-water Slope-channel
System, Angola Block 15," Porter, Sprague, Sullivan,
Jennette, Beaubouef. Garfield, Rossen, Sickafoose, )en·
sen, Friedmann, and Mohrig describe the lower lvlio·
cene slo pe-channel systems from Angola Block 15,
which regional seism ic mapping, exploration drilling,
and appraisal drilling have established as a world-
class development opportunity. One o f the major de·
velopment targets in Block 15 is Burdigalian-aged
slo pe-channel reservoirs, which are part of a system
that traverses across the block in a n east - west direc.
tion and can be continuously mapped on adjacent
seism ic datasets for more than 30- 40 km (18- 24 ml).
The channel system was a sediment fairway for the
delivery of coarse-grained turdidites and mixed mud·
dy and sandy debrites in to the Lower Congo basin.

Map patterns show distinctive changes in sinuosity,
channel confi nement, and degree of amalgamation
bro<tdly rel<tted to concurrent growth of salt-related
structures. The episodic fill of the s lope-channel sys-
tem can be better understood by a hierarchical ar·
rangement of unco nformity-bounded stratal units.
Nested channels form composite channel complexes
that show distinctive trends in lithofacies type and
vertical facies succession . Conventional cores calibrat-
ed to well logs and high-resolution seismic data show
that the lower parts of the channel complexes are
dominated by sandy-muddy debrites, slumps, and in-
jected sa ndsto nes. These facies are typically overlain
by coarse-grained, gravelly, and well-amalgamated
sandy turbidites. The ove rlying facies succession is
more variable, but commonly consists o f interbedded
sandy and muddy turbidites, injected sandstones,
and a range of both muddy and sandy debrites.
North America
Dubois, Byrnes. Bohllng, a nd Doveton docume nt
the reservoir characterization and modeling from
pore to field scale of the giant Hugoton field in "Mul·
tlscale Geologic and Petrophysical Modeling of the
Giant Hugoton Gas field (Permia n), Kansas and Ok-
lahoma." Their work on this mature Permian gas sys-
tem is aiding in defining original gas in place, the
nature and distribution of gas saturation, and reser-
voir properties. The Kansas- Oklahoma part of the field
has yielded 963 billion m 3 (34 teO gas throughout a
70-year period from more than 12,000 wells. Most
remaini ng gas is in lower permeabiUty pay zones o f the
170-m (557-ft)·thick, differentially depleted, layered
reservoir system. The main pay zones have remarkable
lateral continuity. They represent 13 shoaling-upward,
fourth-order marine-continental cycles, comprising
thin-bed ded (2-10-m; 6.6-33-ft), marine cMbonatc
mudstone to grainstone and siltstones to very fine sand-
stones. The pay zones are separated by low-reservoir-
quality eolian and sabkha redbeds. Petrophysical prop-
erties vaty among major lithofacies classes. Neu ral
n etwork p rocedures, stochastic modeling, and auto-
mation fadlitated building a detailed full-field 3-D cel-
lular reservo ir model using a fou r-step workflow:
(1) define lithofacies in co re and correlate to electric
log cu rves (training set); (2) train a neu ral network
and predict li thofacies at noncored wells; (3) po p-
u late a 3-D cellu lar mode l wi th lithofacies using
stochastic methods; and (4) populate m odel with
lithofacies-specific petrophysical properties and fluid
saturations. Both the knowledge gained and the tech-
niques and workflow employed have implications
Introduction I 5
for understanding and modeling reservoir systems
worldwide that have similar geologic age and reser-
voir architectu re.
In "Key Role o f Outcrops and Cores in Carbonate
Reservoir Characterization and Modeling, Lower Perm -
ian Fullerton field, Permian Basin, United States" au-
thors Ruppel and jones discuss the rock-based model
con.struction for fulle rton Clear fork field, which is
a shallow-water platform carbonate reservoir of mid-
dle Permian age in the Permian Basin of west Texas.
Fundamenta l steps in their study included (1) creat-
ing and applying an analogous outcrop depositional
model; (2) describing and interpreting subsurface core
and log data in terms of this initial model; (3) defin-
ing the sequence-stratigraphic arch ite~ture of the res-
ervoir section; (4) developing a ~)'de-based reservoir
framework; and (5) defin ing controls, interrelation-
ships, and distribution o f porosity and permeability.
Data used in th is analysis included co res. th in sec-
tions, 3-D and two-dimensional (2-D) seism ic data,
borehole image Jogs, and outcrop models. Stratal ar-
chitecture, differential dolomitization, karst fill, min-
eralogical variations, and rock-fabric distribution were
incorporated into the model. 'lh ese components were
used to con strain interpretation and d efinition of
flow units, permeability distribution, and saturation.
The rock-based methods demonstrated in this study
provide key insights with broader application into
the formatio n, characterization, and interpre tation
of cMbonate platform reservoirs.
Authors Weissenlle rger, Wierzbicki, and Harland
describe the dee p Pan uke field in "Carbonate Se-
quence Stratigraphy and Pe troleum Geology of the
jurassic Deep Pa nuke f ield, Off~hore Nova Scotia,
Canada ." Deep Panuke, which is located 250 km
(!55 mi) offshore of Halifax, Nova Scotia, Canada,
contains gas-fiUed cavernous porosiry in jurassic car-
bonates of the Abenaki !'ormation. The Abena ki car-
bonates range from Bathonian to Neocomian in age
and were deposited on broad carbonate platform at-
tached to a siliciclastic hinterland. Seven third-order
depositio nal sequences recogn ized in the Abenaki
are correlated with geology and a 2-D seism iC grid;
3-D seismic data are used (Ot deli neation drilling and
reservoir characterization. Lithologies range fro m fo re-
slope to reefal and shoal deposits near the platform
margin. Open and restricted lagoon or tidal-flat de-
posits occur in th e platform interior. Siliciclastics
are concentrated near sequence boundaries or are
distributed along stri ke, close to point sources such
as rive rs cutting through the platform . Th e reser-
voir occurs n car the platform margin in coral and
6 1 Harris and \Veber
stromatoporoid reef and associated skeleta l-peloidal
and occasionally oolitic shoal deposits. It comprises
a range of porosity types, from vuggy limestone and
dolo mite to microporosity in limestone. Geochemi-
cal, isotOJ>e, and petrographic data suggest that the
dolomitization and dissolution can be attributed to
deep burial and hydrothermal fluids.
Sout h America
In "Sedimentology, Sequence Stratigraphy, and
Reservoir Architecture of the Eocene Mirador !'or-
mation, Cupiagua Field, Llanos Footh.ills, Colombia,"
Ramon and Fajardo document the stratigraph.ic archi-
tecture and facies distribution in a h igh-resolution
ti me-space framework to define the 3-D reservoir
zonati on of the Mirador Formation (Eocene) in the
Cupiagua field. The Cupiagua structure is a large, east-
verging. asymmetric anticlinal fold that trends no rth-
northeast in the hanging wall o f the frontal fault. The
Mirador Formation accounts for approximately 55%
of the recoverable oil in Ute field. Three scale.~ of strati·
graphic cycles are recognized based on stacking pat·
tern and general lrend of facies successions: short-
term cycles or progradational and aggradational units
stack systematically into intermediate-term cycles,
which, in turn, are grouped into long-term cycles.
The lower half of the Mirador consists of flood-plain
facies with channel, crevasse splay/ and swamp and
flood-plain facies successions. Bay-head delta and bay-
fill facies occur in the upper half of the Mirador Forma-
tion. The Lower 1\·firador consists or two intermediate-
sca le cycles s howing a seaward-stepping stacking
pattern overla in by a third cycle with a landward-
stepping pattern, and the upper Mirador conti nues
the landward-stepping pattern . This upper unit con-
sists of tluee onlapping cycles composed of a suc-
cession of aggradatio nal channel deposits, p rograda-
tional bay-head delta and bay-fill deposits with a
landward -stepping stacking pattern. The Mirador is
capped by restricted marine shales of the Carbonera
Formation.
ACKNOWLEDGMENTS
As organizers of the core workshop and editors of
the companion voltune, we acknowledge several peo-
ple and our respec.tive companies for their assistance,
without which the workshop and volume would not
have happened. Chevron and ExxonMobil provided
generous financial assistance that enabled numerous
students to attend the core workshop and subsidized
printing costs o f the worksl•op volume. SEPM staff
helped us o ve rco me n umerous o rganizational and
logistica l issues that confronted us as we organized
the workshop. AAPG staff and, in particular, Beverly
Molyneux provided us invaluable help in the editing
and publication of the workshop volume. We thank
all of the authors wh o worked bard to prepare poster
and core displays, oral presentations, and the de-
tailed manuscripts t11at describe their respective fields.
Finally, we thank the many who helped us in the
technical review of the manuscripts for the workshop
publication: Bob Alway, Steve Bachtel, Sherry Becker,
Kelley Bergman, j ohn Bova,joel Colli ns. Bob Dalrym-
ple, Laurence Droz. Ch.ip Feazel, Sean Guidry, Jurgen
Grotseh, jean Hsieh, jon Kaufman, Mike Kozar, Dale
leckie, j ose Matos, jim 1\kGovney, Gary Parker, Carlos
Plrmez, George Pemberton, Linda l>rJce, Gene Rankey,
Rick Sarg, Toni Simo, Krishnan Srinivasan, and Niall
Toomey.
 1
Kenter, J. A. M., P. M. Harris, J. F. Collins, L. J. Weber, G. Kuanysheva, and
D. J. Fischer, 2006, Late Visean to Bashkirian platform cyclicity in the cen-
tral Tengiz buildup, Precaspian Basin, Kazakhstan: Depositional evolu-
tion and reservoir development, in P. M. Harris and L. J. Weber, eds., Giant
hydrocarbon reservoirs of the world: From rocks to reservoir characteriza-
tion and modeling: AAPG Memoir 88/SEPM Special Publication, p. 7 – 54.

Late Visean to Bashkirian Platform

Cyclicity in the Central Tengiz
Buildup, Precaspian Basin,
Kazakhstan: Depositional Evolution
and Reservoir Development
J. A. M. Kenter1 L. J. Weber
Vrije Universiteit, Amsterdam, Netherlands ExxonMobil Development Company,
Houston, Texas, U.S.A.
P. M. Harris
Chevron Energy Technology Company, G. Kuanysheva
San Ramon, California, U.S.A. TengizChevroil, Atyrau, Kazakhstan

J. F. Collins D. J. Fischer
ExxonMobil Development Company, TengizChevroil, Atyrau, Kazakhstan
Houston, Texas, U.S.A.

ABSTRACT

T
he Tengiz buildup, an intensely cored and studied isolated carbonate
platform in the Precaspian Basin, contains a succession of shallow-water
deposits ranging from Famennian to Bashkirian in age. From a reservoir
perspective, Tengiz can be subdivided into platform (central and outer) and rim-
slope (flank) regions. The upper Visean, Serpukhovian, and Bashkirian form the
main hydrocarbon-bearing interval in the platform. Depositional cycles (high-
frequency sequences) in this interval are several to tens of meters thick for the
Visean and Serpukhovian, and decimeter to meter scale for the Bashkirian.
Cycles are made up of a succession of lithofacies overlying a sharp base that
locally shows erosion, calcretes, meteoric diagenesis, and other evidence for
subaerial exposure. At the base of the succession, tight peloidal mudstone and
ash beds are associated with sequence boundaries and are thought to reflect low-
energy conditions developed in deeper platform areas at lowstand and during

1
Present address: Chevron Energy Technology Company, San Ramon, California, U.S.A.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215873M88374

7
8 / Kenter et al.

initial flooding. Above this, beds with in-situ articulated brachiopods signal ini-
tial open-marine but still low-energy conditions. Succeeding crinoid-dominated
intervals represent maximum marine flooding and overlying skeletal-peloidal
grainstones highstand shoaling phases.
Visean and Serpukhovian cycles are generally easy to correlate from well to
well over several kilometers distance. Volcanic ash beds are identified by gamma-
ray spikes, and flooding intervals show as low-porosity zones. In contrast, Bash-
kirian cycles are thinner and incomplete, dominated by thin, peloidal mudstone
intervals alternating with high-energy coated-grain and ooid grainstone, and are
more difficult to correlate. High-frequency icehouse sea level fluctuations ex-
posed the platform during each fall of sea level, and rapid flooding resulted in
incomplete cycles and complex lateral facies changes that may explain relatively
poor lateral continuity of log character.
The distribution of reservoir rock types in the central platform is deter-
mined by burial diagenetic modification of an earlier reservoir system that includes
meteoric alteration and porosity enhancement below major sequence bound-
aries and reduced dissolution along higher order sequence boundaries associated
with the presence of volcanic ash. The lateral continuity of tight layers at se-
quence boundaries probably greatly affected later fluid flow as well as the ulti-
mate distribution of cements, dissolution, and bitumen in the central platform
reservoir.
The burial diagenetic overprint included two major phases of reservoir modi-
fication. First, a corrosion and cementation phase significantly enhanced existing
matrix porosity in the interior central platform while reducing porosity in the
exterior central and outer platform by pore-filling equant calcite cement. This was
followed by bitumen emplacement and associated corrosion. These processes not
only exerted an overall porosity-reducing effect prior to and associated with
bitumen invasion toward the exterior central platform, but also dampened or
flattened the initial cyclic porosity variations and obscured relationships between
pore types and permeability. The bitumen overprint is nearly absent in the
innermost platform wells; bitumen concentrations are highest near the bases of
the cycles, which may imply that the first fill of hydrocarbons migrated through
the flanks laterally into the platform cycles.

INTRODUCTION Tengiz field produces a light, intermediate-sulfur,

Tengiz Field History
The Tengiz field, located in western Kazakhstan, near
the northeastern shore of the Caspian Sea (Figure 1),
produces oil from an isolated carbonate platform
(aerial extent of >110 km2 [42 mi2]) of Devonian and
Carboniferous age. Tengiz was discovered in 1979 by
the Ministry of Oil Industry of the Soviet Union. The
discovery well Tengiz 1 (T-1) reached a total depth
of 4095 m (13,435 ft). Development drilling of the
Tengiz field commenced in 1983, and onsite con-
struction of plant-processing facilities began in 1987.
Field production officially began in April 1991. Since
1993, TengizChevroil, an in-country joint venture
company run by Chevron, has operated Tengiz and
the adjacent Korolev field.
stabilized tank oil of approximately 478 API. As of mid-
year 2005, more than 115 wells have been drilled on
Tengiz. The highest rate wells are located in the plat-
form margin and slope in fractured carbonates with
low (<6%) matrix porosity. Platform wells exhibit
higher porosity (as much as 18%), but matrix perme-
ability is typically low (<10 md).
Regional Paleogeography
The Precaspian Basin is a large Paleozoic basin that
occupies much of the area known today as the north-
ern part of the Caspian Sea and adjacent landmass.
During the Early Carboniferous, the Precaspian Basin
was a major equatorial sag basin in the western part of
Kazakhstan (Ross and Ross, 1985). Carbonate shelves
formed along the basin margins, and isolated, broad
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 9

FIGURE 1. Index map locating North Caspian region and outlining area of the Tengiz field in blue, which is shown in
more detail in Figure 2.

carbonate platforms grew on preexisting basement
highs (Cook et al., 1997). Together with other nearby
fields, Tengiz forms an archipelago of isolated car-
bonate platforms that grew on a regional pre-middle
Devonian structural high near the southeastern mar-
gin of the basin (Figure 2). The initial carbonate plat-
forms in the archipelago formed sometime after the
Middle Devonian and continued more or less unin-
terrupted until the early Bashkirian. The total thick-
ness of carbonate that accumulated at Tengiz during
that time has not been drilled but is represented by
approximately 1.2 s two-way traveltime on the most
recent three-dimensional seismic survey (Figure 3).
Following platform termination in the early Bashkir-
ian, the platforms were encased in a series of lithologies
associated with the closure of the Precaspian Basin.
Initially, a veneer of Moscovian through Artinskian
deep-water carbonates and volcaniclastics accumulat-
ed around and on top of the buildups. These deposits
were followed by thick Kungurian evaporites, includ-
ing halite that completely encased the platforms.
SUPERSEQUENCE FRAMEWORK
In 2001, a joint study by ExxonMobil, Chevron,
and TengizChevroil established a stratigraphic frame-
work for the Tengiz platform based on seismic, bio-
stratigraphic, core, and well-log data. This study was
published in late 2003 (Weber et al., 2003) and stands
10 / Kenter et al.
FIGURE 2. Paleogeographic map showing the archipelago of isolated carbonate platforms, including Tengiz, that grew
on a regional premiddle Devonian structural high near the southeastern margin of the Precaspian Basin.
as the definitive stratigraphic reference for Tengiz.
The framework consists of a hierarchy of sequences
that have continued to be maintained as Tengiz dril-
ling proceeds.
Through the main hydrocarbon-bearing interval
at Tengiz (Famennian – Bashkirian), seven bounding
discontinuity surfaces (sequence boundaries and max-
imum flooding surfaces [MFSs]) are recognized on
seismic data (Weber et al., 2003) (Figure 4; Table 1).
Two additional MFSs, Lvis1_MFS and Lvis2_MFS, are
identified primarily from well control (cores and
well logs). Four supersequences (Tournaisian – lower
Visean, lower Visean – upper Visean, upper Visean –
Serpukhovian, and Bashkirian) extend from the Fa-
mennian supersequence boundary (Fame_SSB) to the
Bash_SSB (Weber et al., 2003). Each supersequence
contains a transgressive sequence set (TSS) and a high-
stand sequence set (HSS) separated by an MFS. The
upper two supersequences are the focus of this chapter.
Three reservoir zones (Visean A, Serpukhovian, and
Bashkirian) are contained in these supersequences and
constitute the platform reservoir part of the Tengiz
buildup.
Thin subaerial exposure caps on subtidal lithofa-
cies form the most common type of cycle top in the
Lvis_SSB to Bash_SSB succession on the Tengiz plat-
form. The implication of this type of cycle boundary
is that rapid base-level fall exposed subtidal litho-
facies without development of an intervening tidal
flat or peritidal succession. Alternatively, anchor points
for tidal flats may not have been available because the
depositional profile near the margin gradually deep-
ened toward the platform break without the pres-
ence of a shallow barrier. In other words, the Tengiz
platform was not conducive to producing extensive
tidal-flat facies. Each cycle is more appropriately
termed a ‘‘high-frequency sequence,’’ capped by a high-
frequency sequence boundary. Cycle boundaries also
occur where subaerial exposure is not evident from
core and thin-section description. Without evidence of
subaerial exposure, upper bounding surfaces of these
cycles define parasequences.
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 11

FIGURE 3. (A) Map of Tengiz showing the location of seismic line shown in (B). The map shows top reservoir (=top
Bashkirian) structure with a contour interval of 100 m (330 ft). (B) Seismic profile showing general depositional areas of
Tengiz buildup (central platform, outer platform, rim slope, or flank) and key reservoir zones (RZ). Vertical scale is
two-way traveltime in milliseconds. See text for discussion.

Upper Visean–Serpukhovian Supersequence
(Visean A and Serpukhovian Reservoir Zones)
The upper Visean – Serpukhovian supersequence
is defined as the interval between the Lvis_SSB and
the Serp_SSB (Figure 4). It includes the Mikhailovsky,
Venevsky, Tarussky, Steshevsky, Protvinsky, and Za-
paltyubinsky regional horizons (Weber et al., 2003).
The Zapaltyubinsky zone is not generally recognized
on the Tengiz platform but is observed in wells that
penetrate the slope and basin. The Zapaltyubinsky
may therefore represent a basinward shift in sedi-
mentation and form a lowstand sequence set (Weber
et al., 2003). However, rim failure processes (see Collins
et al., 2006) have obscured stratal patterns that clear-
ly define the formation of such a depositional body.
On the platform, the upper Visean – Serpukhovian
supersequence is approximately 250 m (820 ft) thick;
however, equivalent slope deposits (see Collins et al.,
2006) may exceed 600 m (1968 ft), where in-situ mi-
crobial boundstone and allochthonous debris aprons
of platform- and slope-derived grainy carbonate and
breccia accumulated (Figure 3). The upper Visean –
Serpukhovian supersequence spans 14 m.y. (Gradstein
et al., 2004).
The Lvis2_MFS separates the TSS of the super-
sequence from the overlying HSS (Figure 5). The TSS
consists of five composite sequences that are domi-
nantly aggradational and do not extend beyond the
platform break of the underlying lower Visean–upper
Visean supersequence. Small-scale backsteps are pos-
sible, but not proven, at or near the platform break
through the TSS. Deeper water, low-energy open-marine,
platform lithofacies in these sequences include crinoid
grainstone-packstone and poorly sorted, thick-walled
brachiopod grainstone-packstone. Shallower water,
high-energy, open-marine to back shoal lithofacies
 12 / Kenter et al.

FIGURE 4. (A) Map of Tengiz showing the location of cross section shown in (C). The map shows top reservoir (=top Bashkirian) structure with a contour interval
of 100 m (330 ft). (B) Legend for log curves shown in cross section of (C). Gamma-ray (SGR and CGR) is shown on the left and porosity (PHIE) on the right. (C) Cross
section through Tengiz showing major stratigraphic surfaces and reservoir zonation scheme. Stratigraphic surfaces, in ascending order, are Tournaisian (Tour_MFS);
early Visean (Evis_SSB); late Visean (Lvis1_MFS, Lvis_SSB, and Lvis13_csb); Serpukhovian (Serp_SSB); and Bashkirian (Bash_SSB). Reservoir zones bracketed by
these surfaces are Visean D (VisD), Visean C (VisC), Visean B (VisB), Visean A (VisA), Serpukhovian (Serp), and Bashkirian (Bash). The Visean A, Serpukhovian, and
Bashkirian are the subject of this study.
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 13

Table 1. Summary of sequence-stratigraphic nomenclature for Tengiz.*

Stage Reservoir Zones Sequence Top Surface Sequence Number

Bashkirian Unit 1 Bashkirian Bash6 Bash_SSB 1

Bash4 Bash4_csb 2
Bash2.5 Bash2.5_csb 3
Bash1 Bash1_csb 4
Serpukhovian Serpukhovian Serp5 Serp_SSB 5
Serp4 Serp4_csb 6
Serp3 Serp3_csb 7
Serp2 Serp2_csb 8
Serp1 Serp1_csb 9
Late Visean Visean A Lvis13 Lvis13_csb 10
Lvis12 Lvis12_csb 11
Lvis11 Lvis2_MFS 12
Lvis10.5 Lvis10.5_sb 13
Lvis10 Lvis10_csb 14
Lvis9.5 Lvis9.5_sb 15
Lvis9 Lvis9_csb 16
Visean B Lvis8 Lvis_SSB 17
Lvis7 Lvis7_csb 18
Lvis6 Lvis6_csb 19
HRZ Lvis5_csb 20
Unit 2 Visean C Lvis4 Lvis1_MFS 21
Lvis3 Lvis3_sb 22
Lvis2 Lvis2_sb 23
Lvis1 Lvis1_sb 24
Early Visean Evis9 Evis9_sb 25
Evis8 Evis8_sb 26
Evis7 Evis7_sb 27
Evis6 Evis6_sb 28
Visean D Evis5 Evis_SSB 29
Evis4 Evis4_sb 30
Evis3 Evis3_sb 31
Evis2 Evis2_sb 32
Evis1 Evis1_sb 33
Tournasian Tournasian Tour6 Tour_MFS 34
Tour5 Tour5_sb 35
Tour4 Tour4_sb 36
Tour3 Tour3_sb 37
Tour2 Tour2_sb 38
Tour1 Tour1_sb 39
Famennian Unit 3 Famennian A Fame2 Fame_SSB 40
Famennian B Fame1 Fame_MFS 41

*This framework was established by Weber et al. (2003), and picks in wells have, since then, been refined as a result of new wells and
progressed knowledge of the field.

include well-sorted grainstone and grainstone-pack-
stone consisting of algae, foraminifera, peloids, and
coated grains.
The upper Visean–Serpukhovian HSS is subdivided
into an aggradational and progradational phase. The
aggradational interval from Lvis2_MFS to Serp1_csb
(composite sequence boundary) is composed of three
composite sequences (Figure 5). Significant progra-
dation is observed in four subsequent composite se-
quences from the Serp1_csb to the Serp_SSB as shallow
platform lithofacies fill available accommodation
space and prograde to the present-day location of the
14 / Kenter et al.

FIGURE 5. North – south cross section through wells T-5044, T-5246, T-5447, T-220, T-6246, and T-6846; well locations
are shown in Figure 4A. Cross section show detailed lithofacies and cycle correlations in the central platform region in
the Visean A (VisA), Serpukhovian (Serp), and Bashkirian (Bash) supersequences. The lithofacies legend was simplified
from that of Weber et al. (2003) by reducing the number of facies types and relating them to the level of energy
during deposition. Sequence stratigraphy is adopted from Weber et al. (2003); triangles identify higher order sequences
(fourth and higher order) correlated from well to well. See text for further explanation and discussion.
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 15
16 / Kenter et al.
platform break. As much as 400–500 m (1600–1900 ft)
of depositional topography was filled on the slope
as shallow-water platform lithofacies prograded over
predominantly slope boundstone of microbial origin
(Figure 3). During the progradational part (Serpuk-
hovian) of the upper Visean – Serpukhovian HSS, the
Tengiz buildup was an overall flat-topped platform
with a substantial slope apron collecting debris primar-
ily sourced from autochthonous boundstone growing
on the outermost platform and upper slope and, to a
lesser extent, from the platform top. The platform in-
terior is dominated by grainy textures (packstone and
grainstone), with packstone dominating in the deeper
subtidal areas and well-sorted grainstone in the higher
energy shallow subtidal parts of the platform.
Bashkirian Supersequence
(Bashkirian Reservoir Zone)
The Bashkirian supersequence occurs between the
Serp_SSB and the Bash_SSB (Figures 4, 5) and spans
approximately 10 m.y. (Gradstein et al., 2004). The
oldest Bashkirian horizon (i.e., Bogdanovsky) and
most of the overlying Syuransky horizon do not ap-
pear to be present on the Tengiz platform (Brenckle
and Milkina, 2001). The Akavassky and Askynbashsky
horizons are generally present on the Tengiz platform
above the Syuransky horizon. Although more work is
necessary, results suggest that the early Bashkirian at
Tengiz is transgressive, i.e., TSS of the Bashkirian su-
persequence (Figure 5). The HSS of the Bashkirian
supersequence is very thin and condensed. Biostrati-
graphic analysis of core and cuttings typically dates
the Bash_SSB as the top of the early Bashkirian. The
late Bashkirian is condensed and represents starved
sedimentation on an isolated platform, probably de-
posited in deep water (Weber et al., 2003).
Throughout most of the central part of the Tengiz
platform, the Bashkirian supersequence is uniform
in thickness (80 – 100 m; 262 –330 ft). The Bashkirian
platform break is generally aggradational through
three composite sequences (Bash 1, 2.5, and 4) and
coincident with the underlying platform break at the
Serp_SSB. In isolated areas along the eastern margin
of the platform, the Bashkirian section thickens to
about 150 m (492 ft) (e.g., T-5056 and T-7252), with
some of this thickening resulting from a facies change.
Well-log correlations in these areas indicate that the
uppermost composite sequence is locally 30–60 m
(98–196 ft) thick (Weber et al., 2003), seismic data
shows locally mounded seismic facies, and cores show
skeletal packstone and local microbial boundstone on
top of Serpukhovian microbial cement boundstone
several hundreds of meters thick. The remaining
thickening could be explained by differential com-
paction between the massive and mechanically strong
boundstone and the compacted central platform. The
latter is supported by the petrographic observation
that most grainstones are overcompacted.
VISEAN A, SERPUKHOVIAN, AND
BASHKIRIAN PLATFORM ARCHITECTURE
The Visean A, Serpukhovian, and Bashkirian reser-
voir zones are significant because they account for
the bulk of the oil production from the platform part
of the buildup. The platform is subdivided into differ-
ent paleogeographic or depositional areas: central plat-
form, outer platform, and rim slope (flank) (Figure 3).
This chapter focuses on details of the central plat-
form part of the Visean A, Serpukhovian, and Bash-
kirian intervals.
To refine the central platform stratigraphy, a care-
ful comparison of core and thin sections in a series of
north –south wells (T-5044, T-5246, T-5447, T-220,
T-6246, and T-6846) produced a detailed lithofacies
correlation across the central platform region in the
Visean A, Serpukhovian, and Bashkirian reservoir zones
(Figure 5). Several general observations can be made
on the cycle stacking pattern, overall depositional
framework, and facies patterns visible in Figure 5; they
are summarized here and discussed in more detail in
subsequent sections.
The succession of Visean A, Serpukhovian, and
Bashkirian reservoir zones are, respectively, the TSS,
HSS, and TSS of the two supersequences previously
mentioned, and each consists of several composite
and higher order sequences (cycles). Cycle stacking
patterns (shoaling cycles as indicated by triangles in
Figure 5) in the Visean A and Serpukhovian reservoir
zones (and, respectively, TSS and HSS intervals) show
a repetition of a thick cycle (15– 20 m; 49 – 66 ft) with
commonly a basal crinoid- and brachiopod-rich sand
followed by a series of thinner cycles (2 – 8 m; 6.6 –
26 ft). Those crinoid- and brachiopod-rich facies are
interpreted as generally deeper water and relatively
low energy and commonly extend all across the plat-
form, which would suggest significant flooding fol-
lowing a lowstand of relative sea level. Ten to twelve
composite sequences, each containing two to six high-
er order sequences (cycles), are shown in Figure 5.
The boundaries of these composite sequences do not
always exactly match those by Weber et al. (2003),
although the number of composite sequences is es-
sentially the same.
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 17
The composite sequences described herein bundle
into four larger sequences that each have a thick and
laterally continuous brachiopod or crinoid interval
near the base. Minor differences are observed be-
tween wells, but in general, the stacking pattern de-
scribed above is very consistent. In the Bashkirian in-
terval (mostly interpreted as TSS), the stacking pattern
is more complex and shows a lower interval with thick
cycles (10 m; 33 ft), followed by an interval domi-
nated by thin (2–4-m; 6.6–13-ft) cycles, and an interval
near the top of intermediate cycles (4 – 8 m; 13 – 26 ft).
No clear stacking pattern is observed, and although
lateral thickness variations are significant, the succes-
sion of thick, thin, and intermediate cycle intervals
seems to be laterally continuous.
The moderate differences between the sequences
established by Weber et al. (2003) and those described
herein may lead to slight modifications of the plat-
form sequence-stratigraphy framework. One poten-
tial modification is changing the MFS in the Visean
A – Serpukhovian supersequence from its position at
Lvis2_MFS as previously discussed to the base of the
thick brachiopod interval observed between Lvis12_csb
and Lvis13_csb, which marks a shift from a series of
thinning cycles below to thick cycle above and an over-
all change of generally deeper facies below to higher
energy shallow-water facies above (Figure 5).
These sequences of the Visean A, Serpukhovian, and
Bashkirian intervals span a transition from greenhouse
to icehouse cyclicity as suggested by observed trends in
cycle thickness and general depositional environment.
Another important observation is that sequences and
cycles (third order and higher order) display a high
degree of lateral continuity and constant thickness
across the central platform area into the outer platform
zones over distances of more than 10 km (6 mi) north
to south. Careful inspection of this railroad pattern
suggests the presence of a very low-angle top lap and
possible truncation of underlying cycles at the Serp_SSB
surface (the cross section of Figure 5 is flattened on this
level). The truncation is variable but generally increas-
ing to the south, indicating northward paleodip.
In general, it can be stated that the central plat-
form area as documented in this chapter is several
kilometers inboard from the physical break between
platform and flank. The outer platform contains areas
that remain poorly defined because of a lack of core
penetrations and/or uncertainty of the physical and
sedimentological character of the central platform-to-
flank transition. Core penetrations close to the western
edge of the platform are uncommon, but those avail-
able, wells T-44, T-6846, and T-6743, suggest moder-
ate but substantial differences from the central plat-
form area and, therefore, a potentially discontinuous
transition from platform to flank. The situation is more
complicated toward the eastern margin, where appar-
ent top-lap geometry of the Serpukhovian and Bash-
kirian platform develops over a Serpukhovian bound-
stone slope. In the underlying Visean A platform, the
transition may be comparable to that recognized from
the western margin, wherein the Serpukhovian and
Bashkirian central to outer platform transition seems to
be gradual and changes from deeper outer platform with
intercalated boundstone intervals into boundstone
upper slope. As a consequence, the lateral facies asso-
ciations below the pseudo-top-lap may be significantly
different from those above. See Collins et al. (2006)
for further discussion related to these transitions.
Lithofacies Types in the
Central and Outer Platform
Core penetrations available for the Tengiz sequence-
stratigraphic framework developed by Weber et al.
(2003) were limited to TengizChevroil wells T-220,
T-5246, T-6846, and incomplete cores from numer-
ous older Russian wells (figure 3 of Weber et al., 2003).
The present study builds on that of Weber et al. (2003)
by adding the information from cores that nearly com-
pletely span the Visean A, Serpukhovian, and Bashkir-
ian in several key wells in the central (T-5044, T-5447)
and southwestern outer platform (T-6246) (Figure 4A).
These complete core penetrations for the first time
allowed a detailed comparison of each individual cy-
cle (and facies types within) in the study window.
Evaluation of core observations, petrography, and
wire-line and core gamma logs (porosity was exclud-
ed during this phase) resulted in a coherent catalog of
10 key lithofacies types that are ranked primarily based
on the relative hydrodynamic energy during deposi-
tion (Table 2). This information was extracted through
environmental indicators like grain size, sorting, ma-
trix type, sedimentary structures, and grain types (bio-
ta). The resulting 10 facies or rock types can be grouped
into three or four composite rock types, reflecting
major differences in energy conditions during depo-
sition. The catalog was also designed to assist in the
characterization of plugs and petrography of the key
wells at 1-ft (0.3-m) intervals and serve as a direct link
between geological parameters to reservoir properties.
Cyclic Lithofacies Variations
In the central platform, lithofacies stack into
shallowing-upward cycles. The two supersequences
and three reservoir zones making up the Lvis_SSB to
18 / Kenter et al.

Table 2. Summary of lithofacies, pore types, porosity, and bitumen content for central platform wells T-220 and
T-6246.*

Facies Fabric EoD

Abbreviation General Description New Facies Simplified

Code Number Fabric Code

Algal grainstone Algal dominated silt to fine sand G, 46 Grainstone High-energy reef flat,
(AG) very well sorted, with minor coated grains backshoal-inner platform
(smaller than 500 mm)
Skeletal intraclast Mixed skeletal-(peloidal) G, moderately well 44 Grainstone High-energy reef flat,
grainstone (SIG) sorted, fine to medium sand, varying shell backshoal-inner platform
hash, minor crinoids, abundant intraclasts,
oncoids, coral (massive) and Chaetetes
fragments, forams, algae
Coated grain Ooid-dominated G, well to very well sorted, 43 Grainstone High energy above
grainstone-rudstone minor crinoids or shell hash, intraclasts, wave base-intertidal
(CgGR) forams; locally rundstone with intraclasts; platform
cross-lamination
Ooid grainstone Skeletal G, fine to medium sand, well to very 42 Grainstone High energy above
(OG) well sorted, coated grains, minor crinoids or wave base-intertidal
shell hash, forams, intraclasts, algal fragments, platform
locally overpacked; locally rudstone with
intraclasts; cross-lamination
Well-sorted Mixed skeletal-(peloidal) G, moderately well 41 Grainstone Moderate to High
skeletal sorted, fine to medium sand, minor to energy above wave
grainstone (SPG) abundant shell hash, moderate crinoids, base-open platform
intraclasts, forams, algae
Poorly sorted Skeletal-peloidal GP, coarse sand, moderately 32 Grainstone- Moderate to high
skeletal-peloid to poorly sorted, mixed crinoids and packstone energy below or around
grainstone-packstone brachiopod shell hash, minor in-situ wave base-open platform
(SPP) (thick-walled) brachiopods, peloids, algae,
forams, whole and fragments of stick corals
Crinoid Skeletal-peloidal GP, medium to coarse sand, 31 Grainstone- Moderate to low energy
skeletal-peloid moderately to poorly sorted, dominant packstone below or around wave
grainstone-packstone crinoids (1 – 2 mm or larger), minor whole base-open platform
(CSP) brachiopods and hash, peloids, forams,
(green and tubular) algal fragments
Brachiopod Skeletal-peloidal GP, coarse sand, poorly 2 Grainstone- Moderate to low energy
skeletal-peloid sorted, brachiopods in-situ (thin vs. packstone below or around wave base
grainstone- thick-walled; 1-5/20-50), bedded, variable open platform (abundant
packstone crinoids, peloids, divers algal fragment, crinoids an thick-walled
(BSP) abundant and divers forams, whole and brachiopods) versus shallow
fragments of stick corals protected platform interior
(minor crinoids and
thin-walled brachiopods)
Peloid Ranging from mud to peloidal PW, 13 Packstone- Low-energy restricted;
packstone- silt-fine sand size with small oncoids, grainstone restricted lagoon to
wackestone calcispheres, thin-walled bivalve supratidal
(PPW) fragments and ostracods
Volcanic ash Volcanic ash 12 Clay mudstone Low-energy restricted
(VA) inner platform to
postdrowning
All
Grain-supported mud-lean; high energy 41 – 46
Grain-supported muddy; intermediate 2, 31, 32
energy
muddy; low energy 13

*Lithofacies, also shown in Figure 5, are based on Weber et al. (2003), but simplified by reducing the number of facies and relating them to the level
of energy during deposition. The vertical arrangement of lithofacies in this table resembles the ideal and most complete succession of lithofacies in a
shoaling cycle or sequence. Pore types are 1 = interparticle; 2 = intraparticle; 3 = fenestral; 4 = shelter; 5 = growth framework; 6 = intercrystalline; 7 =
moldic; 8 = microporosity; 9 = fracture; 10 = channel; 11 = vug; and 12 = enhanced dissolution. Note that bitumen analyses for well T-6246 are not
yet complete. See text for discussion.
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 19

T-220 T-6246

N Porosity Bitumen (% Rock Volume) N Porosity

Dominant Range Mean Range Mean Dominant Range Mean

Pore Types (%) (%) Pore Types (%) (%)

24 11, 1, 6, 2 2.25 – 17.59 13.08 0.16 – 0.31 0.23 54/48 1, 7, 6, 11, 2, 9 0.84 – 14.80 4.90

18 0.18 – 16.35 4.46 0.08 – 0.41 0.28 22/22 1, 2, 7 0.97 – 12.07 4.58

78 11, 1, 7 0.42 – 23.66 10.09 0.13 – 6.86 1.35 36/36 1, 7, 11, 6, 9, 2 0.37 – 14.91 4.43

100 11, 1, 7, 2 0.78 – 21.36 11.79 0.01 – 2.73 0.40 103/78 1, 7, 6, 11, 2, 9 0.41 – 20.11 4.74

188 11, 1, 7, 6 0.19 – 23.18 11.64 0.04 – 6.36 0.23 217/114 1, 11, 2, 6, 7, 8 1.21 – 16.18 6.81

217 11, 1, 6, 7, 2 0.50 – 17.33 12.56 0.06 – 4.04 0.35 262/90 1, 11, 2, 9, 6, 7 0.97 – 15.30 6.91

83 11, 6, 1, 7, 9 0.34 – 14.79 5.37 0.05 – 5.54 2.58 79/14 1, 2, 11, 6, 7, 8 1.30 – 18.48 8.28

199 11, 1, 6, 2 1.45 – 17.53 10.51 0.09 – 5.49 0.68 207/27 11, 1, 6, 9, 2, 7 0.63 – 16.50 8.42

53 11, 9, 1, 6, 7, 2 0.07 – 17.55 3.62 0.11 – 3.72 0.67 49/14 1, 2, 11, 6, 7, 9 1.02 – 15.50 8.16

971 11, 1, 6, 7, 2, 9 0.07 – 23.66 10.44 0.01 – 6.87 0.74 1031 1, 11, 2, 6, 7, 9 0.37 – 20.11 6.89
408 11, 1, 7, 2, 6 0.18 – 23.66 11.15 0.01 – 6.87 0.24 432 1, 11, 7, 6, 2, 9, 8 0.37 – 20.11 5.77

510 11, 6, 1, 2, 7 0.34 – 19.34 10.58 0.05 – 5.54 0.84 549 11, 1, 2, 6, 7, 9, 8 0.63 – 18.48 7.67

53 11, 9, 1, 6, 7, 2 0.07 – 17.55 3.62 0.11 – 3.72 0.67 48 1, 2, 11, 6, 7, 9 1.02 – 15.50 8.16
20 / Kenter et al.
Bash_SSB interval comprise more than 50 cycles rang-
ing in thickness from 2 to 15 m (6.6 to 49 ft) (Table 1;
Figure 5). Commonly, cycles are bounded by thin vol-
canic ash layers less than a few centimeters in thick-
ness that have been altered by burial diagenesis to
resemble shale partings. Closely spaced volcanic ash
partings are sometimes observed near bounding sur-
faces. Thicker ash beds do occur in some parts of the
section (e.g., just above the Lvis_SSB and the Serp_SSB)
and may exceed 1 m (3.3 ft) in thickness. Laminated
ash beds were likely deposited (and preserved) during
periods of low energy, either during exposure of the
platform top or during the initial flooding and devel-
opment of low-energy lagoonal deposits. The pres-
ence of thick (decimeter to meter) intervals of vol-
canic ash indicates major periods of exposure and/or
deep initial flooding. These are therefore key sur-
faces that guided the selection of the supersequence
boundaries.
When presented in the context of their deposi-
tional setting related to energy (i.e., bottom agitation)
at the time of deposition, lithofacies roughly form a
succession of low to higher energy from the base to
top of a depositional cycle (see Table 1). One caveat in
relating the lithofacies with energy level is that there
is no one-to-one correspondence. For example, simi-
lar low-energy facies occur both in deeper subtidal as
well as in more restricted shallow depositional envi-
ronments. For discussion purposes, such a succession
is shown in Table 1 and partly demonstrates litho-
facies stacking from base to top. A brief description
is provided in the sections immediately following,
whereas the implications for the general depositional
system (spatial distribution of lithofacies types and
cycle thickness) are discussed in subsequent sections.
In brief, the description of lithofacies types and
associated depositional energy, from base to top, is
as follows: cycles generally start with a thin peloid
packstone-wackestone that exhibits subaerial expo-
sure features and are commonly interbedded with
thin volcanic ash layers. These lithofacies are over-
lain by brachiopod and/or crinoid skeletal-peloid
grainstone-packstone, which are, in turn, overlain by
poorly sorted skeletal-peloid grainstone-packstone
and/or well-sorted skeletal grainstone. This succes-
sion is, in turn, overlain by coated-grain grainstone-
rudstone and/or ooid grainstone and locally a skeletal
intraclast grainstone or algal grainstone. Obviously,
deviations from this generalization exist, and cycle
thickness, stacking patterns, and relative contribution
of each facies type vary within and across the major
sequences.
Peloid Packstone-wackestone (PPW) and
Volcanic Ash (VA)
Peloid packstone-wackestone caps many shallowing-
upward carbonate cycles and can overlie any of the
other lithofacies. Usually less than 0.5 m (1.6 ft) thick,
this lithofacies exhibits rhizoliths, laminated crusts,
alveolar fabric, uncommon fenestrae, low-diversity
microfauna (ostracods and calcispheres), and uncom-
mon megascopic biota (Figure 6). Peloid packstone-
wackestone includes the development of a pedolithic
zone that occurs at most cycle boundaries, which is
interpreted to include both the top of a cycle prior to
exposure and the exposure event itself, as well as the
initial reflooding event with intermittent exposure.
Peloid packstone-wackestone nearly always has volca-
nic ash interbeds or dispersed volcanic ash (Figure 6A)
and, probably as a result, is tight and well cemented.
Volcanic ash is locally also dispersed within high-
energy facies immediately below the Serp_SSB and
below the Serp1_csb in well T-5447 (Figure 5). Those
intervals are similarly tight and well cemented. The
frequent occurrence of volcanic ash interbeds argues
for nearly continuous fallout during deposition of
the sequences; ash layers accumulated only during
subaerial or low-energy conditions, thereby marking
cycle boundaries on the gamma-ray logs. The pres-
ervation of volcanic ash is related to terrestrial fall-
out deposition during exposure and during initial
low-energy conditions while the platform was sub-
merged, probably in a shallow, restricted lagoon.
Brachiopod and Crinoid Skeletal-peloid
Grainstone-packstone (BSP and CSP)
An inferred rise in relative sea level resulted in
slightly deeper conditions and deposition of skeletal
grainstone to packstone at the base of a cycle. The basal
part of the cycles can be either brachiopod skeletal-
peloid grainstone-packstone or crinoid skeletal-peloid
grainstone-packstone, but brachiopod skeletal-peloid
grainstone-packstone is dominant.
Brachiopod skeletal-peloid grainstone-packstone is
interpreted to represent the continued flooding of the
platform in relative low-energy conditions. Thick accu-
mulation of such deposits could indicate deeper water
or ongoing restriction caused by emergent outer plat-
form shoals or barriers. The facies is initiated during ini-
tial flooding of the platform and is most robust during
times of maximum flooding. Brachiopod skeletal-peloid
grainstone-packstone is poorly sorted, composed of large,
commonly in growth position, thick-walled brachio-
pods, variable crinoids, diverse algal fragments and
foraminifera, and/or corals (Figure 7). Brachiopod
FIGURE 6. (A) Core photos from well T-5246 show peloid packstone-wackestone (PPW) alternating with volcanic ash layers (VA) below the Lvis_SSB. The boundary
that shows evidence for subaerial exposure is overlain by poorly sorted skeletal-peloid grainstone-packstone (SPP) with elevated bitumen cement in the lower
70 cm (27 in.) (see Table 2 and Figure 5 for additional lithofacies information). Photomicrographs show the range of PPW textures. Common skeletal grains are
tubular algal fragments (B, D), calcispheres (B, E), and minor coated skeletal grains and aggregrate grains (C, D). Width of photomicrographs is 4.20 and 1.58 mm
(0.16 and 0.062 in.) (B, C – E, respectively).
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 21
 22 / Kenter et al.

FIGURE 7. (A) Core photos from well T-5246 show shoaling succession of skeletal-peloid grainstone-packstone (SSP) and well-sorted skeletal grainstone (SPG)
immediately below the Lvis_9.0_csb. The boundary is overlain by peloid packstone-wackestone (PPW) and a thick interval of brachiopod-skeletal-peloid
grainstone-packstone (BSP). Concentrations of in-situ brachiopods and local development of intraparticle and shelter porosity are clearly visible in the BSP
interval. Photomicrographs (B – E) show the generally poorly sorted texture and microfractured nature of BSP. Brachiopods are dominant constituents and range
from small thin-walled (less than 2 cm [0.8 in.] long and 1 mm [0.04 in.] thick) to large and thick-walled species (respectively, 5– 15 cm [1.9– 5.9 in.] and 1– 10 mm
[0.04 and 0.40 in.]). Note the sutured, stylotized contacts of the brachiopods. Other common grain types are crinoid ossicles, peloids, algal fragments, and
benthic foraminifera. Width of photomicrographs is 4.20 mm (0.16 in.).
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 23
skeletal-peloid grainstone-packstone intervals are gen-
erally continuous across the platform, appear to thin
toward the outer platform, and have the highest fre-
quency and volume in the upper part of the Visean A
(between Lvis_13 and Lvis_10.5), where they generally
form the basal parts of most cycles (Figure 5). Intervals
of brachiopod skeletal-peloid grainstone-packstone
occur sporadically in the lower part of the Visean A
platform. Two very distinct and thick-bedded (4–10-m;
13–33-ft) intervals, one right above the Lvis9_csb and
one below the Lvis13_csb, blanket the entire platform.
Crinoid skeletal-peloid grainstone-packstone is coarse
and poorly to moderately sorted crinoid-dominated
skeletal-peloidal sand. The crinoids commonly have
destructionally micritized outer surfaces; other com-
mon components are mostly broken brachiopods, green
and red algal fragments, and foraminifera (Figure 8).
Crinoid skeletal-peloid grainstone-packstone inter-
vals are generally not continuous across the platform
and seem to preferentially occur near the platform
margin (wells T-6846, T6246, T-8, and T-44). Crinoid
skeletal-peloid grainstone-packstone is interpreted as
being deposited in moderate to low energy below or
around wave base and may represent the time of maxi-
mum flooding. These intervals represent the maxi-
mum open-marine influx during the cycle, and only
one interval (right above Lvis_10.5) covers the entire
platform (Figure 5).
Poorly Sorted Skeletal-peloid Grainstone-packstone
(SPP) and Well-sorted Skeletal Grainstone (SPG)
Poorly sorted skeletal-peloid grainstone-packstone
and well-sorted skeletal grainstone generally occur
above brachiopod skeletal-peloid grainstone-packstone
and crinoid skeletal-peloid grainstone-packstone, exhibit
progressively better sorting and less mud, and therefore
have been interpreted as shallow-water wave-agitated
sand shoal deposits (Figures 9, 10, respectively). These
grainstone lithofacies dominate the cycles for most of
the Visean A reservoir and commonly show lateral
shifts from one facies into the other. In the T-5246 well,
well-sorted skeletal grainstone dominates the cycles,
and there seems to be a general subtle increase away
from the central platform (Figure 5). Immediately above
the Lvis2_MFS, a southward thinning interval of well-
sorted skeletal grainstone covers the entire platform.
Coated-Grain Grainstone (CgGR), Ooid Grainstone,
(OG), and Skeletal-intraclast Grainstone (SIG)
In a complete cycle, the previously described litho-
facies are overlain by cross-bedded or uniformly lami-
nated, well-sorted coated-grain and/or ooid grainstone,
which are interpreted as high-energy shallow subtidal
open-platform to intertidal deposits (Figures 11, 12,
respectively). Ooids are generally small, ranging from
0.5 to 1.0 mm (0.02 to 0.04 in.), and were most prob-
ably calcitic in origin from their radial structure.
Coated grains are commonly skeletal fragments with
coatings that are micritic and mainly constructional.
Coated-grain grainstone has uncommon occurrences
in the Visean A but is common in the Serpukhovian
reservoir (Figure 5), where it is concentrated in the
northern part of the platform amplifying a general
trend from higher energy in the north to lower en-
ergy (and/or open marine) toward the south. Coated-
grain and/or ooid grainstones are locally overlain by a
relatively thin interval of coarse-grained lithoclastic
rudstone-grainstone, skeletal-intraclast grainstone,
with centimeter-size subangular to well-rounded intra-
clasts and Chaetetes (demosponge) fragments, which
are commonly coated by micrite (Figure 13). Some in-
tervals display clast-supported flat pebbles, indicative
of beach environments. Dominant grain types range
from coated grains to ooids and algal grains.
Algal Grainstone (AG)
Finally, very well-sorted and cross-bedded, silt to
fine sand-size, algal grainstone occurs near the top of
the Bashkirian sequences and locally in the Visean A
sequence (Figures 5, 14). Algal grainstone intervals
were also observed in the outermost platform (Serpu-
khovian) of the T-4556 well and may indicate the oc-
currence of this facies in a high-energy outer platform
setting. Elsewhere on the platform, algal grainstone
alternates with more open-marine well-sorted skel-
etal grainstone or coated-grain grainstone-rudstone;
in these instances, algal grainstone is interpreted as
being deposited in the intertidal zone near the top of
cycles, perhaps sourced from the outer platform and
filling the remaining accommodation space.
Lithofacies Mosaics
Figure 15 shows hypothetical depositional models
of higher order depositional cycles for both the Bash-
kirian sequence (Figure 15A) as well as the Visean A
and Serpukhovian sequences (Figure 15B). The mod-
els show grain compositional and textural trends from
the central platform through the outer platform and
into the upper flank and indicate some associated
bathymetric relief during the Visean A and Serpu-
khovian. In the central platform, Visean A and Serpu-
khovian platform cycles are vertically and laterally
relatively predictable and show a succession, from
base to top, that reflects depositional energy. Near
 24 / Kenter et al.

FIGURE 8. (A) Core photos from well T-5246 show succession of well-sorted coated-grain grainstone (CgGR) overlain by peloid packstone-wackestone (PPW)
and capped by the Lvis_10.5_csb. The boundary is, in turn, overlain by crinoid skeletal-peloid grainstone-packstone (CSP) and poorly sorted skeletal-peloid
grainstone-packstone (SPP) (see Table 2 and Figure 5 for additional lithofacies information). The CSP interval is only moderately stained by bitumen here in
T-5246, whereas toward the outer platform, it is commonly completely stained black. Crinoid skeletal-peloid grainstone-packstone is coarse and poorly to
moderately sorted crinoid-dominated skeletal-peloidal sands (B) with destructionally micritized outer surfaces (C) and common components as (mostly broken)
brachiopods, (green and red) algal fragments and foraminifera (B – E). Common grain types are crinoid ossicles, brachiopod hash and whole valves, peloids, algal
fragments (notably Koninckopora), benthic foraminifera. Width of photomicrographs is 4.20 mm (0.16 in.).
FIGURE 9. (A) Core photos from well T-5246 show a thick interval of poorly sorted skeletal-peloid grainstone-packstone (SPP) at the top. The lower half of the core
contains a shoaling succession of well-sorted skeletal grainstone (SPG) capped by Lvis_9.5_csb. The boundary is overlain by peloid packstone-wackestone (PPW)
with a thin volcanic ash ( VA) interbed near the top (see Table 2 and Figure 5 for additional lithofacies information). The skeletal-peloid grainstone-packstone (SSP) and
well-sorted skeletal grainstone (SPG) are generally located above the B- and CSP; because they exhibit progressively better sorting and less mud, they have been
interpreted as shallow-water wave-agitated sand shoal deposits. Common grain types found in SPP shown in (B – E) are crinoid ossicles, brachiopod hash, peloids, algal
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 25

fragments, and benthic foraminifera. Grains are commonly coated by thin micrite rims and are poorly sorted. Width of photomicrographs is 4.20 mm (0.16 in.).
 26 / Kenter et al.

FIGURE 10. (A) Core photos from well T-5246 show shoaling succession of well-sorted skeletal grainstone (SPG) capped by Lvis_10.0_csb. The boundary is overlain
by thin peloid packstone-wackestone (PPW), well-sorted skeletal grainstone (SPG), and peloid packstone-wackestone (PPW), with large thin-walled brachiopods
(see Table 2 and Figure 5 for legend). Well-sorted skeletal grainstones (SPG) are generally located above the BSP and CSP lithofacies, exhibit better sorting and less
mud, and have been interpreted as shallow-water wave-agitated sand shoal deposits. Photomicrographs (B – E) show common grain types of SPG: crinoid ossicles,
brachiopod hash, peloids, algal fragments, and benthic foraminifera. Width of photomicrographs is 4.20 mm (0.16 in.) (B, C, E) and 1.58 mm (0.06 in.) (D).
FIGURE 11. (A) Core photos from T-5246 show shoaling succession of a thick interval of cross-beds, but laminations are obscured by bioturbation, well-sorted
coated-grain grainstone (CgGR), which is overlain by a thin peloid packstone-wackestone (PPW). The Lvis_10.5_csb is overlain by coarse and poorly to moderately
sorted crinoid-dominated skeletal-peloidal sands (CSP) (see Table 2 and Figure 5 for legend). In a complete cycle, mixed skeletal-peloidal sands (mostly SPG) are
overlain by cross-bedded or uniformly laminated, well-sorted coated-grain and/or ooid grainstone (CgGR and OG), interpreted as high-energy shallow subtidal
open platform to intertidal deposits. Common grain types in CgGR shown in (B – E) are well-sorted peloids, coated skeletal grains, algal fragments (tubular), and
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 27

benthic foraminifera; crinoid ossicles and brachiopod hash are minor grains. Width of photomicrographs is 4.20 mm (0.16 in.) (D) and 1.58 mm (0.06 in.) (B, C, E).
 28 / Kenter et al.

FIGURE 12. (A) Core photos from T-5246 show part of an approximately 5 – 6-m (16 – 19-ft)-thick interval of cross-bedded, well-sorted ooid grainstone (OG). An
exposure surface, with small intraclasts, separates a dark, bitumen-rich interval above from the underlying better cemented, yellow-cream-colored interval
in the rightmost core slab (see also Table 2 and Figure 5 for legend). In the Bashkirian, a complete cycle of mixed skeletal-peloidal sands (mostly SPG) is generally
overlain by cross-bedded ooid grainstone lithofacies (OG) interpreted as high-energy shallow subtidal open platform to intertidal deposits. Photomicrographs
(B – E) show the dominant grain types of ooids and coated skeletal grains; algal fragments (tubular) and benthic foraminifera also occur. Width of photomicrographs
is 4.20 mm (0.16 in.) (B, D) and 1.58 mm (0.06 in.) (C, E).
FIGURE 13. (A) Core photos from T-5246 show a shoaling succession of well-sorted skeletal grainstone (SPG), overlain by a relatively thin interval of coarse-
grained skeletal-intraclast grainstone (SIG), and capped by the Serp1_csb. The boundary is overlain by thin intervals of PPW and SPG, brachiopod-skeletal-peloid
grainstone-packstone (BSP), and another interval of SPG (see Table 2 and Figure 5 for legend). (B– D) Skeletal-intraclast grainstone (SIG) typically has centimeter-
size subangular to well-rounded intraclasts and Chaetetes fragments, which are commonly coated by micrite. Other grains are coated grains, ooids, and algal fragments.
Width of photomicrographs is 4.20 mm (0.16 in.).
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 29
 30 / Kenter et al.

FIGURE 14. (A) Core photos from T-5246 show a thick interval of very well-sorted and cross-bedded, silt- to fine sand-size algal grainstone (AG) underlain
and overlain by well-sorted coated-grain grainstone (CgGR); an interbed of volcanic ash (VA) occurs at the bottom (see Table 2 and Figure 5 for legend).
Photomicrographs (B-E) show dominant grain types of AG: algal fragments, peloids, and uncommon benthic foraminifera. Typical algal species are Donezella
(B, C), paleoberesellids such as Kamaena (D), and beresellids such as Beresella (E). Paleoberesellids are typically Visean and beresellids and Donezella occur in the
Serpukhovian to Bashkirian. Width of photomicrographs is 4.20 and 1.58 mm (0.16 and 0.06 in.) (B, C-E, respectively).
FIGURE 15. Models of hypothetical higher order depositional cycles for the Bashkirian (A) and the Visean A – Serphuhovian (B) showing grain composition and
textural trends from the central Tengiz platform through the outer platform and into the flank. TST = transgressive systems tract; HST = highstand systems tract.
See text for discussion.
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 31
32 / Kenter et al.
the base, a thin, peloid packstone-wackestone that
exhibits subaerial exposure features is commonly in-
terbedded with thin volcanic ash layers. These litho-
facies are overlain by brachiopod bioherms and/or
crinoid skeletal-peloid grainstone-packstone, which
are, in turn, overlain by poorly sorted skeletal-peloid
grainstone-packstone and/or well-sorted skeletal grain-
stone. These lithofacies reflect lowstand and early
transgressive restriction providing the environment
for trapping volcanic ash fallout (peloid packstone-
wackestone) and maximum flooding during the cri-
noid skeletal-peloid grainstone-packstone. This succes-
sion is succeeded by highstand deposits that indicate
shoaling to (nearly) sea level: coated-grain grainstone-
rudstone and/or ooid grainstone and locally a skeletal
intraclast grainstone or algal grainstone. In the outer
platform toward the east, cycles lack the lagoonal base
and have a microbial boundstone interval (type 3 micro-
bial boundstone; see Collins et al., 2006) instead, which
is superceded by thicker poorly sorted mixed skeletal-
peloidal and crinoid sand bodies. Some cycles include
thin deposits of high-energy coated-grain grainstone-
rudstone and/or ooid grainstone and skeletal intra-
clast grainstone or algal grainstone. Brachiopod bio-
herms, crinoid skeletal-peloid grainstone-packstone,
and poorly sorted skeletal-peloid grainstone-packstone
and/or well-sorted skeletal grainstone are absent.
The overlying Bashkirian cycles (Figure 15A) are
thinner and, in the central platform, show alternating
lagoonal peloid packstone-wackestone with subaerial
exposure features and interbedded with thin volcanic
ash layers, coated-grain grainstone-rudstone and/
or ooid grainstone and skeletal intraclast grainstone.
Above the Bash2.5_MFS, coated-grain grainstone-
rudstone and/or ooid grainstone alternate with algal
grainstone intervals, and volcanic ash beds are nearly
absent. However, toward the outer platform areas, low-
energy transgressive deposits (brachiopod skeletal-
peloid grainstone-packstone, crinoid skeletal-peloid
grainstone-packstone, poorly sorted skeletal-peloid
grainstone-packstone) thicken and may even include
microbial boundstone intervals (T-4556), while high-
energy cycle tops (coated-grain grainstone-rudstone,
ooid grainstone) are thinning or absent (T-6743).
Obviously, deviations from the generalizations of
Figure 15 exist, as will be discussed in the following
sections. Cycle thickness, stacking patterns, and relative
contribution of each facies type vary within and across
the major sequences. As an example, the bathymetric
profile changes toward the Visean A–Serphukhovian
outer platform to a deeper setting that brings microbial
boundstone on the platform during flooding. In con-
trast, the Bashkirian sequence shows a flat-topped plat-
form with parallel facies zones dominated by higher
energy grainy lithofacies types. High-frequency cycles
are much thinner in the Bashkirian, at meter scale,
and show mostly alternating coated grain grainstone,
ooid grainstone, and algal grainstone with sedimen-
tary structures. The Bashkirian rim was probably flat
during deposition and shows thinning of platform
facies that border a nearly starved slope system.
Visean A and Serpukhovian:
Trends and Depositional Environments
Visean A cycles consist almost entirely of grainy
facies (packstone and grainstone) that feature depo-
sitional environments ranging from restricted low-
energy brachiopod skeletal-peloid grainstone-packstone
to moderate- to low-energy deeper subtidal envi-
ronments characterized by more poorly sorted and
open-marine poorly sorted skeletal-peloid grainstone-
packstone as well as crinoid skeletal-peloid grainstone-
packstone and high-energy well-sorted grainstone shoals
(well-sorted skeletal grainstone and uncommon coated-
grain grainstone-rudstone).
General trends in lithofacies distribution, cycle
thickness, and stacking suggests the presence of three
intervals in the Visean A and Serpukhovian reservoir
zones, each displaying certain depositional trends
(Figure 5). These are the intervals between the Lvis_SSB
and Lvis10.5_csb, Lvis10.5_csb and Lvis13_csb (which
straddles the second-order Lvis2_MFS), and Lvis13_csb
to the Serp_SSB.
In the Lvis_SSB to Lvis10.5_csb interval, the plat-
form cycles are dominated by the poorly sorted, open-
marine poorly sorted skeletal-peloid grainstone-
packstone, with minor intercalations of brachiopod
skeletal-peloid grainstone-packstone and/or crinoid
skeletal-peloid grainstone-packstone. Crinoid skeletal-
peloid grainstone-packstone intercalations generally
occur near the base of cycles, whereas well-sorted skel-
etal grainstone intercalations appear near the top.
Cycle caps are thin, decimeter-thick, micropeloidal
packstone-mudstone intervals that have, in nearly all
cases, one or two thin volcanic ash interbeds, but
generally display moderate subaerial exposure. Cycle
thickness ranges from 1–2 m (3.3–6.6 ft) near the
base to more than 15 m (49 ft) in the middle of the
interval with an average of 5–8 m (16– 26 ft). Poorly
sorted skeletal-peloid grainstone-packstone intervals
are dominant in the central platform around wells
T-220 and T-5848 but gradually are partially replaced
by higher energy sand shoals (well-sorted skeletal
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 33
grainstone) toward the outer platform areas (well
T-5246 to the north and wells T-6246 and T-6846 to
the southwest). Crinoid-rich intervals represent maxi-
mum, open-marine flooding and the deepest litho-
facies. These occur preferentially in the outer platform
toward the southwest and seem to thin and pinch out
toward the central platform. One interval, however,
runs contrary to this trend and appears to blanket
the entire platform at the top of the Lvis_SSB to
Lvis10.5_csb interval. Similarly, one nearly 10-m (33-ft)-
thick and very continuous brachiopod skeletal-peloid
grainstone-packstone interval covers the platform right
above the Lvis9_csb boundary. This general deposi-
tional picture suggests an overall low-energy platform,
subtidal or below wave base, with slightly shallower
outer platform sand shoals and a dominant southern
open-marine influence. The cycles in this succession
never filled the total accommodation space, despite
the relatively low-amplitude sea level oscillations.
The Lvis10.5_csb to Lvis13_csb interval, which strad-
dles the supersequence flooding surface (Lvis2_MFS),
shows a similar central versus outer platform facies
pattern as the previous interval but has much thinner,
2–5-m (6.6–16-ft)-thick cycles and frequent brachio-
pod skeletal-peloid grainstone-packstone intervals.
These brachiopod interbeds seem to thin away from
the platform center and are replaced by higher energy
mixed skeletal-peloidal sands to the south and high-
energy, well-sorted sands to the north. In addition,
the section has a crinoid-rich base, a thick interval of
high-energy sand shoals in the middle, and a thick
brachiopod interval (more than 10 m) near the top.
Each of these intervals is continuous across the entire
platform. The crinoid-dominated base indicates max-
imum deepening of the platform coinciding with the
Lvis_MFS surface. Although relative depositional con-
ditions did not change and accommodation space
was still underfilled, it suggests that sea level fluctua-
tions increased in frequency.
The Lvis13_csb to the Serp_SSB interval in the HSS
of the supersequence has similar thin cycles, 2– 5 m
(6.6– 16 ft) in thickness, but those show more pro-
nounced subaerial exposure near the cycle tops. Fur-
thermore, there is an overall trend of higher energy
coated-grain and ooid sand shoal deposits with ele-
vated amounts of dispersed volcanic ash toward the
north (wells T-5447, T-5044, and T-5444) and slightly
deeper, lower energy, environments toward the south,
near the outer platform (wells T-6246 and T-6846).
Cycles were nearly completely filled in the northern
platform area, multiple ash layers occur near sequence
boundaries, and there is an increased expression of
exposure toward the Serp_SSB. These observations
suggest the development of a slightly deeper and lower
energy outer platform toward the southeastern mar-
gin (possibly caused by structural tilting). The thin-
ning of cycles, increasing exposure of cycle tops, and
lateral facies differentiation during the Serpukhovian
coincide with the development of an unusual deeper
water outer platform toward the northeastern to east-
ern margins of the Tengiz platform, where upper
slope facies consisting of algal-microbial, skeletal-
microbial, and microbial boundstone prograded sev-
eral kilometers out into the basin on a preexisting
Visean A and B shelf edge. In these regions of the outer
platform, which generally have lower energy and
probably deeper water facies, exposure expression is
reduced or absent (see Figure 15).
Bashkirian: Trends and
Depositional Environments
Bashkirian cycles are thinner and less systematic
than those recognized in the Visean A and Serpu-
khovian, most likely because of rapid sea level fluc-
tuations produced during global icehouse conditions.
The Bashkirian consists of three intervals, of which
the lower (Serp_SSB to Bash1_csb) is essentially a 15–
20-m (49–66-ft)-thick ooid sand body covering the
entire platform (Figure 5). This unit is nearly tabular in
shape, has locally thin, mixed skeletal-peloidal open-
marine intercalations, and is cyclic with thin mud-
stone cycle caps and associated exposure surfaces. In
the central platform, the lower half of the interval
commonly contains high bitumen content and has
low porosity; toward the southwestern outer platform
(T-6246) and north (T-5246), nearly the entire section
has bitumen and strongly reduced porosity. The mid-
dle part of the Bashkirian is bounded above by a
stratigraphic surface (Bash2.5_MFS) that was not part
of the original framework published by Weber et al.
(2003). The Bash1_csb to Bash2.5_MFS is dominated
by thin (a few meters) shoaling cycles of sorted skel-
etal subtidal open-marine sands, overlain by high-
energy intertidal coated grain to ooid sand shoals,
and capped by thin peloidal packstone-mudstone in-
tervals with general extensive subaerial exposure and
intercalated centimeter- to decimeter-thick volcanic
ash beds. In the uppermost part of the Bashkirian
(Bash2.5_MFS to Bash_SSB), ooid to coated grain sands
occur in the basal part of shoaling cycles and are
overlain by silt-size (high-angle), cross-bedded, well-
sorted algal sands. Lagoonal packstone-mudstone inter-
vals and associated volcanic ash beds are uncommon or
34 / Kenter et al.
lacking completely. The algal fragments are silt-size,
dominant, and probably tubular red algae in origin.
The waning of ooid domination and introduction
of algal silt above the Bash2.5_MFS may indicate a
general deepening stratigraphic trend toward the
Bash_SSB, which is consistent with an interpretation
of the subsequent drowning of the platform (Weber
et al., 2003). Cross-bedding was possibly generated by
oceanic currents sweeping the deeper water platform
top during the progressively longer highstands; the
platform was still exposed during shorter lowstands
associated with high-frequency and high-amplitude
(as much as 40–60-m [131–196-ft]) sea level changes
of the icehouse world. Expressions of subaerial ex-
posure, such as dissolution, recrystallization, and pres-
ervation of ash layers are also nonuniform from cycle
to cycle. Despite this, Bashkirian cycles retain some
similarities to Visean A and Serpukhovian cycles, such
as the presence of high-energy shoals and deeper
platform skeletal grainstone and the association of
semirestricted to restricted facies near sequence bound-
aries. In the Bashkirian, high-energy shoals are rep-
resented by oolites, pisolites, and coated grains, whereas
the deeper and restricted facies are similar to those
present in the Serpukhovian and Visean A.
Interplay between Minor Tilting
and Deposition
A remarkable observation from the cross section in
Figure 5 is that the railroad-track pattern of sequence
and cycle boundaries below the Serp_SSB is nearly
perfectly parallel, with the exception of the section in
T-6846. Another observation is the presence of a minor
angular unconformity of several degrees between the
Serp_SSB and the underlying cycle breaks, wherein
both the Serp_SSB and cycle breaks dip to the south a
few degrees. In addition, the Serp_SSB appears to have
variable erosional relief of as much as 5 m (16 ft)
across the platform as indicated by the apparent in-
creased erosion in well T-5447. These observations sug-
gest several things. First, fairly gradual and constant
vertical aggradation across the central platform re-
sulted in a flat platform top with minimal local relief as
a result of immediate compensation of possible varia-
tions in depositional topography. In addition, the plat-
form was tilted to the south prior to the Serp_SSB, sub-
sequently eroded (based on the missing Bogdanovsky
and Suryansky biozones), generating as much as 5-m
(16-ft) relief, and finally, tilted a few more degrees
along the same axis during the post-Bashkirian. Well
T-6846 shows about a 5-m (16-ft) thicker section rela-
tive to other wells of Figure 5 between the Lvis10.5_csb
and Lvis_SSB, a nearly 5-m (16-ft) thicker section be-
tween the Lvis13_csb and Lvis10.5_csb, probable minor
erosion at the Lvis13_csb, and several additional meters
missing section below the Serp_SSB. Although not
proven, these differences may indicate that the south-
western margin tilted downward some 10 m (33 ft) or
more prior to the Serp_SSB and was eroded at the time
of the Serp_SSB following minor tilting to the north.
RESERVOIR QUALITY AND DIAGENESIS
Pore Types
Air permeability data from approximately 11,000
plugs clearly suggest the presence of three general
reservoir environments based on plug-scale matrix
properties: central platform, outer platform, and rim
slope (flank) (Collins et al., 2006). The Visean A, Ser-
pukhovian, and Bashkirian central platform reservoir
can be regarded as having distinctly different gross
reservoir properties that are generally well behaved as
a whole. Porosity-permeability crossplots and poros-
ity histograms suggest different behavior between
the northern wells (T-220 and T-5246) and southern
wells (T-6246 and T-6846) in the central platform
(Figure 16). Trends of generally lower mean porosi-
ties and better behaved porosity-permeability relation-
ships occur in the southern wells, whereas bimodal
porosity distributions occur in the northern wells.
Pore-type characterization of nearly 2000 plugs (and
associated thin sections) from the nearly bitumen-
free T-220 well and the moderately bitumen-cemented
T-6246 well show the presence of varied pore types
like moldic, vugs, microporosity, interparticle poros-
ity, intercrystalline, and intraparticle porosity and
the general absence of fractures (Figure 17).
The porosity classification scheme of Choquette
and Pray (1970) defined numerous carbonate pore
types, either fabric selective (including interparticle,
intraparticle, intercrystalline, and moldic) or not fab-
ric selective (including fracture and vug). There were
implications for porosity-permeability relationships,
of course, but this was not emphasized. Lucia’s (1995,
1999) landmark study showing the direct link between
rock type, porosity type, and reservoir quality (porosity-
permeability relationships or transforms) changed the
context of some of these classic pore types. Lucia (1995,
1999) recognized interparticle pore types and vuggy
pore types. Interparticle pores, which have a more
straightforward porosity-permeability transform that
can be tied to rock type, grain size, and sorting, are
instances where the pore space occurs between par-
ticles and is not significantly larger that the particles.
FIGURE 16. Subtle porosity-permeability differences are observed between four wells in the central platform; see Figure 4 for well locations and Figure 5 for
their facies and stratigraphy. K-Phi plots (top) and porosity histograms (bottom) suggest different behavior between the northern wells (T-220 and T-5246) and
southern wells (T-6246 and T-6846) with generally lower mean porosities and better behaved permeability-porosity relationships in the southern wells and
bimodal porosity distributions in the northern wells.
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 35
36 / Kenter et al.

FIGURE 17. Pore and cement types recognized in the T-220 and T-6246 wells are generally similar. However, relative
contributions by vuggy, intercrystalline, and moldic pore types are significantly higher in T-220 for all facies types
than in T-6246, which may be related to the larger average grain size and higher contribution by the mud-lean rock
types in T-6246. Similarly, cement types in T-6246 show a higher abundance (frequency) of equant calcite and syntaxial
cements than in T-220. This may suggest a larger volume of deep burial cements filling pores at T-6246, but clearly, more
quantitative data are required for such assessment. Pore types across the bottom of the figures are 1 = interparticle;
2 = intraparticle; 3 = fenestral; 4 = shelter; 5 = growth framework; 6 = intercrystalline; 7 = moldic; 8 = microporosity; 9 = fracture;
10 = channel; 11 = vug; and 12 = enhanced dissolution. Cement types are 1 = fibrous; 2 = bladed; 3 = equant to rhombic;
4 = coarse crystalline; 5 = botryoidal; 6 = syntaxial; 7 = micritic; 8 = meniscus; 9 = microstalactitic; 10 = poikilotopic; 11 =
neomorphic spar; and 12 = grain rims. Plotted are relative dominance of pore or cement types in terms of visual presence
with the number indicating their ranking. See text for discussion.

Vuggy pore types are typically larger than the grains or
occur within the grains, and these can be separate or
touching. Separate vugs are connected through the in-
terparticle pore system, whereas touching vugs them-
selves form an interconnected system.
At Tengiz, separate vugs that add to porosity but
not permeability include moldic and intraparticle
porosity, so these must be subtracted from total po-
rosity to use interparticle transforms. Touching pore
types include some vugs (solution-enlarged interpar-
ticle pores) and fractures, which add porosity and
permeability independently from the interparticle
component, such that interparticle transforms must
be modified. Clearly, real-world reservoirs in lime-
stone can be even more complex with the addition of
microfractures, intercrystalline porosity, or micropo-
rosity connecting separate vugs, and/or the problem
of assessing in two dimensions with petrography the
three-dimensional complexity of porosity. The fol-
lowing sections describe the pore types commonly
observed in Tengiz samples and their relative abun-
dance in the central platform.
Pore types were identified in the Tengiz cores and
thin sections generally following the terminology
of Choquette and Pray (1970) to first assess their
relation to lithofacies and then explore the relation
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 37
between lithofacies and reservoir quality (Table 2).
Common pore types in the Tengiz central platform are
vugs, interparticle, microporosity, intercrystalline,
moldic, and intraparticle with generally the highest
permeability occurring in vuggy-dominated matrix.
The most common pore type is vuggy, which is, in part,
fabric selective, being found in matrix, grainy areas
(solution-enhanced interparticle pores), or associated
with stylolites (Figure 18A–C). The second most common
pore type is interparticle porosity (Figure 18D–H),
followed by microporosity (Figure 19A, B).
Vuggy porosity includes moldic porosity when
considered in terms of porosity-permeability trans-
FIGURE 18. Photomicrographs illus-
trating the dominant pore types in
the central platform area. (A) Vugs
are nonfabric-selective pores com-
monly larger than the dominant grain
size (although different interpreta-
tions exist in the literature). Examples
of vugs shown here are the result of
enhanced dissolution in a grainstone
following early equant rim cementa-
tion. (B) Vugs developed in a pack-
stone-grainstone, dissolving both
grains and matrix but leaving crinoid
ossicles and brachiopod fragments
intact. (C) Overpacked ooid grain-
stone with vugs that postdate early
cementation but predate blocky cal-
cite and bitumen emplacement. Vugs
are a dominant pore type in the
central platform and seem to be
mostly related to deep-burial disso-
lution instead of initial meteoric pro-
cesses. (D) Advanced stage of en-
hanced dissolution of moldic and
interparticle porosity (and minor in-
traparticle) leading to the destruction
of cement bridges and development
of vugs. (E) Skeletal and coated grain
grainstone dominated by interparti-
cle porosity with minor cement fill.
(F) Overpacked ooid grainstone with
minor blocky calcite and bitumen
occluding solution-enhanced inter-
particle porosity. (G) Skeletal and
peloid grainstone with rims of equant
spar lining interparticle pores and
minor moldic porosity. (H) An exam-
ple where the interparticle porosity is
completely occluded by equant to
rhombic spar and bitumen cement.
Width of photomicrographs is 4.20
and 1.58 mm (0.16 and 0.06 mm)
(A – G, H, respectively). See text for
discussion.
forms (Lucia, 1995, 1999). Figure 19C – E shows suc-
cessive stages of the development of moldic porosity
at Tengiz: in its early phase where micrite grains are
gradually being dissolved following partial fill of in-
terparticle porosity by equant to rhombic spar ce-
ment (Figure 19C); a phase of progressive dissolution
shown by corrosion of the mold boundaries, nearly
breaking through at contacts and corroding earlier
cement (Figure 19D); and an advanced stage where
enhanced dissolution left only micrite envelopes. In-
traparticle porosity, which is also included as vugs in
terms of porosity-permeability transforms, is shown
in Figure 19D–F.
38 / Kenter et al.
Examples of an advanced stage of dissolution of
moldic and interparticle porosity leading to the de-
struction of cement bridges and development of vugs
with minor blocky calcite and bitumen occluding
porosity are shown in Figure 18D and F. Commonly,
moldic porosity co-occurs with interparticle poros-
ity as shown in Figure 18G; interparticle porosity is
completely occluded by equant to rhombic spar and
bitumen in some samples (Figure 18H). Microporos-
ity in grains or matrix is difficult to observe in Tengiz
samples because pore sizes are much smaller than the
thickness of a typical thin section. Microporosity in-
cludes fine intercrystalline porosity because both are
FIGURE 19. Photomicrographs illus-
trating additional dominant pore types
in the central platform area. (A, B)
Examples of microporosity, pore size
smaller than 25 Mm, developed in
grains (A) and matrix (B). This type of
pore is generally difficult to observe
in thin sections because of their thick-
ness, commonly about 25 – 45 Mm.
(C) Moldic porosity in its early phase:
micrite grains are gradually being
dissolved following partial fill of inter-
particle porosity by equant to rhombic
spar cement. (D) Moldic porosity
showing evidence for enhanced disso-
lution shown by corrosion of the mold
boundaries, nearly breaking through
at contacts and corroding earlier ce-
ment. (E) Moldic porosity showing
further stage of enhanced dissolution
leaving only micrite envelopes. Bitumen-
stained micrite appears to slow down
dissolution. Interparticle porosity is
partially filled with calcite cement.
(F) Intraparticle porosity developed
in benthic foraminifera. (G) Intraparticle
porosity developed in a green alga,
Koninckopora. (H) Intraparticle porosity
developed in a Chaetetes fragment (not
all pores are filled by blue epoxy). Width
of photomicrographs is 4.20 mm (0.16 in.)
(A, B, F, G) and 1.58 mm (0.06 in.)
(C, D, E, H). See text for discussion.
generally beyond normal petro-
graphic resolution and, therefore, dif-
ficult to distinguish. Fractures and
fissures are generally uncommon in
the central platform, and where pre-
sent, they are mostly occluded by
cement. In some few cases where the
fractures remain open, they range
from microhairline fissures to sharp
microfractures (Figure 20A–F). Stylolites are generally
stained by bitumen and commonly associated with
small vugs (Figure 21A – C). Examples demonstrate
that such dissolution postdates stylolite formation.
Parameters Controlling Porosity
and Permeability
The relative contributions by vuggy, moldic, and
intercrystalline porosity are significantly higher for
all facies types in well T-220 than in well T-6246. In
contrast, the relative contribution by interparticle po-
rosity is higher at T-6246, a change that is most likely
related to the higher percentage of grainy, mud-lean
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 39
rock types in that well (Figure 22). Nevertheless, at a
gross scale, none of the individual pore types or com-
bination of pore types seem to have a significant first-
order relationship with primary depositional facies
nor have any significant control on porosity or per-
meability. This suggests that the central platform area
could be regarded as one single reservoir region wherein
changing pore types and lithofacies do not significantly
impact true reservoir connectivity. This suggestion is
rather surprising and seemingly contradicts Lucia’s (1995,
1999) reservoir rock-type classification that suggests
pore types control permeability behavior.
Although the observations above suggest minimal
first-order control by pore type or facies, several second-
order relationships should be acknowledged. Porosity
distributions and pore-type contributions for the
muddy lagoonal facies (peloid packstone-wackestone),
moderately sorted grainy facies with mud matrix
(brachiopod skeletal-peloid grainstone-packstone,
crinoid skeletal-peloid grainstone-packstone, and
poorly sorted skeletal-peloid grainstone-packstone)
and mud-lean well-sorted grainy facies (well-sorted
skeletal grainstone, coated-grain grainstone-rudstone,
FIGURE 20. Photomicrographs illus-
trating dominant pore types in the
central platform area. (A, B) Generat-
ing thin sections from core plugs
with open fractures is difficult if not
impossible. Shown here are examples
of thin fractures that are filled by
blocky calcite cement. (C, D) Examples
of hairline fissures or fractures filled
by bitumen. Note that the fissure in
(C) postdates blocky calcite interparti-
cle cement fill but predates bitumen
emplacement and confirms the very
late timing of bitumen invasion.
(E, F) An exception to the rule: vague
outline of open fractures that have
been enhanced by late-stage dissolu-
tion in a skeletal-peloidal grainstone
with high interparticle porosity. Frac-
tures may connect vugs as shown
in these samples. Width of photomi-
crographs is 4.20 mm (0.16 in.). See
text for discussion.
ooid grainstone, skeletal intra-
clast grainstone, and algal grain-
stone) that make up the cyclic
pattern shown in Figure 5 do show
subtle but significant differences
(Figure 22). Pore type contribu-
tions in the grainy facies groups
are very similar, but muddy la-
goonal facies have a higher contribution of intra-
particle and fracture porosity. The grainy facies in
well T-6246 have a higher contribution of interpar-
ticle porosity but also differ in the overall distribu-
tion of porosity. These distributions are remarkably
different between the two wells with bimodal and
generally higher porosities in grainy facies and low
porosities in lagoonal facies in well T-220, whereas
the porosity distributions in well T-6246 show highly
reduced and flattened distributions for the mud-lean
facies compared to T-220 (Figure 23). Similarly, the
mud-rich grainy facies have reduced porosity, but the
lagoonal mudstone has higher porosity, displaying
a bimodal distribution that is possibly related to a
higher contribution of fracture porosity. The above
suggests that T-220 porosity has a relatively higher
contribution by pore types that are the result of cor-
rosive processes, like vug and moldic porosity, an ob-
servation that can be made for T-5246 as well.
In addition to pore type, observations were made
on the type of pore-filling agents, which resulted in
two significant trends. First, the distribution of pore-
filling cement types for both wells T-220 and T-6246
40 / Kenter et al.
appears to be comparable, with the exception of a rel-
atively higher contribution by equant and syntaxial
calcite at T-6246 (Figure 17). Another obvious difference
between the T-220 and T-6246 wells is the amount of
bitumen; Figure 5 shows qualitatively higher amounts
of bitumen in T-6246 (and T-6846) than in T-220. For
well T-220, bitumen volume is mostly associated with
muddy grain-rich to mud-rich facies (Figure 24A) and
dominant pore types like interparticle, intercrystalline,
and vug with minor contributions by intraparticle,
moldic, and fracture (Figure 24B). Bitumen content
shows a negative relationship with porosity (Figure 24C)
and permeability (Figure 24D).
FIGURE 21. Photomicrographs illus-
trating the dominant pore types in the
central platform area. (A) Pressure
solution stylolite with dark-brown
micrite and possibly bitumen. (B) As
in (A), but dissolution associated with
stylolite. (C) As in (A, B), but here,
dissolution postdating stylolite forma-
tion is clearly visible; note that the vug
is not breaking through the bitumen
seam. (D) Poikilotopic syntaxial over-
growth cements on crinoid ossicles are
locally dominant and reduce interpar-
ticle porosity. However, invasion by
bitumen and associated dissolution was
able to corrode cement and crinoid
ossicles. (E) Coated grainstone where
equant to rhombic spar is occluding in-
terparticle pore space, probably re-
placing earlier marine cements, and
only sparse molds and remaining
interparticle pores remain. (F) Micri-
tized coated grainstone with evidence
for minor corrosion of grain boundaries
followed by burial blocky calcite that
occludes part of the interparticle po-
rosity. Micrite is locally invaded by bi-
tumen. (G) Detail of partially open
fracture that was occluded by blocky
calcite cement and followed by bitumen.
(H) Porosity in poorly sorted grain-
stone was occluded by cement and
completely recrystallized to fine cal-
cite spar cement. Width of photomi-
crographs is 4.20 mm (0.16 in.)
(A – C, E – H) and 1.58 mm (0.06 in.)
(D). See text for discussion.
Figure 25 better shows the influ-
ence of bitumen volume on the po-
rosity versus permeability relation-
ship in well T-220 (no such data are
yet available for T-6246). Increasing
bitumen content above a threshold
of approximately 0.5% rock volume
is negatively correlated with K-Phi. Although bitumen
commonly occurs in grainy facies with matrix pore types,
it also occurs with late diagenetic pore types like vugs
as well as with nonmatrix pore types like fractures.
Bitumen content significantly impacts the cyclic and
stratigraphic porosity variations as shown by Figure 26;
it also causes the observed decrease in porosity toward
the outer platform. Porosity logs corrected for bitumen
volume are easier to correlate and confirm that the dis-
tribution of bitumen increases toward the outer plat-
form (Figure 26). Clearly, bitumen reduced the porosity
toward the outer part of the central platform and even
more so into the outer platform itself (Figures 4, 5).
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 41

FIGURE 22. Pore-type contributions for muddy lagoonal facies (13 = PPW) and moderately sorted grainy facies with mud
matrix (2 = BSP, 31 = CSP, 32 = SPP) and mud-lean, well-sorted grainy facies (41 = SPG, 42 = CgGR, 43 = OG, 44 =
SIG, 46 = AG) in both T-220 and T-6246 wells. Pore types are 1 = interparticle; 2 = intraparticle; 3 = fenestral; 4 = shelter;
5 = growth framework; 6 = intercrystalline; 7 = moldic; 8 = microporosity; 9 = fracture; 10 = channel; 11 = vug; and
12 = enhanced dissolution. Plotted are relative dominance of pore types in terms of visual presence with the number
indicating whether pore types were first, second, or third in importance. See text for discussion.

Volcanic ash also produces low-porosity (tight) zones.
These tight zones tend to occur around cycle bound-
aries and within intervals that have dispersed volcanic
ash. One mechanism explaining these observations
associates volcanic ash with inhibited downward fluid
percolation during early diagenesis associated with
sequence and cycle boundaries. This condition gener-
ally reduced the effect of meteoric dissolution. In ad-
dition, silica leached from the volcanic ash was, in
some cases, redeposited as chert and silica cement be-
low the sequence boundaries, thereby adding to the
porosity reduction of the associated lagoonal inter-
vals. Finally, the dissolution associated with the bitu-
men emplacement did not affect those tight zones.
Processes and Timing of Diagenesis
The present-day distribution of reservoir quality in
the central and (southwestern) outer platform in the
Lvis_SSB to Bash_SSB platform succession (Visean A,
Serpukhovian, and Bashkirian reservoir zones) was
determined by the combined effect of late diagenetic
modification, specifically bitumen invasion and as-
sociated corrosion, of an earlier reservoir system whose
porosity and permeability distribution was established
by highly cyclic depositional and early diagenetic
processes (Figure 27). Early diagenesis includes me-
teoric alteration in cyclic platform facies associated
with major sequence boundaries and reduced disso-
lution along sequence and high-order cycle bound-
aries associated with the presence of volcanic ash.
Diagenetic processes associated with subaerial expo-
sure produced both an increase and reduction of po-
rosity (Figure 27). Brown fibrous, mycrostalactitic,
meniscus, and pendant cements are commonly as-
sociated with meteoric vadose and marine vadose
diagenesis, and these are observed in Tengiz cores to
several meters below a suspected exposure surface
(Figure 28A – D). Possible examples of alveolar fabrics
indicating soil formation below an exposure surface
are shown in Figure 28E. Early dissolution related to
 42 / Kenter et al.

FIGURE 23. Porosity distributions for the major facies types in T-220 and T-6246: muddy lagoonal facies (13 = PPW), moderately sorted grainy facies with mud
matrix (2 = BSP, 31 = CSP, 32 = SPP), and mud-lean well-sorted grainy facies (41 = SPG, 42 = CgGR, 43 = OG, 44 = SIG, 46 = AG). See text for discussion.
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 43

FIGURE 24. Relationship between bitumen (Bit rock), rock fabric (PP fabric), pore type (pore 1), porosity (Phi), and
permeability (K ) in well T-220 (no such data are yet available for T-6246). Bitumen content is mostly associated with
grainy mud-lean to mud-rich facies (A) and dominantly in interparticle, intercrystalline, and vug pore types with minor
contributions by intraparticle, moldic, and fracture pores (B). Bitumen contents above 0.5% rock volume shows a
negative relationship with porosity (C) and permeability (D). Pore types are 1 = interparticle; 2 = intraparticle;
3 = fenestral; 4 = shelter; 5 = growth framework; 6 = intercrystalline; 7 = moldic; 8 = microporosity; 9 = fracture; 10 =
channel; 11 = vug; and 12 = enhanced dissolution. Textures are 1 = rudstone; 2 = grainstone-rudstone; 3 = grainstone;
4 = grainstone-packstone; 5 = packstone-grainstone; 6 = packstone; 7 = packstone-wackestone; 7 = wackestone-packstone; 8 =
wackestone; 9 = wackestone-mudstone; 10 = mudstone-wackestone; 11 = mudstone; and 12 = boundstone. See text for
discussion.

exposure may have led to early compaction as shown
by the frequent occurrence of fitted fabrics (Figure 28H),
where the thin cement rim is included in the com-
paction and postdated by blocky calcite cement and
bitumen.
Later stage diagenesis includes pressure solution
culminating in stylolites with concentrations of dark-
brown micrite and bitumen (Figure 21A, B). Dissolution
is sometimes observed postdating the stylolite forma-
tion (Figure 21C). Poikilotopic syntaxial overgrowth
cements on crinoid ossicles are locally a dominant
cement, reducing interparticle porosity (Figure 21D).
Later burial processes also reduced porosity by cemen-
tation and recrystallization (Figure 21E – H). Exam-
ples show progressive occlusion of pore space during
this process, which includes cementation of pore space
created by fracturing (Figure 21G).
Burial dissolution events that added porosity are
of great interest and paramount importance to the
Tengiz reservoir and its interpretation with respect
to porosity modeling (Figure 29A–F). A very common
observation in samples that have interparticle poros-
ity is enhanced (late-burial?) dissolution that enlarges
and connects interparticle pore space into larger vugs.
These vugs are lined by bitumen, attesting to the timing
of the dissolution relative to bitumen emplacement
(Figure 29A). Grain boundaries are commonly corrod-
ed by this dissolution event, as is syntaxial overgrowth
cement in some cases (Figure 29B, C). Extensive late
deep burial dissolution prior to oil migration has been
44 / Kenter et al.
reported as a major reservoir porosity generation mech-
anism in, at least, several other carbonate fields (Moore,
2001; Esteban and Taberner, 2003; Zampetti et al.,
2003; Sattler et al., 2004). Most of these authors sug-
gest corrosion by mixing of formation fluids with an
external fluid at higher temperatures, and such pro-
cess may be responsible for the burial corrosion in the
Tengiz central platform as well. Similarly, the cor-
rosion event postdating the bitumen invasion (and
cementation) may represent yet another such leach-
ing phase with different fluids. Fluid-inclusion and
stable isotope studies will shed light on the origin of
the substantial burial corrosion at Tengiz; see Collins
et al. (2006) for discussion of preliminary geochem-
ical data and a probable dissolution model.
Recrystallization is also an important process in the
central platform diagenesis. As an example, Figure 29D
is a grainstone that was recrystallized to equant and
blocky spar. This recrystallization followed early par-
tial cementation of interparticle porosity and was,
in turn, followed by late dissolution that produced
moldic porosity. Note that early moldic porosity was
FIGURE 25. Influence of bitu-
men volume on the porosity
(PHI) versus permeability (K )
relationship in well T-220 (no
such data are yet available for
T-6246). Increasing bitumen
content is negatively correlat-
ed with K-Phi above a certain
threshold of about 0.5% rock
volume. Although bitumen is
common in grainy facies with
matrix pore types, it also oc-
curs with late secondary pore
types like vugs indicative of
dissolution as well as with non-
matrix pore types like fracture
porosity. See text for discussion.
occluded by blocky calcite
(see micrite envelope), where-
as this later dissolution phase
generated the open molds that
are lined by bitumen. More
common are finely recrystal-
lized fabrics associated with
intercrystalline and minor
moldic porosity, as well as
the presence of scattered bi-
tumen (Figure 29E, F). Al-
though the recrystallization
phase seems to postdate (cor-
rode and truncate) blocky calcite cementation, the
exact timing remains unclear because of the destruc-
tive nature of the process.
The important diagenetic overprint in the Tengiz
central platform as outlined in Figure 27 can be sum-
marized as follows. Syn- and postdepositional disso-
lution in meteoric environments generated moldic,
enlarged interparticle and intraparticle, and micro-
porosity pore types, as well as vuggy microkarst. Ce-
mentation of lagoonal intervals associated with the
presence of discrete and dispersed volcanic ash signifi-
cantly reduced porosity and created permeability baffles
around cycle and sequence boundaries. Intermediate
shallow burial produced commonly observed fitted
fabrics and low- to medium-amplitude stylolites. The
lateral continuity and persistence of tight layers around
sequence and cycle boundaries as well as in local high-
energy intervals probably greatly affected later fluid
flow and, thus, the ultimate distribution of cements,
dissolution, and bitumen in the central platform res-
ervoir. The late diagenetic overprint was responsible
for the enhanced dissolution, increasing any existing
FIGURE 26. Subtle but significant differences in petrophysical and geological properties are observed between these four wells in the central platform (see Figure 4
for well locations and Figure 5 for additional correlations). K-Phi plots and porosity histograms suggest different behavior between northern central platform wells (T-220
and T-5246) and southern central platform wells (T-6246 and T-6846), with generally lower mean porosities and better behaved permeability-porosity relationships
in the southern wells and bimodal porosity distributions in the northern wells. Bash = Bashkirian; Serp = Serpukhovian; VisA = Visean A. See text for further discussion.
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 45
46 / Kenter et al.

FIGURE 27. Paragenetic sequence for the central platform. Significant syndepositional, intermediate (shallow)-burial,
and deep (late)-burial events are listed in chronological order. In the lower part of the figure, those events that are
particularly important to porosity creation and destruction are listed. MO = moldic; BP = interparticle; IP = intraparticle;
MP = microporosity; MVUG = micro vug. See text for discussion.

porosity in the matrix and forming vugs and minor
fractures, whereas equant calcite cement reduced po-
rosity away from the interior central platform. The sub-
sequent bitumen invasion exerted an overall porosity-
reducing effect increasing toward the outer platform.
These late-burial events that increased and reduced
porosity significantly dampened or flattened initial
vertical, nearly cyclic, porosity variations linked to
primary depositional properties.
Spatial Distribution of Central
Platform Reservoir Quality
In the absence of any significant first-order con-
trol by pore type or lithofacies, the following sec-
tions describe the reservoir quality in the context of
the transition from central to outer platform envi-
ronments. This lateral change is likely a result of the
early cyclic depositional and diagenetic processes, in-
cluding the preservation of volcanic ash, coupled with
late-burial dissolution and bitumen emplacement.
Figure 30 presents a model that summarizes po-
rosity generation and reduction events in the cen-
tral Tengiz platform from the standpoint of gamma-
ray and porosity logs. Porosity changes are linked to
the major diagenetic phases following deposition of
alternating thin, low-energy lagoonal intervals with
volcanic ash beds and thick intervals of high-energy
grainstone. Detailed log and core correlation shows
that lagoonal interval and ash beds are not always
associated and/or continuous between well locations,
and cycles may vary slightly in thickness over kilo-
meters distance. Three important diagenetic stages
that affect the reservoir quality and distribution are
observed. First, porosity curves with a blocky shape
are the result of early-burial processes, including
low porosity around cycle boundaries related to the
presence of thin ash beds and corrosion and cemen-
tation related to meteoric diagenesis (Figure 30A).
Locally thick grainstone intervals have low porosity
and show the presence of elevated levels of dispersed
ash concentrated in frequent stylolites. Second, a
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 47
burial event led to corrosion and enhanced blocky
porosity curves in the interior central platform and
porosity reduction in the exterior central platform;
alternatively, corrosion in the exterior central plat-
form was lower compared to the interior central plat-
form (Figure 30B). Finally, bitumen that was emplaced
during a late-burial event acts as cement and reduces
porosity (Figure 30C). Bitumen volume increases from
the interior to the exterior central platform and margin
and obscures correlation using the porosity curves.
Some evidence suggests that another corrosion phase
is associated with the bitumen emplacement; in the
FIGURE 28. Photomicrographs illustrat-
ing some of the dominant cement
types in the central platform area. (A,
top to the right) Brown fibrous pendant
cement attached to lower part of bra-
chiopod fragments is interpreted as
the result of meteoric diagenesis and,
therefore, suggests subaerial exposure
above. (B) Mycrostalactitic or pendant
cements, which are commonly associ-
ated with vadose diagenesis, are ob-
served to several meters below a suspect
exposure surface. (C, D) Meniscus and
pendant cements, indicating vadose
to marine vadose diagenesis occlude
much of the primary porosity in these
samples. Burial dissolution and/or cor-
rosion of preserved interparticle porosity
are followed by blocky calcite cement
and bitumen. (E) Possible example of
alveolar fabrics indicating soil formation
below exposure surface. (F) Possible
vadose cementation followed by en-
hanced corrosion or dissolution of grain
boundaries and cement, blocky calcite
cement, and bitumen. (G, H) A combi-
nation of minor cementation and com-
paction leads to fitted fabrics or over-
packed fabrics. Here, a thin cement
rim is included in the compaction pro-
cess and followed by burial blocky
calcite spar cement and bitumen. Width
of photomicrographs is 4.20 mm
(0.16 in.) (A, C, D, E, F) and 1.58 mm
(0.06 in.) (B, G, H). See text for
discussion.
rim and flank area, observations indi-
cate a corrosion event following the
bitumen cement (Collins et. al.,
2006).
Packstone to mudstone intervals
around the cycle boundaries in the
central platform wells typically dis-
play significantly lower porosity and permeability rela-
tive to the overlying grainstone-packstone lithofacies
(Figure 31, sequences between Lvis10.5_csb-Lvis9.5_csb
in wells T-5246 and T-220). The latter have porosity
ranges between 0 and 24% with a mean of 10.5 and
11.0% (Table 2). The packstone to mudstone inter-
vals are generally associated with gamma-ray spikes
caused by the presence of volcanic ash and are gen-
erally tight with a mean porosity of 3.5%. Within the
central platform, the cyclic signature of alternating
blocky high-porosity and low-porosity zones associ-
ated with gamma-ray spikes is dominant. Lateral and
48 / Kenter et al.

FIGURE 29. Photomicrographs illustrating additional pore types of the central platform area. (A) Enhanced late-burial(?)
dissolution joining interparticle pores into larger vugs prior to bitumen lining the pore walls. The primary grain
boundaries appear to have been corroded by the dissolution. In addition, following bitumen emplacement, grains were
partially dissolved by an additional dissolution event. (B) Bitumen corroding both the crinoid ossicle boundaries as well as
those of the syntaxial overgrowth cement. (C) As in (B), but advanced stage of bitumen emplacement corroding grain
boundaries and leading to pseudopackstone to wackestone fabric. (D) Recrystallization is a dominant process in the
central platform diagenesis. Here, an example where the grainstone fabric was recrystallized to equant and blocky spar
following early partial cementation of interparticle porosity and followed by late dissolution generating the moldic
porosity. Note that early moldic porosity was occluded by blocky calcite (see micrite envelop), whereas a later disso-
lution phase generated the open molds that are lined by bitumen. (E, F) Finely recrystallized fabrics associated with
intercrystalline and minor moldic porosity, as well as the presence of scattered bitumen, are very common in the
central platform. Although the recrystallization phase seems to postdate (corrode and truncate) blocky calcite cemen-
tation, the exact timing remains unclear because of the destructive nature of the process. Width of photomicrographs
is 4.20 mm (0.16 in.) (C, F) and 1.58 mm (0.06 in.) (A, B, D, E). See text for discussion.

vertical changes in distribution and type of the in-
termediate grainstone to packstone lithofacies have
only a minor effect on the present-day porosity. The
interval on Figure 31 in well T-5246 represents high-
energy grainstone, and although the interval in well
T-220 is lower energy packstone-grainstone facies,
no visible difference is observed on the porosity
log.
A remarkable deviation in the central platform
occurs in the lower part of the Bash1_csb, where
mostly well-sorted ooid grainstone is invaded and
cemented by bitumen all across the platform. The re-
sulting porosity range is 0–10%, with a mean of 4%,
whereas in bitumen-poor ooid grainstone, the poros-
ity ranges between 0 and 21%, with a mean of 12%
(Table 2). The fact that bitumen (as much as 3.5%
 Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 49

FIGURE 30. Hypothetical log correlation scenarios in the central platform, with gamma ray shown on the left and porosity
on the right for each well(s). Porosity generation and reduction events that are linked to the major diagenetic phases
are indicated for a stratigraphy that includes thin, low-energy lagoonal intervals alternating with volcanic ash beds (yellow
and green horizons) and thick intervals of high-energy grainstone (white). Detailed correlation shows that lagoonal
intervals and ash beds are not always associated and/or continuous between well locations, and cycles (triangles) may
vary slightly in thickness over kilometers distance. (A) Porosity curves with a blocky shape are the result of early-burial
processes, including porosity loss around cycle boundaries related to the presence of thin ash beds and corrosion and
cementation related to meteoric diagenesis (porosity is blue). Locally thick grainstone intervals have low porosity and
show the presence of elevated levels of dispersed ash concentrated in frequent stylolites. (B) Late-burial events leading
to corrosion and enhanced blocky porosity curves in interior central platform (red) and porosity reduction in the
exterior central platform (yellow); alternatively, corrosion in the exterior central platform was lower compared to the
interior central platform. (C) Bitumen, which is emplaced during late burial, acts as cement and reduces porosity.
Bitumen volume increases from the interior to the exterior central platform and margin and obscures correlation using
the porosity curves.

rock volume) invaded this far into the central plat-
form may be related to the continuous underlying
pressure baffle, at the Serp_SSB, which preexisted
prior to the bitumen and associated dissolution phase
(Figure 32A).
Bitumen occurs within fine matrix and as a pore-
filling phase, and two spatial trends are observed (see
Figure 5, 25): (1) a general increase from exterior to
interior central platform wells and (2) elevated bitu-
men levels near the base of grainy facies between
50 / Kenter et al.

FIGURE 31. Cross section showing lithofacies, gamma-ray (left), and porosity (right) logs; well locations are shown in Figure 4, and lithofacies legend is shown in

and T-6846 wells. The cyclic and blocky high-low porosity character is visible in nearly bitumen-free wells (T-5246 and T-220) and represents, respectively, alternating
tight lagoonal intervals. In addition, some ooid grain-

grainy facies and lagoonal mud intervals with volcanic ash. Although the interval in T-5246 represents high-energy and essentially mud-free grainstone, and that
Figure 5. The lateral facies and thickness continuity of the sequences between Lvis10.5_csb and Lvis9.5_csb are shown between the T-5246, T-220, T-6246,
stone (right above the Serp_SSB) and crinoid grain-
stone-packstone intervals (mostly between the Lvis_SSB
and Serp_SSB) show increased bitumen content all
across the central platform. The late timing of bitu-
men relative to other diagenetic overprint suggests
the introduction through the flanks and into the
platform part of the Tengiz buildup, with the largest
lateral extent in the most permeable units and resid-

in T-220 is lower energy and packstone-grainstone facies, no visible difference is observed on the porosity log. See text for discussion.
ual deposition controlled by gravity, that is, on top of
the immediate underlying impermeable cycle bound-
ary (see also Figure 26). Additional intervals with
elevated bitumen extend to T-220 but disappear up-
section and in the direction of T-5246 (Figures 5, 32B).
In the north (well T-5044) and southwest (wells
T-6246 and T-6846) parts of the central platform,
overall bitumen content in the Lvis_SSB to Bash_SSB
grainy facies reduces porosities to a mean of 5.8–
7.6%, which is significantly reduced from the 10.6 –
11.2% recorded in well T-220 (Figure 5). In general,
increased bitumen content is commonly confined to
the lower part of a cycle. The bitumen-cemented
interval exhibits a sharp, lower contact with under-
lying, well-cemented peloid packstone-wackestone
lithofacies, has a gradational upper contact, and ap-
pears dark gray to black in core. Although the blocky
log character changes to a more irregular and spiky
shape with increasing bitumen content, no obvious
relationship exists between the amount of porosity
reduction and primary lithofacies type. Instead, the
presence of the tight peloid packstone-wackestone
baffles seems to control the occurrences of bitumen-
cemented interval (see Figures 31, and 33, well T-220
versus well T-6246, no bitumen but lateral facies
change). The bitumen effect progressively increases
away from the platform center and, at the scale of
the plug sample, reduces the porosity and perme-
ability as suggested by the relationship shown in
Figure 24.
In addition to porosity and permeability reduction
by bitumen and volcanic ash layers, the presence of
dispersed volcanic ash in high-energy grainstone fa-
cies locally also has an effect (Figure 34). Volcanic ash
dispersed in the well-sorted skeletal grainstone and
coated-grain grainstone-rudstone lithofacies of the
Serp1_csb sequence at well T-5447 reduces porosity
to nearly zero and generates a nearly flat porosity
curve, whereas the same facies interval at well T-220
retains high porosity. Similarly, the tight and later-
ally continuous interval associated with the Serp_SSB
is probably also associated with elevated levels of dis-
persed volcanic ash (see Figure 5).
FIGURE 32. Cross sections showing lithofacies, gamma-ray (left), and porosity (right) logs; well locations are shown in Figure 4, and lithofacies legend is shown
in Figure 5. The lateral facies and thickness continuity of sequences between the Serp_SSB and Lvis13_csb are shown for the T-220 and T-6246 wells in (A) and
between the Bash2.5_MFS to Bash4_csb for the T-5246, T-5447, T-220, and T-6246 wells in (B). The lower part of the Bash1_csb, mostly well-sorted ooid grainstone
(OG), is invaded and cemented by bitumen all across the platform. Additional intervals with elevated bitumen extend to T-220 but disappear upsection and toward
T-5246. See text for discussion.
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 51
52 / Kenter et al.

section increases (as much as 2 – 4% rock volume bitumen in some parts of T-6246), and generally, elevated levels of bitumen are commonly confined to the lower
FIGURE 33. Cross section showing lithofacies, gamma-ray (left), and porosity (right) logs; well locations are shown in Figure 4, and lithofacies legend is shown in

southwestern (interpreted as outer platform) margin and away from the central platform to the north (T-5044), overall bitumen in the Lvis_SSB to Bash_SSB
IMPLICATIONS FOR RESERVOIR MODELING

Figure 5. The lateral facies and thickness continuity of sequence Lvis9.5_csb to Lvis9_csb are shown for wells T-5246, T-5447, T-220, and T-6246. Toward the
The stratigraphic framework introduced for the
central platform part of the Tengiz buildup in the
preceding sections is the basis for the layering scheme
that is being used in reservoir models for the field. Full-
field models use the layering portrayed on Figure 4,
whereas more detailed models focusing on parts of
the platform incorporate more detailed layers as
shown in Figures 5 and 25. The lithofacies summa-
rized in Table 2 have been grouped and used to
subdivide layers in the reservoir models into regions
where porosity and permeability might be expected to
vary. The complex diagenesis that occurs in the central
platform and the outer platform has overprinted the
original (depositional) porosity to an extent, as sum-
marized in Figure 30, that it cannot be ignored when
trying to interpret reservoir quality from wire-line logs
and correlate it within a reservoir model. Ongoing studies
of the detailed analysis of the types and amounts of
porosity, as well as pore-filling agents are being used to
better understand the log signatures of porosity and
improve porosity-permeability transforms. Much work
remains in the detailed characterization of the Tengiz
reservoir and in correctly portraying the significant
geologic characteristics of the Tengiz platform in
reservoir models.

SUMMARY AND CONCLUSIONS

An integrated (and ongoing) study of core, well

logs, and discrete petrophysical measurements provid-
ed a first comprehensive assessment of depositional
evolution and reservoir quality in the late Visean to
Bashkirian interval in the Tengiz central platform.
The study resulted in the following observations and
tentative conclusions.
Depositional cycles (=higher frequency sequences)
are up several to tens of meters in thickness and made
part of the cycle. See text for discussion.

up of a succession of lithofacies overlying a sharp

base with variable evidence for subaerial exposure. At
the base of the cycles, tight peloidal mudstone and
volcanic ash beds are associated with sequence bound-
aries and represent flooding events. They are overlain
by brachiopod- and crinoid-dominated intervals that
represent maximum marine flooding and near the top
by skeletal-peloidal grainstone that are interpreted as
the highstand-shoaling phase. Nonskeletal grainstones
are most common in the Bashkirian. Visean and Ser-
pukhovian cycles are generally easy to correlate lat-
erally, whereas Bashkirian cycles are best explained by
the transition to high-frequency icehouse sea level
Late Visean to Bashkirian Platform Cyclicity in the Central Tengiz Buildup / 53

FIGURE 34. Cross section showing lithofacies, gamma-ray (left), and porosity (right) logs; well locations are shown in Figure 4, and lithofacies legend is shown

bitumen-reducing porosity, dispersed volcanic ash in high-energy grainstone facies locally has an even greater effect on porosity in the central platform. Volcanic
fluctuations, which generate incomplete cycles and

ash dispersed in the SPG and CgPR lithofacies in the Serp1_csb sequence at T-5447 reduces porosity to nearly 0%, generating a flat porosity curve, whereas
in Figure 5. The lateral facies and thickness continuity of sequences between Serp_SSB and Lvis13_csb are shown for wells T-5447 and T-220. In addition to
complex lateral facies changes.
The distribution of reservoir quality in the central
platform was initially determined by early diagen-
esis associated with cyclic deposition, which resulted
in a stacked system of high-porosity intervals bor-
dered by tight higher order sequence boundaries with
reduced dissolution associated with the presence of
volcanic ash. This diagenetic phase includes low-
porosity, thick (discontinuous and not platformwide)
grainstone intervals with high amounts of dispersed
volcanic ash. The lateral continuity of tight cycle
boundaries probably greatly affected later fluid flow
as well as the ultimate distribution of cements, dis-
solution, and bitumen in the central platform res-
ervoir. Burial diagenetic overprint, an initial corrosion
phase followed by a pore-filling phase, and finally,
bitumen invasion and associated corrosion, produced
an overall porosity reduction toward the exterior cen-
tral platform. Bitumen content decreases from the
exterior to the central platform and is highest near the
bases of the cycles. This suggests that the first fill of
hydrocarbons migrated through the flanks laterally
into the platform cycles. Both corrosion and bitumen
emplacement flattened the initial cyclic porosity va-
riations and further complicated porosity and perme-
the same facies interval at T-220 retained high porosity. See text for discussion.

ability relationships.

ACKNOWLEDGMENTS

The authors thank the following individuals, who

provided insightful comments or suggestions and,
in some cases, data to greatly improve this manuscript.
Perhaps most importantly, Michael Clark (Chevron)
is thanked for initiating an intense coring and core
analysis program that formed the basis for most of
our geologic studies and for his strong support of the
sequence-stratigraphic framework. Ben Robertson
(Chevron) provided important discussion during the
continued coring program and supporting geological
analyses of outcrop analogs and literature reviews.
Numerous coworkers in Chevron, ExxonMobil, and
TengizChevroil assisted with our core and log studies
and participated in lively discussions on many topics.
Giovanna Della Porta (Cardiff University) and Frans
van Hoeflaken (independent consultant) are acknowl-
edged for critically reviewing the petrographic and
lithofacies descriptions. Steve Jenkins (TengizChev-
roil) offered numerous comments and observations,
which tested the ideas expressed in this study in
their practical application for three-dimensional
54 / Kenter et al.
geologic modeling of the Tengiz platform. Comments
by AAPG reviewers helped to greatly improve the
manuscript. Finally, we thank TengizChevroil and its
shareholder companies (Chevron, ExxonMobil, Kaz-
munaigaz, and BPLukArco) for permission to publish
this study.
REFERENCES CITED
Brenckle, P. L., and N. V. Milkina, 2001, Foraminiferal
timing of carbonate deposition on the Mississippian –
early Pennsylvanian Tengiz platform, Kazakhstan:
Paleoforams 2001 — International Conference on
Paleozoic Benthic Foraminifera, Abstracts, Ankara,
August 20 – 24, p. 2001.
Choquette, P. W., and L. C. Pray, 1970, Geologic no-
menclature and classification of porosity in sedimen-
tary carbonates: AAPG Bulletin, v. 54, p. 207 – 250.
Collins, J. F., J. A. M. Kenter, P. M. Harris, G. Kuanysheva,
D. J. Fischer, and K. L. Steffen, 2006, Facies and
reservoir-quality variations in the late Visean to Bash-
kirian outer platform, rim, and flank of the Tengiz
buildup, Precaspian Basin, Kazakhstan, in P. M. Harris
and L. J. Weber, eds., Giant hydrocarbon reservoirs of
the world: From rocks to reservoir chacracteriza-
tion and modeling: AAPG Memoir 88/SEPM Special
Publication, p. 55 – 95.
Cook, H. E., W. G. Zempolich, V. G. Zhemchuzhnikov,
and J. J. Corboy, 1997, Inside Kazakstan: Cooperative
oil and gas research: Geotimes, v. 42, no. 11, p. 16 – 20.
Esteban, M., and C. Taberner, 2003, Secondary porosity
development during late burial in carbonate reservoirs
as a result of mixing and/or cooling of brines: Journal of
Geochemical Exploration, v. 78 – 79, p. 355 – 359.
Gradstein, F. M., J. G. Ogg, and A. G. Smith, 2004, A
geologic time scale: International Commission on
Stratigraphy (ICS) under: www.stratigraphy.org.
Lucia, F. J., 1995, Rock-fabric/petrophysical classification
of carbonate pore space for reservoir characterization:
AAPG Bulletin, v. 79, p. 1275 – 1300.
Lucia, F. J., 1999, Carbonate reservoir characterization:
New York, Springer-Verlag, 222 p.
Moore, C. H., 2001, Carbonate reservoir-porosity evolu-
tion and diagenesis in a sequence stratigraphic frame-
work: Developments in Sedimentology, v. 55, p. 444.
Ross, C. A., and J. R. P. Ross, 1985, Carboniferous and Early
Permian biogeography: Geology, v. 13, p. 27 – 30.
Sattler, U., V. Zampetti, W. Schlager, and A. Immenhau-
ser, 2004, Late leaching under deep burial conditions:
A case study from Miocene Zhujiang carbonate reser-
voir: South China Sea: Marine and Petroleum Geology,
v. 21, p. 977– 992.
Weber, L. J., B. P. Francis, P. M. Harris, and M. Clark, 2003,
Stratigraphy, lithofacies, and reservoir distribution,
Tengiz field, Kazakhstan in W. M. Ahr, P. M. Harris, W. A.
Morgan, and I. D. Somerville, eds., Permo-Carboniferous
carbonate platforms and reefs: SEPM Special Publica-
tion 78 and AAPG Memoir 83, p. 351 – 394.
Zampetti, V., W. Schlager, J. H. Van Konijnenburg, and
A. J. Everts, 2003, Architecture and growth history of a
Miocene carbonate platform from 3D reflection data:
Luconia Province, offshore Sarawak, Malaysia: Marine
and Petroleum Geology, v. 21, p. 517 – 534.
 2
Collins, J. F., J. A. M. Kenter, P. M. Harris, G. Kuanysheva, D. J. Fischer, and
K. L. Steffen, 2006, Facies and reservoir-quality variations in the late Visean to
Bashkirian outer platform, rim, and flank of the Tengiz buildup, Precaspian
Basin, Kazakhstan, in P. M. Harris and L. J. Weber, eds., Giant hydrocarbon
reservoirs of the world: From rocks to reservoir characterization and mod-
eling: AAPG Memoir 88/SEPM Special Publication, p. 55 – 95.

Facies and Reservoir-quality

Variations in the Late Visean to
Bashkirian Outer Platform, Rim,
and Flank of the Tengiz Buildup,
Precaspian Basin, Kazakhstan
J. F. Collins G. Kuanysheva
ExxonMobil Development Company, TengizChevroil, Atyrau, Kazakhstan
Houston, Texas, U.S.A.
D. J. Fischer
J. A. M. Kenter1 TengizChevroil, Atyrau, Kazakhstan
Vrije University, Amsterdam, Netherlands
K. L. Steffen
P. M. Harris ExxonMobil Development Company,
Chevron Energy Technology Company, Houston, Texas, U.S.A.
San Ramon, California, U.S.A.

ABSTRACT

T
engiz field is an isolated carbonate buildup in the southeastern Precaspian
Basin, containing a succession of shallow-water platforms ranging in age
from late Famennian to early Bashkirian. Platform backstepping from Tour-
naisian through late Visean resulted in approximately 800 m (2625 ft) of bathy-
metric relief above the Famennian platform. This was followed by as much as
2 km (1.2 mi) of Serpukhovian progradation, which formed a depositional wedge
around the older platforms referred to as the Serpukhovian rim and flank.
Rim and flank facies include lower slope mudstone, volcanic ash, and platform-
derived skeletal packstone to grainstone interbedded with boundstone breccia;
middle-slope poorly bedded to massive boundstone breccia with subtypes based
on clast composition, size, and packing; upper-slope in-situ microbial bound-
stone; and outer-platform to shallow-platform skeletal, coated-grain, and ooid
packstone to grainstone. The upper-slope microbial boundstone represents the

1
Present address: Chevron Energy Technology Company, San Ramon, California, U.S.A.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215874M881469

55
56 / Collins et al.

dominant source of clasts in the middle- and lower-slope breccias. Periodic large-
scale failure of the rim during both Serpukhovian and Bashkirian time resulted in
a high degree of lateral facies discontinuity.
Solution-enlarged fractures, large vugs, and lost circulation zones produced
mainly during late diagenesis form a high-permeability, well-connected reser-
voir in the rim and flank. This diagenetic overprint is associated with the pres-
ence of bitumen and extends upward into overlying Serpukhovian and Bash-
kirian platform facies and inward into adjacent late Visean platforms, where it
has substantially altered reservoir properties that remained after early diagenesis
related to cyclic depositional processes.

INTRODUCTION STRATIGRAPHY

The Tengiz field is located in onshore western Ka-
zakhstan at the northern end of the Caspian Sea’s
eastern shoreline, among an archipelago of isolated
carbonate buildups that grew on a regional pre-Middle
Devonian structural high (Figure 1). Tengiz was dis-
covered in 1979 by the Oil Ministry of the former
Soviet Union and, since 1993, has been operated by
TengizChevroil (TCO), an in-country joint-venture
company. The field is covered by a three-dimensional
(3-D) seismic survey encompassing about 1000 km2
(385 mi2), acquired in 1998. As of year-end 2005, more
than 115 wells have penetrated the reservoir at Tengiz
(Figure 2), which produces an intermediate sulfur
oil (478 API). The highest initial production rates are
from wells located in the buildup rim and flank areas,
in fractured carbonates with low (<6%) matrix poros-
ity. Wells located toward the center encounter higher
porosity (as much as 18%) but lower matrix perme-
ability (<10 md) and lower initial rates.
The Tengiz buildup formed sometime after the Mid-
dle Devonian and grew more or less uninterrupted
until the early Bashkirian by aggradation of a series of
grain-dominated platforms. The total thickness of car-
bonate is unknown, but is represented by approxi-
mately 1.2 s TWT on 3-D seismic data. The deepest
borehole to date bottomed in the middle Famennian at
6032 m (19,790 ft) subsea, more than 2100 m (6890 ft)
below the top of the reservoir and more than 700 m
(2297 ft) below the top of the Famennian platform.
Following platform demise in the Bashkirian, the
Tengiz buildup was buried initially by a thin veneer
of Moscovian through Artinskian carbonates and vol-
caniclastics and then by thick Kungurian evaporites
(mainly halite) that completely encased the platform.
Post-Permian salt diapirism produced a series of salt
pillars and clastic-filled withdrawal basins above the
platforms, which locally adversely affect 3-D seismic
image quality (Figure 3A).
In 2001, a joint study by ExxonMobil, Chevron,
and TCO established a stratigraphic framework for
the Tengiz platform based on seismic, biostratrigraphic,
core, and well-log data that existed up to that time.
This study was published (Weber et al., 2003) and
stands as the definitive stratigraphic reference for Ten-
giz. The published framework includes a hierarchy
of second-order, third-order, and fourth-order sequences
that have been maintained as Tengiz drilling pro-
ceeds. This architecture is a locally applicable, fit-for-
purpose framework, although it more or less con-
forms to the second-order eustatic cycles of Ross and
Ross (1987).
This chapter describes facies and reservoir-quality
variations in the Visean A through Bashkirian plat-
form margins at Tengiz (Figure 3B). This interval forms
the late highstand of a second-order supersequence
set that began with deposition of a series of back-
stepping platforms in the Tournaisian. By the end
of the Visean, the platform area had decreased from
210 km2 (81 mi2) in the late Famennian to about
90 km2 (35 mi2). During the Serpukhovian, the Tengiz
platform prograded as much as 2 km (1.2 mi) basin-
ward, filling much of the accommodation space cre-
ated during backstepping, and forming a microbial
boundstone-cored depositional wedge known as the
rim and flank around the late Visean and older plat-
forms. This wedge attained a maximum thickness of
more than 800 m (2625 ft). Following a brief deposi-
tional hiatus on the platform (Brenckle and Milkina,
2003), early Bashkirian platform carbonates accumu-
lated more or less aggradationally above the Serpu-
khovian. Because this chapter also discusses reservoir
connectivity between the Serpukhovian rim and flank,
the overlying Bashkirian outer platform region, and
the adjacent Visean A platforms, a brief description of
how platform facies correlate to rim and flank facies
is included here. For a detailed account of the nature
FIGURE 1. North Caspian region and location of the Tengiz field.
The Outer Platform, Rim, and Flank of the Tengiz Buildup / 57
58 / Collins et al.
of the Visean A, Serpukhovian, and Bashkirian cen-
tral platform regions, the reader is referred to Kenter
et al. (2006).
Visean A, Serpukhovian, and
Bashkirian Platform Cycles
Visean A, Serpukhovian, and Bashkirian fourth-
order central platform sequences generally adhere
to the model shown in Figure 4. These sequences
feature depositional environments ranging from high-
energy grainstone shoals to slightly deeper environ-
ments characterized by more poorly sorted, open-
marine grainstone to packstone, or in some cases,
mudstone and faunally restricted, fine-grained car-
bonate sand. Outer platform cycles vary over the in-
terval but are generally characterized by an increase
in algal grains and micritic intraclasts. Because of the
unique nature of microbial facies in the rim and
flank, greater attention is given to Serpukhovian
outer-platform cyclicity. These cycles are commonly
poorly expressed and range in thickness from less
FIGURE 2. Structure contours
on top of the Tengiz buildup.
Current well locations and the
path of the seismic profile in
Figure 3 are indicated. Contour
interval = 100 m (330 ft).
than 10 m (33 ft) to several
tens of meters. They consist
of dipping, bedded to mas-
sive skeletal packstone to
rudstone containing abun-
dant large crinoids and bra-
chiopod fragments, minor
colonial coral detritus, and
occasional platform-derived
breccias. In the central plat-
form, fourth-order sequence
boundaries are associated
with subaerial exposure and
deposition of low-energy or
restricted facies containing
thin volcanic ash layers (Ken-
ter et al., 2006). In the Ser-
pukhovian outer-platform
cycles, these features are re-
duced or absent at the se-
quence boundaries.
DEPOSITIONAL ENVIRONMENTS
General Features of the Tengiz Rim and Flank
The terms ‘‘platform,’’ ‘‘rim,’’ and ‘‘flank’’ were
originally derived from seismic features and are now
used for drilling purposes to denote geographic re-
gions in the field. The platform region comprises
boreholes that penetrate the stack of backstepping
platforms, whereas boreholes that encounter any part
of the adjacent progradational wedge are described
as rim and flank wells. As mentioned earlier, the
wedge arises as a result of Serpukhovian prograda-
tion, but includes rocks of Serpukhovian and Bash-
kirian age (see Figure 3). The rim area refers to the
inner, structurally elevated rim crest or raised rim,
whereas the flank refers to the outer sloping surface.
The raised rim is interpreted to have formed mainly
by differential compaction caused by a mechanical
contrast between grainy platform facies and more
rigid upper-slope microbial boundstone facies in the
rim and has resulted in as much as 50 ms of seismic
relief and as much as 200 –250 m (656 – 820 ft) of
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 59

FIGURE 3. Tengiz stratigraphic framework and the Serpukhovian rim and flank region. The seismic line (A) shows the
east-west asymmetry in the shape of the wedge and in the width of the structurally elevated portion (raised rim). The
seismic profile also shows the absence of Moscovian – Artinskian volcanic sediments on the upper flank of the buildup,
placing an anhydrite layer at the base of the salt in direct contact with the reservoir, as well as the source rock interval
associated with subsalt facies in the same interval in surrounding lows. These are two potentially important factors in rim
and flank diagenesis. The schematic framework (B) is derived from Weber et al. (2003), with details from this study
added in the rim and flank area. In particular, the diagram shows the rim and flank to consist of an early debris apron
that accumulated around the base of the buildup, followed by progradation of a boundstone-cored upper wedge (black
dashed line).
60 / Collins et al.

FIGURE 4. Idealized Tengiz fourth-order platform sequence, showing textural and compositional variations associated
with platform bathymetry for the Bashkirian, Serpukhovian, and late Visean. Outer-platform, rim, and flank facies
illustrated apply particularly to the Serpukhovian. For other intervals, outer-platform facies, in general, are characterized
by an increase in algal grains, micritic intraclasts, and coarse crinoid-brachiopod debris. Sea level lowstands are
associated with exposure of shallow-platform shoals, contemporaneous formation of an outer margin complex of
algal-intraclast grainstones and possible coral-algal patch reefs, and local development of restricted conditions on the
platform. Partly restricted-marine conditions that persisted during the early transgression were dominated by bra-
chiopods, including the formation of outer-platform brachiopod mud mounds. Widths of facies belts are not shown
to scale.

structural relief1 (see Figure 2). Evidence for platform
compaction includes:
1) a lack of volumetrically significant deep lagoonal
facies in the central areas of the Visean A, Ser-
pukhovian, and Bashkirian platforms
2) upper Serpukhovian and Bashkirian platform
cycles in the raised rim that contain shallow fa-
cies similar to age-equivalent cycles in the cen-
tral platform
3) significant grain-to-grain compaction observed
in thin sections from the platform facies
4) observation of high early cement volumes in the
rim and flank facies, indicating early develop-
ment of mechanical rigidity
5) a lateral facies progression in Serpukhovian cy-
cles that deepens basinward from the platform
1
Weber et al. (2003) estimated that approximately two-thirds of the
raised rim relief was caused by differential compaction, and one-third was
caused by a bioconstructed margin associated with upper Bashkirian cycles.
into the rim, based on comparisons between Ten-
giz and other microbial boundstone-cored plat-
forms (Kenter et al., 2005).
The break in slope separating the raised rim and
flank areas does not necessarily indicate the limit
of Serpukhovian progradation. In some regions, this
break appears to be controlled by faulting or mass
wasting. In addition, differential compaction of a
basinward-thickening wedge of middle- to lower-
slope facies may be responsible for the slight basin-
ward dip of the raised rim observed in some areas.
Rim and Flank Depositional Environments
The rim and flank contain carbonate facies depos-
ited in water depths ranging from a few meters to sev-
eral hundred meters. Facies descriptions come mainly
from cores2 and from Formation MicroImager (FMI)
2
Approximately 1200 m [3937 ft] of core from 15 wells available as of
year-end 2005 were examined for this study.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 61
logs calibrated to cores. Facies interpretation from FMI
logs was hampered by use of oil-based mud during
drilling, which tended to reduce conductivity con-
trasts between porous and nonporous rock, and by
frequent hole washouts and zones of lost circulation
that adversely affected image quality. Rim and flank
facies are associated with a depositional profile con-
sisting of the following environments, in the order
of increasing paleowater depth:
1) cyclic shallow-platform deposits
2) outer-platform skeletal-intraclast rudstone, grain-
stone, packstone, and minor boundstone
3) upper-slope microbial boundstone
4) middle slope welded to mosaic boundstone
breccia
5) lower-slope detrital boundstone breccia and grainy
platform-derived deposits
6) lower-slope and debris apron boundstone breccias
and bedded, fine-grained periplatform deposits.
These are partly based on environments and fa-
cies recognized in outcrops of the Asturias plat-
form margin in the Cantabrian region of northern
Spain and from the Capitan margin in west Texas
and New Mexico (Figure 5). The Tengiz outer plat-
form, slope, and debris apron facies are described
in the ensuing paragraphs and are also summa-
rized in Appendix 1. The cyclic, shallow-platform
facies are described in Kenter et al. (2006).
Outer-platform Facies
The Asturias and Capitan margins were both formed
by deeper water microbially influenced boundstone
margins (Kirkland et al., 1998; Della Porta et al., 2003)
and support the interpretation of considerable origi-
nal outer-platform relief above the Tengiz upper slope
during the Serpukhovian. The Asturias margin had
outer-platform relief of between 20 and 50 m (66 and
164 ft) during times of progradation (Bahamonde
et al., 2000; Kenter et al., 2005). Estimates of the re-
lief across the Capitan outer platform margin show
variations through time from 10 m to more than
60 m (33 to more than 197 ft) (Kerans and Harris,
1993; Tinker, 1998). Total paleobathymetric relief
from the shallow Tengiz platform to the start of
the upper-slope microbial boundstone at the edge
of the outer platform was at least tens of meters
(Weber et al., 2003) and probably higher. Apparent
thickening in the outer platform of the uppermost
Serpukhovian fourth-order cycles, plus the limited
presence of subaerial exposure at the tops of outer-
platform cycles, indicates that relief exceeded at
least some of the ambient low-amplitude eustatic
variations, estimated to be between 25 and 50 m
(82 and 164 ft) (Ross and Ross, 1987, 1988).
Whereas the succession of facies comprising cy-
cles in the central platform area is well documented
in cores (see Kenter et al., 2006), the descriptions of
Serpukhovian outer-platform cycles contained here
are generalizations based on a limited number of
complete cycles present in the current suite of cores.
The basal parts of some outer-platform cycles were
observed to contain breccias composed of colonial
coral detritus or, in other cases, rounded, platform-
derived grainstone clasts (Figure 6A, B). The breccias
contain either microbial matrix textures and cements
or a matrix of coarse skeletal debris. They may repre-
sent lowstand deposits, indicating updip platform
exposure, or development of outer-margin coral patch
reefs. In other sequences, brachiopod mud mounds a
few meters thick (Figure 6C) are encountered near
the base. The middle and upper parts of cycles are
characterized by wackestones, packstones, and rud-
stones containing large crinoids, brachiopod shell
fragments, and solitary rugose corals (Figure 6D), in-
dicating an increase in water depth. This facies is
associated with dipping beds (Figure 6E), indicat-
ing deposition on a paleoslope, and may also ex-
hibit microbial matrix textures. Cycle tops tend to be
characterized by a reduction in crinoid-brachiopod-
coral grains and an increase in coated-grain or algal-
intraclast grainstone, resembling some shallow-
platform grainstones. Subaerial exposure effects like
those described from the central platform (Kenter
et al., 2006) are commonly absent, but large disso-
lution voids filled with multilayered geopetal sedi-
ment and multiple generations of carbonate cement
are found, as are brecciated grainstone intervals
with evidence of clast rotation and mixing. Iso-
lated, sediment-filled dissolution voids are also scat-
tered throughout the section, but appear to increase
in frequency at or near the tops of cycles. These may
be associated with longer term exposure at superse-
quence boundaries (Serp_SSB and Bash_SSB), or they
may be unrelated dissolution features whose distri-
bution was influenced by stratigraphic permeability
variations.
Upper- to Middle-slope Facies
The upper slope consists mainly of in-situ microbial
boundstone and boundstone breccia. Similar facies
are recognized from the Asturias margin (Bahamonde
et al., 2000; Della Porta et al., 2003) as mentioned
62 / Collins et al.

FIGURE 5. Outcrop analogs with facies and depositional relationships similar to the Tengiz outer platform, rim, and
flank. Like Tengiz, the platform margins illustrated are characterized by dipping, basinward-deepening platform facies
that pass downward into upper-slope, microbially influenced boundstones and foreslope facies with reef-derived breccias.
Based on these profiles, the structurally raised rim visible on Tengiz seismic data (see Figure 3) is largely an artifact
of postdepositional differential compaction (see text for discussion). Microbial boundstone types A, B, and C noted in the
Asturias example closely match boundstone types recognized in Tengiz cores.

previously, from Triassic platforms in Italy (Blendin-
ger, 2001) and from Famennian platforms in the
Canning basin (Playford, 1984; Kerans et al., 1986;
Stephens and Sumner, 2003). Tengiz cores and FMI
logs suggest that the in-situ boundstone interval is
perhaps 150– 200 m (492–656 ft) thick. In-situ mi-
crobial boundstones (Figure 7) are light colored in
core, with textures that include relatively featureless
micritic to peloidal fabrics (type A) or irregular lami-
nar fabrics and amalgamated semiconcentric laminar
masses (type B). A third boundstone type (type C),
characterized by massive to peloidal fabric with com-
mon skeletal grains (bryozoans, algae, forams, ostracods,
pelecypods, crinoids, corals, and sponges), represents
the transition between upper-slope and outer-platform
facies. Small- to medium-size fenestrae filled with banded
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 63

FIGURE 6. Serpukhovian outer-platform facies. Breccias composed of platform-derived clasts (A) or colonial coral
fragments (B) occur toward cycle bases. The breccia in (A) contains cement and clotted fabric of possible microbial origin,
whereas the matrix in (B) consists of coarse crinoidal rudstone. Brachiopod mud-mounds (C) as much as a few meters
thick also sometimes occur near cycle bases. The middle and upper parts of typical outer-platform cycles contain
abundant packstone and wackestone with large crinoids, brachiopods, and solitary corals (D), as well as intervals with
dipping beds (E), indicating deposition on a paleoslope.

cements are commonly abundant, and larger cement-
lined cavities filled with internal sediment or skeletal
accumulations are observed. Boundaries between mi-
crobial masses are sometimes formed by networks of
elongate cavities that are also frequently cement filled.
The boundstone interval at Tengiz could be ex-
tended down to more than 300 m (984 ft) if certain
types of breccias are included. The upper slope inter-
val in the Asturias margin is comparable in thickness
and includes both in-situ microbial boundstone, and
nearly in-situ boundstone breccia generated by rap-
id marine cementation of incipient boundstone fail-
ure (Bahamonde et al., 2000; Della Porta et al., 2003).
To date, similar breccias held together by thick ma-
rine cements have not been observed in abundance
at Tengiz, although core coverage is still limited.
However, breccias consisting of irregularly shaped to
subangular boundstone clasts with stylolitic (welded
breccia) or fracture (mosaic breccia) contacts are found
in the Tengiz middle to upper slope (Figure 8) and
may be analogous to the Asturias intact breccias. The
matrix consists mainly of several generations of calcite
cement, secondary microbial encrustations, micro-
bial cement, and minor amounts of platform-derived
skeletal sediment, thin-bedded marly volcanics, or
argillaceous carbonate. Very little open primary pore
space was observed in these breccias.
Formation MicroImager logs over upper-slope mi-
crobial boundstone and breccia intervals show little
evidence of cyclicity, suggesting that sea level varia-
tions that affected the platform had little or no effect
on production of microbial boundstone. As a result of
deposition below the range of sea level fluctuations
and productivity across a wide range of water depths,
64 / Collins et al.

FIGURE 7. Serpukhovian microbial boundstone facies, shown in order of increasing paleodepth. (A) Type C boundstone
(skeletal-micritic) with corals and crinoids, and numerous cement-filled, irregular-shaped fenestrae. The matrix is
skeletal rich with micritic to peloidal microbial texture. (B) Type C boundstone containing numerous bryozoans (arrows)
and thick-banded cements of probable marine origin. (C) Type B boundstone (laminar cemented), characterized by
irregular to rounded, concentrically laminated microbial masses. These boundstones contain as much as 75% cement.
(D) Type A boundstone (peloidal-micritic), featuring a dominantly peloidal to micritic matrix, scarce fauna, and numerous
fenestrae or tubular voids.

including subphotic depths, the boundstones were
apparently able to prograde continuously seaward
during all stages of sea level rise and fall.
Middle- and Lower-slope Facies
The middle and lower slope consists of mainly de-
trital facies (Figure 9), including sedimentary bound-
stone breccias; coarse skeletal rudstones; fine-grained
breccias; massive or poorly bedded allochthonous
grainstones; and thin-bedded to laminated grainstones,
packstones, wackestones, mudstones, argillaceous car-
bonates, volcanic ashes, and cherts. Clast size and pack-
ing varies in the breccias, with textures ranging from
matrix-supported floatbreccia to welded breccia with
stylolite-bounded clasts (Figure 9D – F). Clast sizes
range from less than 1 cm (0.4 in.) to several meters.
The matrix contains variable amounts of bioclastic
debris, fine-grained carbonate, argillaceous mudstone,
and volcanic ash. The most visible skeletal components
are large crinoids and brachiopod shell fragments
similar to outer-platform facies. Relationships be-
tween clast size, clast packing, matrix grain size, and
matrix composition have not been fully evaluated,
but they appear to be poorly correlated. Clasts are
dominantly upper-slope microbial boundstone frag-
ments, although platform-derived clasts and reworked
slope clasts are increasingly observed in breccias near
the base of the section. Some detrital breccias are mas-
sive and unbedded over very thick intervals, whereas
others are layered and contain intervals of bedded,
grainy periplatform facies. Formation MicroImager logs
again do not indicate an ordered cyclicity that can be
related to the platform sequences.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 65

FIGURE 8. Upper- to middle-slope diagenetic and sedimentary breccias. (A) Fractured microbial boundstone healed
by banded, probably marine cements. (B) Partly cemented mosaic breccia containing earlier cemented voids dissected
by later fractures and voids containing bitumen and equant calcite spar (arrows). Larger open voids are mostly a product
of leached cement fills in the secondary breccia. Note also that the early, cemented fracture in (A) is also intersected by
a later, bitumen-filled fracture. (C) Mosaic to welded breccia containing microbial boundstone clasts with stylolitic contacts.
(D) Sedimentary floatbreccia with irregularly shaped boundstone clasts, and crinoid-rich, skeletal-lithoclast rudstone
matrix.

DEPOSITIONAL HISTORY OF THE
TENGIZ RIM AND FLANK
The additional facies control provided by FMI logs
(Figure 10) proved helpful in interpreting the rim
and flank regions. Attempts to extrapolate fourth-
order platform correlations into the rim and flank on
ordinary wire-line logs proved difficult because of the
presence of apparently nonstratiform porosity vari-
ations and diagenetic effects manifested as anoma-
lous radiogenic responses on gamma-ray logs. Addi-
tionally, rim and flank stratigraphy is complicated by
large-scale failure and mass wasting because of insta-
bility, both during and after deposition. Seismic and
biostratigraphic control indicates that the interval con-
tains breccias of Serpukhovian age as well as Bashkirian-
aged breccias with Serpukhovian clasts. Boundstone
breccia is the dominant facies present in rim and
flank cores recovered to date; therefore, its distribu-
tion is a critical element of reservoir prediction. This
requires a deeper understanding of depositional and
failure processes at the Tengiz margin.
The analogous Asturias and Capitan platforms are
also associated with middle- and lower-slope breccia
deposits. The Asturias breccias occupy a basinward-
thickening middle- to lower-slope wedge (see Figure 5),
whereas the Capitan margin is associated with both
middle- and lower-slope breccias and basinal debris
flows out in front of the shelf (Melim and Scholle,
1995). These dominantly progradational margins have
maximum upper-slope angles of 30 – 358 (Tinker,
1998; Della Porta et al., 2003; Kenter et al., 2005).
Slope angles for the Tengiz rim and flank are known
from structural and seismic profiles of the Bashkirian
66 / Collins et al.

FIGURE 9. Middle to lower slope and debris apron facies. Middle- to lower-slope facies include coarse skeletal (crinoid-
brachiopod) rudstone (A), skeletal wackestone to packstone with dipping beds (B), and thin-bedded volcanics,
mudstone, packstone, and fine grainstone (C). Debris apron facies include coarse breccia with meter-size boundstone
clasts (D), packbreccia with platform-derived, crinoid-rich matrix (E), and float- to packbreccia interbedded with
allochthonous grainstone and packstone (F). Note the relatively low dip angles in (F).

surface and from dipmeter data. In the steeper sloped
upper flank, the Bashkirian is thin, consists of mostly
fine-grained deeper water facies, and, thus, forms a
prominent seismic reflector that closely approximates
the Serpukhovian profile. Measured angles vary be-
tween 20 and 258 for apparent accretionary slopes,
whereas angles approaching 308 are measured in areas
of obvious detachment or failure. These measure-
ments represent angles determined from a compacted
body of rock and do not necessarily reflect original
depositional dips. Because of the chaotic nature of
boundstone and breccia facies at Tengiz, FMI dip mea-
surements commonly produce inconsistent results,
but some intervals indicate that preserved bedding
dips of 25 –308 in the upper slope may be typical, and
angles of 35 – 458 occur locally. Directionality is vari-
able, however, and dips do not always point directly
away from the platform. Multiple azimuth changes
have been recorded vertically in a single well or in
comparing two nearby wells. Formation MircoImager
logs also suggest that facies distribution is more com-
plex than expected for simple progradation. Whereas
the Asturias and Capitan margin profiles have regu-
lar shapes, the Tengiz flank has an irregular shape on
seismic profiles, even considering geometries created
by depositional or compactional drape over under-
lying backstepping platforms. This shape varies sys-
tematically around the perimeter of the buildup and,
combined with facies obtained from core and FMI
logs, allows the rim to be divided into subregions gov-
erned by different depositional processes.
Allochthonous and Accretionary Rim Sectors
The rim and flank area consists of two distinct
regions, the allochthonous and accretionary sectors
(Figure 11). The allochthonous sector refers to the
southwestern and northwestern parts of the rim. It
is characterized by a lower, distally thickened apron
that onlaps the Tournaisian and Visean platforms,
which itself appears to be downlapped by a narrow
progradational wedge (Figure 12A). The raised rim is
reduced or, in some places, completely absent along
this trend (Figure 12B). The accretionary sector re-
fers to the northern, northeastern, and southeastern
regions of the rim, where it is characterized by a con-
spicuous raised rim crest and a considerably wider
progradational wedge (Figure 13A).
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 67

FIGURE 10. Formation MicroImager log to core facies calibrations. FMI facies, indicated by different colors in the depth
columns (depths are in meters), include boundstone-breccia (BBr), boundstone-horizontal fabric (Bhz), breccia, rud-
stone, grainstone, and laminated fine-grained (lam-FG). BBr facies, characterized by a chaotic image lacking horizontal
layering (C – E), includes both boundstones and breccias from core, which, except occasionally, were not distinguish-
able on FMI logs. Bhz facies corresponds to poorly bedded outer platform skeletal rudstone and type C microbial
boundstone (F). FMI rudstone is characterized by nodular texture and some horizontal layering (A, D) and includes
lithoclast-skeletal rudstones, fine-grained breccias, and poorly bedded skeletal packstone, wackestone, and floatstone
from core. Formation MicroImager grainstones, identified by a low-contrast image with faint layering (B), correspond
to generally tightly cemented allochthonous grainstones. Lam-FG facies, marked by frequent high-contrast layers (A),
are equivalent to thin-bedded, deep-water volcanics, mudstones, and grainstones in core. In some intervals, image quality
is adversely affected by bitumen, which produces a snowy texture overprinted on the facies (C).

In the allochthonous sector, the lower apron con-
sists of bedded to massive boundstone breccia in-
terlayered with bedded platform-derived sediments,
suggesting relatively continuous failure of a bound-
stone margin. In the northern half of the accretion-
ary sector, the flank slope is typically irregular. This is
partly caused by a series of local slump scars and ap-
parently bottomless faults, but also to a prominent mid-
slope bulge. The lateral continuity of this latter fea-
ture (see Figure 11) could be accounted for by slight
downslope movement of upper-slope material or al-
ternately by downlap of a late-stage progradational
wedge onto a protruding debris apron (Figure 13B),
analogous to the allochthonous sector. Seismic data
68 / Collins et al.
suggest the existence of an early debris apron in the
accretionary sector in another way as well. The north-
ern and southern margins of the Tengiz platform rep-
resent transitional regions between the allochthonous
and accretionary sectors. Both areas are characterized
by an upper wedge that gradually widens eastward
while maintaining a downlapping relationship to lat-
erally continuous debris apron deposits (Figure 14).
The accretionary and allochthonous sectors dem-
onstrate internal and comparative differences with
respect to the total volume of rock in the rim and flank,
as well as the relative proportions in the upper wedge
and lower apron. This asymmetry could result from
both extrinsic and intrinsic controls. For example,
one intrinsic control might be the pre-Serpukhovian
profile created by the stacking pattern of Tournaisian
and Visean platforms. In areas of significant backstep-
ping, a bench or gentle slope might have provided
stability for progradation (see Figure 13A), whereas in
areas where platform stacking is more vertical, insta-
bility and rim failure might dominate. Alternatively,
FIGURE 11. Dip magnitude on
the top of the reservoir (Bashkir-
ian surface) measured from 3-D
seismic data, showing the extent
of the accretionary and alloch-
thonous sectors. The map shows
a thin band of elevated dip
(dashed white line) that is di-
rected toward the platform inte-
rior caused by differential com-
paction between platform facies
and upper-slope microbial
boundstones and indicates the
position of the platform margin
prior to progradation. The zone
of low dip between this band
and the slope break is the raised
rim and indicates the limit of
progradation (possibly modified
by failure processes). The bound-
ary between the accretionary
and allochthonous sectors occurs
where the band of compactional
dip intersects the slope break
(white arrows). In the accre-
tionary sector, the slope break is
marked by several local scarps
indicative of localized failure
(yellow lines). In the allochtho-
nous sector, the raised rim is thin
or absent, indicating either nearly
complete rim failure or a poorly
developed progradational wedge.
The locations of seismic profiles
in Figures 12–15 are also shown.
primary microbial production may have varied around
the platform. In the accretionary sector, elevated pro-
ductivity could have resulted in faster accumulation
of a debris apron and, thus, earlier initiation of upper
wedge progradation. In a lower productivity alloch-
thonous sector, the debris apron required a longer
time to accumulate and was perhaps still in progress
at the end of the Serpukhovian, accounting for the
more limited upper wedge development.
Tectonic activity is a possible external mechanism
that could also account for rim asymmetry. Rim fail-
ure could have been restricted to, or enhanced in the
allochthonous sector because of alignment with re-
gional stress regimes or as a result of directly facing
pressure waves from the tectonically active southern
margin of the Precaspian Basin (see Figure 1). Whether
the apron is mostly an early or a late feature is there-
fore still somewhat uncertain because of these pos-
sibilities. The significant role of rim failure and the
timing uncertainty precluded erecting a stratigraphic
framework in the rim and flank that can be specifically
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 69

FIGURE 12. Seismic profiles

from the allochthonous rim
sector. In (A), a slightly raised
rim indicates a narrow pro-
gradational wedge of upper-
slope microbial boundstone
that formed after deposition
of an extensive debris apron.
The figure in (B) shows a re-
gion where the raised rim is
absent, indicating failed pro-
gradation. Formation Micro-
Imager logs from T-7040
indicate the upper wedge de-
posits at this location most
likely consist of coarse brec-
cia. See Figure 11 for profile
locations.

forms a mound-shaped pile

tied to Serpukhovian and Bashkirian fourth-order
platform cyclicity; however, core control, FMI facies,
biostratigraphic data, and, to a lesser extent, wire-line
logs were still used to separately constrain the dis-
tribution of facies in the lower apron and the upper
wedge as a result of their relative spatial separation.
Upper Wedge Facies Distribution
From previous discussions, the raised rim effect
is interpreted to indicate the presence of mechani-
cally rigid upper-slope microbial boundstone facies
in the upper wedge. In places along the allochtho-
nous sector, the upper wedge is partly or completely
detached from the platform, and the raised rim effect
is absent. In some cases, the wedge appears to have
detached more or less intact, whereas in others, it
that is almost continuous with
the distal apron (Figure 15).
To date, the upper wedge is
penetrated by 14 wells with
core slabs and/or FMI images
(T-463, T4346, T-4635, T-70,
T-7040, T-5056, T-5857, T-7252,
T-6457,T-4556,T-5059, T-5454,
T-7450, and T-7052). These
wells indicate facies consistent
with a prograding boundstone-
cored wedge in areas of accre-
tion in both sectors and facies
dominated by breccia in areas
of detachment in the alloch-
thonous sector (Figure 16).
Specific cored examples in-
clude cyclic shallow platform
overlying deeper water outer-platform facies (T-4556
and T-7052); upper-slope in-situ boundstone facies
(T-7450 and T-5056); and upper-slope boundstone
and welded-to-mosaic breccia from a detached but
relatively intact upper wedge (T-4635, Figure 15A).
Other cored examples include middle-slope breccias
(T-5056, T5857, and T-7252), and middle- to lower-
slope clast- and matrix-supported boundstone breccia
interbedded with platform-derived crinoid-brachiopod
rudstone to packstone (T-463). The Bashkirian in the
upper wedge consists of shallow-platform facies in
the raised rim area and, on the basis of mainly short
cores and core chips from older wells, deeper water
facies composed of laminated to thin-bedded mud-
stones, packstones, fine grainstones, and volcanics in
the flank.
70 / Collins et al.
Lower Apron Facies Distribution
The lower apron is penetrated by four wells in the
allochthonous sector with cores and/or FMI data
(T-16, T-6337, T-3938, and T-3948). Assuming that an
early apron is also present in the accretionary sector,
five other wells potentially have representative core
and/or FMI data (T-47, T-5963, T-5059, T-5660, and
T-6261). In particular, the FMI facies from T-5660 and
T-5059, combined with three separate cored inter-
vals from nearby T-5056 in the raised rim, indicate a
possible distinction between apron breccias and upper
wedge facies in the accretionary sector (Figure 17).
Apron facies are best known from cores in the al-
lochthonous sector, which reveal that, in general,
the apron consists of mixtures of upper-slope-derived
FIGURE 13. Seismic profiles
from the accretionary rim
sector. The flank profile in
(A) is smooth and lacks high-
amplitude internal reflections.
Here, a bench formed by a
backstepping of the Visean A
and B platforms may have
provided stability for early
upper wedge progradation.
The flank profile in (B) shows a
distinct midslope bulge. For-
mation MicroImager data
from T-5059 and T-5660 dem-
onstrate that the upper part of
the Serpukhovian interval in
the bulge contains a prograd-
ing boundstone sequence,
whereas the lower part con-
tains bedded boundstone
breccias (see Figure 17). The
change occurs near a zone of
high-amplitude, subparallel
reflectors on the seismic line,
which have been interpreted
to indicate areas in the accre-
tionary sector where deposi-
tion of early debris aprons
filled in some of the accom-
modation space as a prerequi-
site to upper wedge pro-
gradation. See Figure 11 for
profile locations.
boundstone breccia and
platform-derived grainy fa-
cies, but which also indicate
significant lateral variations
in bed thickness, clast size,
and amount of platform-
derived grainy material ver-
sus breccia. In two separate cored intervals from
T-6337, for example, both clast- and matrix-supported
breccias are frequently interbedded with fine-grained
periplatform mudstone, packstone, and grainstone.
Overall clast size is larger compared to the T-463 core
from the upper wedge, but the platform-derived frac-
tion is finer grained. By contrast, a long, cored inter-
val at the T-3948 core consists almost entirely of
poorly bedded boundstone breccia made up of very
large clasts, with very few interbeds of fine-grained
platform-derived material. The T-3938 FMI log sug-
gests that the apron interval there also consists of
massive boundstone breccia with few interbeds of
platform-derived facies. The lateral extent of this mas-
sive, coarser apron breccia facies appears to coincide
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 71

FIGURE 14. Seismic profiles

from the northern (A) and
southern (B) rim and flank
regions. These areas are tran-
sitional between the alloch-
thonous and accretionary sec-
tors and show both a lower
debris apron and significant
upper wedge progradation,
suggesting that variations in
wedge development did not
necessarily occur at the expense
of the apron. See Figure 11 for
profile locations.

sent there. Cores across the

base of rim and flank brec-
cias have been recovered at
T-6337, T-3938, T-3948, and
T-7252. Formation Micro-
Imager logs from these and
other wells demonstrate that
the breccias were deposited
abruptly on top of a con-
tinuous interval containing
mainly thin, dipping beds
of deeper water facies. Bio-
stratigraphic analyses from
the cores indicate that in both
sectors, the onset of breccia
deposition occurred close
to the Visean–Serpukhovian
boundary.
Biostratigraphic data also
demonstrate that Bashkirian-
aged sediments containing
Serpukhovian boundstone
clasts occupy as much as
300 m (984 ft) at the top of
with chaotic seismic character and, therefore, may be the apron breccias in the allochthonous sector, indi-
regionally mappable within limits imposed by image cating that significant postdepositional rim failure
quality (see Figure 16). occurred in that region. The Bashkirian interval ap-
Apron breccias in the allochthonous sector con- pears to be associated with a distinctive pattern on
tain increasing numbers of transported slope and gamma-ray and resistivity logs (Figure 18), and well
platform clasts toward the base of the interval. Cores correlations suggest that Bashkirian breccias are vari-
from T-3938 and T-3948 demonstrate that some- able in thickness but laterally extensive (see Figure 16).
what finer grained breccias with abundant platform The apparent absence of equally thick Bashkirian de-
and slope clasts are associated with a resistivity de- posits or Bashkirian-aged breccias in the accretionary
crease on petrophysical logs. This low-resistivity zone sector is an important distinction that perhaps indi-
is not observed on logs from the accretionary sector, cates that relative depositional stability was achieved
but because this could be caused by a change in po- in this region prior to the end of the Serpukhovian.
rosity or other diagenetic effect, its absence does not Cross sections illustrating the depositional concepts,
necessarily indicate that similar breccias are not pre- core and FMI facies control, and typical porosity and
72 / Collins et al.
gamma-ray expression of the lower apron and upper
wedge in the Tengiz allochthonous and accretionary
rim sectors are provided in Figures 18 and 19.
FIGURE 15. Seismic profiles illustrating
different modes of upper wedge fail-
ure. The figure in (A) shows a com-
pletely detached but largely intact upper
wedge. Core from T-4635 contains mainly
type C boundstones and breccias, indi-
cating original deposition in shallower
water. The figure in (B) shows a narrow
upper wedge with a downslope bulge,
indicating possible partial failure. The
debris apron in this area consists of
poorly bedded boundstone breccia, as
indicated on FMI logs from T-3938.
Biostratigraphy from drill cuttings re-
vealed a thick Bashkirian-aged section
at the top of the apron breccias. The
figure in (C) shows local fault detach-
ment in the upper wedge. In addition,
chaotic seismic character in the debris
apron is separated from the upper
wedge slope by a zone of relatively
continuous dipping reflections from a
distinct lower slope facies. T-3948 cores
reveal that the apron at this location
consists of massive boundstone breccia
with very large clasts. See Figure 11
for profile locations.
DIAGENESIS
As a long-lived shallow platform,
Tengiz was expected to have a com-
plex distribution of pore types be-
cause of its multifaceted stratigraphic
hierarchy and associated facies vari-
ations, repeated exposures to ma-
rine, meteoric, and shallow burial
diagenesis, and, because of the age
range involved, the influences of
both greenhouse and icehouse glob-
al themes. Core and thin-section ob-
servations, combined with prelimi-
nary geochemical analyses, provide
insights into processes of porosity
creation, destruction, and preserva-
tion in outer platform, rim, and flank
reservoirs and also illuminate rela-
tionships that determine reservoir-
quality distribution in the rim and
flank. These relationships suggest
that present-day reservoir-quality var-
iations are strongly influenced by a late diagenetic
overprint involving fracturing, dissolution, and bi-
tumen cementation.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 73

FIGURE 16. Facies distribu-

tion in the rim and flank. The
Visean A outline indicates
the platform configuration
just prior to Serpukhovian
progradation. The raised rim
area (upper wedge) repre-
sents the preserved extent of
Serpukhovian platform pro-
gradation. The outer limit of
the upper-slope boundstone
was determined by an as-
sumed vertical thickness of
about 200 m (656 ft) below
the slope break. The lower
apron contains Serpukhovian
and Bashkirian coarse brec-
cias, mapped from seismic
character and well-log corre-
lations. Serpukhovian brec-
cias occur on both sides of the
platform, whereas Bashkirian
apron deposits are generally
restricted to the allochtho-
nous sector, indicating depo-
sitional instability in that area
throughout the Serpukhovian
and Bashkirian. Facies control is
indicated for rim and flank
wells only.

tant for platforms like Ten-

Early Dissolution, Cementation, and
Bitumen Formation
Early pore types consist of primary and early dia-
genetic matrix pores and larger voids from at least
one significant dissolution event. Existing cores sug-
gest that this early dissolution was more prevalent in
the outer-platform and upper-slope facies compared
to the middle and lower slope. Early dissolution may
have occurred during platform margin karst events
associated with major sequence boundaries or as a
result of Kohout-type convection driven by large fluid-
density contrasts between platform-top groundwa-
ters and seawater (Kohout et al., 1977; Saller, 1984;
Sanford et al., 1998). Such contrasts, controlled by
temperature and salinity, could be especially impor-
giz with considerable verti-
cal height.
Early voids commonly con-
tain several cement genera-
tions. These cements, starting
with inclusion-rich calcites
of probable marine origin and
culminating with medium-
to coarse-grained clear equant calcite of burial origin,
form a sequence that destroyed much of the early po-
rosity (Figure 20). The residual pore space commonly
contains bitumen, either as a final fill phase or as beads
and partial linings on the equant calcite. The early
cement sequence is therefore referred to collectively as
the prebitumen sequence. Petrographic relationships
between bitumen and equant calcite (Figure 20C)
suggest that they were emplaced approximately con-
temporaneously. Equant calcite and bitumen occur
in some dissolution features as the only phases, in-
dicating a second fracturing and dissolution event
prior to bitumen emplacement. Fluid-inclusion ho-
mogenization temperatures indicate that the equant
calcite was precipitated at temperatures in the 80–958C
74 / Collins et al.

FIGURE 17. Selected core and FMI images from T-5660 (top), showing facies representing the upper wedge sequence
overlying a possible debris apron sequence in the accretionary sector. The upper wedge sequence consists of distal bedded
rudstone and fine breccia toward the base (E, F), which are overlain by successively coarser breccias (B – D). Above the
cored interval, the section consists of massive boundstone or breccia (A). The apron (G – J) consists of interbedded lam-FG,
rudstone, and breccia FMI facies (the latter presumably containing microbial boundstone clasts). For definitions of FMI
facies, see Figure 10. The cross section (bottom, see Figure 16 for well locations) shows that the inferred apron section
in T-5660 is associated with reduced radioactivity on the spectral gamma-ray (SGR) log compared to the wedge interval
and may be correlated to nearby wells on this basis. Other log curves shown include computed gamma-ray (CGR,
calculated with uranium subtracted) and deep resistivity (DRES).
FIGURE 18. Facies and stratigraphic relationships in the allochthonous rim sector. Wells are shown in approximate order of increasing distance from the edge
of the platform (see Figure 16 for well locations). The thick, allochthonous Bashkirian section at T-6337 is supported by biostratigraphic analysis of the cored
interval. Comparison of spectral gamma-ray (SGR) and computed gamma-ray (CGR, calculated with uranium subtracted) logs indicates uranium enrichment
throughout the rim and flank interval. Note the low-resistivity zone on the deep resistivity (DRES) curves in T-3938 and T-3948 at the base of the apron. Also note the
similarities in the FMI facies and gamma-ray logs in T-6337 and the interpreted apron sections in T-5660 and T-5059 from the accretionary sector (Figures 17, 19).
For FMI facies definitions, see Figure 10.
The Outer Platform, Rim, and Flank of the Tengiz Buildup / 75
 76 / Collins et al.

FIGURE 19. Stratigraphic and facies relationships in the accretionary rim sector. Wells are shown in order of increasing distance from the platform (see Figure 16
for well locations). Raised rim wells contain Serpukhovian middle-slope breccia and upper-slope boundstone overlain by upper Serpukhovian to Bashkirian outer-
and shallow-platform facies. Wells closer to the platform (T-4556 and T-7450) are associated with reduced uranium enrichment, as shown by comparing the
spectral gamma ray (SGR) with the computed gamma ray (CGR, calculated with uranium subtracted). Uranium enrichment also tends to be higher in upper
wedge boundstone breccia and associated facies in flank wells, above the apron facies interpreted in T-5059 and T-5963. FMI facies and deep-resistivity (DRES)
logs are shown for some wells. For FMI facies definitions, see Figure 10.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 77

range (Appendix 2); thus, prebitumen cementation
was completed shortly before hydrothermal tempera-
tures were achieved in the reservoir. Equant calcite is
a volumetrically important cement in the rim and
flank and is referred to as 908 calcite on the basis of the
fluid-inclusion data.
Fractures and Later Dissolution (Corrosion)
Larger fractures observed in core can be separated
into sets according to their fill phases. Earlier frac-
tures, like early dissolution voids, are extensively filled
with prebitumen cement phases. Later fractures con-
tain only bitumen, or equant calcite accompanied by
bitumen, and are filled to partly open. Systematic frac-
tures that are mainly open and lack cement are present,
representing an even later generation following bitu-
men deposition, and at least one set of fractures ap-
pears to be aligned with the present-day stress regime
(D. G. Carpenter, 2005, personal communication). Partly
open and open fractures are commonly associated
with local solution enlargement along fracture walls,
and with bitumen cementation and corrosion in the
surrounding rock matrix. Matrix corrosion occurs
as microporosity in the micritic components of the
rock, for example, skeletal grains and microbial fa-
brics (Figure 21A). In facies containing micritic ma-
trix, microporosity accompanied by small vugs can
sometimes be extensively developed. In addition to

FIGURE 20. Earlier stages of diagenesis in the Tengiz rim and flank. Voids are filled by a prebitumen cement sequence
(A and B) that includes inclusion-rich, probably marine calcite (1), prismatic to equant zoned calcite and inclusion-rich
dolomite (2), and clear blocky equant calcite (3). This sequence is commonly followed by bitumen (4). Prebitumen cements
filled much of the primary pore space, as well as porosity created during early diagenetic dissolution and fracturing.
Some voids are filled only with equant calcite and bitumen (C and D), indicating a separate, later fracturing and dissolution
event. Even later dissolution is suggested by corrosion along crystal growth planes (5) in the equant calcite. Fluid-
inclusion homogenization temperatures of 80 – 958C for the equant calcite indicate the existence of prehydrothermal
conditions in the reservoir prior to bitumen formation. Scale bars = 1 mm.
78 / Collins et al.

matrix corrosion adjacent to solution-enlarged fractures,
bitumen and corrosion are noted between cements
and walls of partly filled fractures. Bitumen and cor-
rosion are also sometimes observed within inclusion-
rich prebitumen cements or along cleavage and growth
planes of individual cement crystals.
Bitumen and matrix corrosion are both observed
to decrease away from vugs and enlarged fractures,
suggesting that bitumen, corrosion, and large-scale
dissolution may be genetically linked. Petrographic
observations indicate that significant corrosion and
solution enlargement postdates the early cement se-
quence. A key question is whether most of this occurred
before, during, or after bitumen formation or a combina-
tion of the three. Examples where bitumen-associated
microporosity and bitumen-free microporosity occur in
the same thin section are common (Figure 21B), and in
some cases, dissolution is observed that appears to post-
date bitumen formation (Figure 21C). This suggests
that some matrix corrosion probably accompanied
bitumen emplacement, and that some dissolution may
also have occurred later along the same pathways.
However, many other instances indicate mainly pas-
sive fill of open void space by bitumen, accompanied
by only minor micrite corrosion.
Other postbitumen diagenesis includes fractures,
noncarbonate cement such as fluorite and quartz (based
on fluid-inclusion temperatures; see Appendix 2), and
possibly some calcite spar (Figure 21D). The existence
of syn- and postbitumen cementation and dissolution

FIGURE 21. Bitumen-related and postbitumen diagenesis in the Tengiz rim and flank. Bitumen is dominant in some
voids and fractures (A, 4) and is commonly associated with corrosion of microbial fabrics (A, 5). Some micrite corrosion
appears to be free of bitumen (B, 6) and possibly postdates bitumen formation. Postbitumen diagenesis also includes
dissolution (selective dissolution of dolomite rhombs in a partial dedolomite, C) and late fracturing (D). The example
in (D) also shows postbitumen calcite cement. A key question in the Tengiz rim and flank is how much postbitumen
dissolution and fracturing contributes to present-day reservoir quality and distribution. Scale bars = 1 mm, except
where indicated.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 79
implies that an active hydrothermal system was prob-
ably a factor in the development of the rim and flank
reservoir. Two scenarios are proposed: hydrothermal
fluid mixing and geothermal convection. Mechanisms
for hydrothermal mixing includes fluids with differ-
ent H2S concentrations or thermochemical sulfate
reduction (TSR) reactions between hydrocarbons and
sulfate minerals. These possibilities are suggested by:
1) the elevated H2S content (13%) of hydrocarbons
in the reservoir
2) the presence of known TSR byproducts in the
reservoir
3) the thermal regime, estimated from fluid-inclu-
sion studies, under which these products were
precipitated
4) the presence near or within the reservoir of
geological conditions and ingredients favorable
for TSR reactions.
Geothermal convection is supported by two-
dimensional numerical flow models of Tengiz using
temperatures derived from basin modeling, known
thermal rock and fluid properties, and a simplified
layering scheme based on bulk reservoir permeabil-
ities. It is also possible that both convective and H2S-
related processes contributed to rim diagenesis and
reservoir quality at different times or at the same time.
Hydrothermal Fluid Mixing and
H2S-related Diagenesis
According to basin models, the Tengiz reservoir
attained hydrothermal temperatures during the early
stages of post-Triassic rapid subsidence (Figure 22).
During and after this time, basin-generated hydro-
thermal fluids may have periodically charged the res-
ervoir, mixing with formation fluids. Thin sections
and quantitative x-ray diffraction data from core plugs
reveal the presence of anhydrite, dolomite, chalcedony,
quartz, barite, fluorite, fluorapatite, and pyrite as po-
tential hydrothermal species in the rim and flank.
Most of these minerals exist in quantities less than
4% by weight. Species that may have a hydrothermal
origin and have been measured as much as 50% and
higher by weight in individual samples include fluo-
rapatite, anhydrite, and dolomite (anhydrite and do-
lomite may also have other origins). Chalcedony and
quartz are also locally abundant. Homogenization
temperatures of 110 – 1258C have been obtained from
examples of quartz and fluorite (Appendix 2).
Significant calcium carbonate dissolution can oc-
cur in reservoirs where fluids with different H2S
saturations are mixed (Hill, 1995), or when H2S and
CO2 generated by TSR mix with oxygenated forma-
tion water in a carbonate reservoir (for example, see
Hill, 1990). In the absence of oxygenated formation
waters, the amount of carbonate dissolution that
occurs in the presence of H2S and CO2 is controlled
by pH of the formation waters and the presence of
iron oxides or aqueous sulfides (Stoessell, 1992). Car-
bonate dissolution commonly occurs because of car-
bonic and/or sulfuric acid generated during these
reactions (Hill, 1987; Machel, 1987, 1989). Once H2S
saturation became established in the reservoir, either
by migration of H2S and CO2 gases or after migration
of hydrocarbons, additional dissolution may have
occurred as a result of periodic mixing with external
hydrothermal fluids during even later burial.
The high burial temperatures, elevated H2S levels,
and the presence of bitumen, dedolomite, and anhy-
drite suggest the possible involvement of TSR in dia-
genesis of the rim reservoir. In the Tengiz region, po-
tential reactants include hydrocarbons generated from
Carboniferous to Permian subsalt source rocks adjacent
to the buildups (Lisovsky et al., 1992) and an anhy-
drite layer at the base of the thick Permian halite for-
mation that encases the buildups. Figure 23 shows two
possible regions where reactions might have occurred:
1) basinal areas between platforms where hydrocar-
bons were being generated; in this scenario, car-
bonate dissolution occurred after migration of
H2S and CO2 from reaction sites into the reservoir
2) in the reservoir itself, with carbonate dissolution
resulting from reactions that occurred when
liquid hydrocarbons entered the reservoir
The latter mechanism is possible because the ab-
sence of Moscovian–Artinskian volcanics on the up-
per slope of the Tengiz flank places the anhydrite layer
in direct contact with the reservoir. It has the advan-
tage of also accounting for, as a common by-product
of TSR, at least some of the ubiquitous bitumen in the
rim and flank. Reservoir temperatures of 80 –958C
achieved prior to bitumen emplacement (based on
prebitumen equant calcite) indicate that bitumen
formation probably occurred as the reservoir entered
the hydrothermal regime.
Carbonate reservoirs exposed to TSR commonly
contain anhydrite, dedolomite, sulfide minerals, ele-
mental sulfur, and uranium and iron enrichment (Hill,
1990, 1995). In particular, anhydrite and dedolomite
of apparent nonevaporative or nonsubaerial origin
are commonly observed in cores and thin sections
80 / Collins et al.
from the Tengiz rim, and GR logs from rim and flank
penetrations typically show uranium enrichment
(see Figures 18, 19). The presence of dedolomite sug-
gests the involvement of fluids with excess Ca2+, pos-
sibly from TSR-related anhydrite dissolution.
Geothermal Convection
Groundwater circulation can be important in the
diagenetic history of carbonate reservoirs, particu-
larly for isolated platforms (Jones et al., 2004). Platform
margins are favorable sites for convective circulation
during burial, as well as during deposition. Deposi-
tional or early postdepositional convective processes
were previously briefly cited with respect to early
dissolution at the Tengiz margin. Geothermal con-
vection can occur during burial under certain condi-
tions (Machel and Anderson, 1989; Morrow, 1998;
Wendte et al., 1998; Heydari, 2000). Two-dimensional
numerical fluid-flow simulations indicate that large-
scale geothermal convection is feasible in the Tengiz
FIGURE 22. Temperature and burial-
history plots for the Tengiz rim and
flank reservoir from basin modeling.
The vertical depth range indicates the
suite of samples from different parts of
the reservoir interval used in fluid-
inclusion analyses (Appendix 2). Bitu-
men formation in the reservoir began,
probably under hydrothermal condi-
tions, after cementation by equant
calcite (top) during Mesozoic subsi-
dence. Potential times of burial fracture
generation include stresses generated
during salt diapirism, uplift (approxi-
mately 220 – 175 Ma), rapid Mesozoic
subsidence, and the present-day stress
regime (bottom).
rim and flank (G. D. Jones and Y.
Xiao, 2005, personal communica-
tion), mostly because of the high ver-
tical permeability possible in frac-
tured reservoirs and because of a
relatively large center-to-edge tem-
perature gradient in the buildup
caused by the thick, conductive en-
velope of halite around the exterior.
Once convection cells are estab-
lished in a closed system, areas of net
dissolution and net cementation can
occur while maintaining average
saturation equilibrium in the cell
with respect to the reservoir. Be-
cause of the inverse relationship at
constant pH and pCO2 between calcium carbonate
solubility and temperature, dissolution occurs where
saturated or nearly saturated fluids are cooling (rising
cell limbs), and cementation occurs where such fluids
are warming (descending limbs). However, this pro-
cess results in mass transfer of carbonate from one
area to another, and closed-system cells may, over
time, be forced to readjust in shape or shut down in
response to changing permeability patterns.
Numerical models further indicate that convec-
tion cells, if they remain active long enough, can
transfer sufficient calcium carbonate to significantly
affect porosity in the reservoir (G. D. Jones, 2005, per-
sonal communication). Several factors may have facil-
itated long-term active convection in the rim and flank
of Tengiz. The absence of Moscovian to Artinskian
sediments on the upper slope of Tengiz (Figure 23) may
have allowed continuous seawater exchange for rim
and flank convection throughout an extended period
following deposition. Second, Tengiz appears to have
FIGURE 23. Evolution of the Tengiz rim and flank high-permeability reservoir, based on data and analyses contained in the text. Prior to burial (A), the reservoir
consisted of fractures and large vugs caused by syndepositional instability and platform margin karsts. Initial large-scale dissolution and cementation patterns
may have been controlled by Kohout convection and by meteoric and mixing processes. During early burial (B), they may have been controlled by geothermal
convection, promoted by still-open fractures and by deposition of thermally conductive halite. Conductivity loss and differential compaction caused by salt
diapirism may have reduced or eliminated geothermal convection (C), although compactional stresses may have rejuvenated the fracture system. Prebitumen
cements had significantly reduced nonmatrix porosity and permeability by the end of this stage. New tectonic stresses again reactivated the fracture system,
possibly facilitating movement and distribution of hydrothermal fluids in the reservoir (D). Diagenesis related to thermochemical sulfate reduction (TSR) is
believed mainly responsible for solution enlargement of open fractures and lost circulation zones (LCZ), either from migration of CO2 and H2S gas, or as a direct
result of the liquid hydrocarbon charge. These fluid charges may also be responsible for enhanced dissolution in adjacent platform facies, forming the
transitional reservoir zone (TRZ). Note that overall fracture density, but not necessarily openness, increased in reservoir over time.
The Outer Platform, Rim, and Flank of the Tengiz Buildup / 81
82 / Collins et al.

porosity (fractures, large vugs,

and lost circulation zones) in
some facies, the rim and
flank are considered a dual-
porosity reservoir. Differ-
ent reservoir facies in the
rim and flank can be rec-
ognized based on both ma-
trix and nonmatrix pore
types.

Matrix Reservoir Types

Air permeability data
combined with thin sec-
tions from core plugs were
used to guide the recogni-
tion of matrix reservoir rock
FIGURE 24. Tengiz rim and flank paragenesis. Cementation events are indicated in types in the Tengiz rim and
black, dissolution events are indicated in white, and events that involve mainly re- flank. This data, while not
placement or recrystallization are shown in gray. Approximate absolute timing has quantitatively accurate for
been established for zoned prismatic spar, dolomite cement, and subhedral equant characterization at reservoir
calcite (150 ± 25 Ma; see Figure 22 ) from fluid-inclusion results. Based on petrographic conditions, nevertheless rep-
relationships, these cements were emplaced just prior to bitumen deposition (see
Figure 20) and the time when hydrothermal temperatures were attained in the res-
resent a relatively consistent
ervoir. All other diagenetic products are shown in relative order only, as part of either laboratory test that is useful
the prebitumen or postbitumen sequences. Other phases might be constrained by for distinguishing reservoir
processes related to burial history, which has an implied time scale (see Figures 22, 23). rock types based solely on
rock properties. The classi-
had a complex fracture history (Figure 24). Multiple fication scheme used for analysis of Tengiz plugs was
fracture events and/or repeated reactivation of pre- modified from that of Lucia (1999) and recognizes
vious fractures in the rim and flank may have main- five matrix pore-type classes (interparticle porosity,
tained or increased vertical permeability over time, microporosity, touching vugs, separate vugs, and frac-
allowing geothermal convection to remain active dur- tures) based on independent permeability behavior
ing burial. as a function of porosity. Applied to Tengiz, these
data permit the distinction of three main reservoirs
RESERVOIR QUALITY AND DISTRIBUTION based on matrix properties:

The distribution of reservoir rock types in carbon-
ate reservoirs is a product of interactions between
depositional facies and diagenetic processes. Reser-
voir quality should tend to average out somewhat
when comparing deeper well penetrations or wells
from the same environment in the field because these
wells encounter a wider range of stratigraphic ages,
facies, and potential diagenetic themes in the Tengiz
spectrum. Reservoir quality at Tengiz has so far shown
a stronger relationship to specific well locations than
to facies or stratigraphy. A partial explanation for
these geography-based reservoir-quality variations lies
in understanding the relative contributions of disso-
lution, cementation, and fracturing to matrix and non-
matrix porosity in the outer platform, rim, and flank.
Because of the significant contribution of nonmatrix
1) The rim and flank reservoir (Figure 25) encom-
passes upper-, middle-, and lower-slope facies
containing mainly microbial boundstone or
breccia (but also including some interbedded
platform-derived facies).
2) The transitional reservoir zone (Figure 26) con-
sists of shallow-platform or outer-platform facies
with significant bitumen cement and matrix
dissolution. This term is used instead of outer
platform because the reservoir type is not re-
stricted to that facies. It should not be confused
with the term ‘‘transition zone,’’ commonly used
to describe certain conditions in hydrocarbon
columns.
3) The central platform reservoir includes the
central regions of the Tengiz platform that are
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 83

FIGURE 25. General characteristics of the rim and flank reservoir. (A) Local solution enlargement of a partly cemented
fracture-breccia in microbial boundstone. Dissolution is found along the fracture walls and between cement crystals.
Bitumen-filled microporosity (darker color) extends a few centimeters from the fractures, but matrix porosity is otherwise
low. A lost circulation zone is associated with this feature. (B) Microbial boundstone with low matrix porosity and
systematic bitumen-filled hairline fractures. (C) Probable fractured zone in skeletal-lithoclast rudstone with extensive
matrix corrosion, resulting in core rubble. The dark color is caused by pervasive bitumen cementation. (D) Microbial
boundstone (clast in breccia) with bitumen and miroporosity developed along microbial filaments. (E) Microbial
boundstone (clast in breccia) with vugs formed by localized solution enlargement between microbial masses.

relatively unaffected by bitumen plugging and
associated effects (see Kenter et al., 2006).
It can be demonstrated that much of the Visean A,
Serpukhovian, and Bashkirian platform reservoir is
in pressure communication with the rim and flank.
As shown in Figure 23, the transitional reservoir zone
refers to parts of the outer regions of these platforms
that may have been altered by the incursion of vari-
ous diagenetic fluids active in the rim. For example,
the extent of this region may represent the equili-
bration limit of corrosive hydrothermal fluids that
charged the reservoir or the innermost edges of active
convection cells. The greatest distinction in matrix
properties is between the transitional reservoir zone,
affected by bitumen plugging and matrix dissolu-
tion, and the central platform reservoir (Figure 27).
Because similar depositional facies are present in both
the outer and central regions for shallow-platform cy-
cles, bitumen and dissolution are mostly responsible
for the observed differences. In some cycles, early po-
rosity and permeability in the transitional reservoir
zone may have been adversely affected by the presence
of deeper, outer-platform facies or by the presence of
microbial or marine cements (for example, see Figure 6).
Figure 27 shows a more subtle contrast between the
transitional reservoir zone and the rim and flank
reservoir. Because bitumen and dissolution affected
both of these reservoirs, the difference is more likely
caused by the grainy nature of transitional reservoir
zone facies, which probably had higher matrix
porosity prior to bitumen and dissolution, in con-
trast with rim and flank facies that are dominated by
early-cemented microbial boundstone and bound-
stone breccia.
Transitional Reservoir Zone
The porosity range for the transitional reservoir
zone determined from plug data is intermediate be-
tween the central platform range and that of the rim
and flank reservoirs (Figure 27A). The data also show
a large amount of permeability scatter around the
central platform trend (Figure 27B). The transitional
reservoir zone matrix is characterized by larger pore
sizes because of solution enlargement of early in-
terparticle and moldic porosity and because smaller
84 / Collins et al.

FIGURE 26. Characteristics of the transitional reservoir zone, which affects the outer regions of the Visean A, Serpukhovian,
and Bashkirian platforms. Facies include skeletal packstone-rudstone with brachiopods and large crinoids (D, F, J, and K),
sponge or coral boundstone (E and H), and occasional high-energy ooid grainstone (G). These examples have in common
dissolution enlargement and bitumen cementation in the matrix. Dissolution enhancement includes matrix vugs
(D, arrows) and enlarged fractures or stylolites (D, J, and K). Bitumen (darker colors in core examples) is somewhat
fabric selective in (E – H). The example in (F), interpreted as a karst feature, shows differences in relative bitumen
abundance relative to the fracture and to the clasts in the cavity. In thin section, bitumen is commonly associated with
minor corrosion (cement corrosion in B and C, arrows). Bitumen and solution enlargement have offsetting effects
on porosity and permeability, as shown in (A), in which both bitumen-cemented pores (arrows) and solution-enlarged
pores are visible.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 85

pores tend to be bitumen filled (Figure 26A, B).
Larger pores exhibit only partial fills or incomplete
linings. The reduced porosity range in the transi-
tional reservoir zone, thus, appears to be caused by
bitumen cementation of smaller pores, whereas the
permeability scatter appears to be caused by a com-
bination of bitumen plugging and solution enhance-
ment that resulted in a fundamental change in matrix
pore size, shape, and connectivity. For the same po-
rosity value or porosity range, the transitional reser-
voir zone has slightly better average matrix perme-
ability compared to the central platform reservoir.
Thus, whereas bitumen plugging appears to have re-
duced the porosity range in the transitional reservoir
zone, an expected reduction in permeability over that
range has been offset by dissolution.
Because significant matrix porosity likely remained
in the transitional reservoir zone after depositional
and early diagenetic processes, cyclic controls on
platform porosity probably influenced patterns of
later dissolution and cementation. The net effect of
porosity changes caused by bitumen plugging and
matrix dissolution overprinted across facies results
in a noticeable change in the expression of cyclicity
on porosity logs. In particular, dissolution some-
what disguises the loss of porosity observed at se-
quence boundaries in the central platform (Kenter
et al., 2006) and adds high-frequency vertical po-
rosity variations in the cycles. Log character thus
provides a means of constraining the width of the
transitional reservoir zone along the outer edge of
the platform. Based on current well and core control,
the width appears to vary vertically and may range
from less than 500 m (1640 ft) to more than 2 km
(1.2 mi).
Rim and Flank Reservoir
The data from the rim and flank reservoir in
Figure 27 indicate a further reduction in matrix po-
rosity range but permeability scatter similar to the
transitional reservoir zone. Although it was originally
anticipated that matrix reservoir quality might depend
on the amount of grainy matrix in different facies,
comparisons of matrix-rich breccias, more tightly
packed breccias containing little or no matrix, and
in-situ microbial boundstones do not reveal greatly
different porosity ranges (Figure 28). It is also sur-
prising, given that matrix dissolution features are less

FIGURE 27. Plug porosity and permeability data comparing the matrix reservoir from the central platform, transitional
reservoir zone (TRZ), and rim and flank areas. The central platform trend includes almost 5000 plugs (only a partial set
is plotted) and demonstrates a single trend despite a variety of facies, textures, and reservoir rock types represented. The
rest of the data consists of approximately 1700 plugs from the rim and flank and 700 from Bashkirian and Serpukhovian
outer-platform facies representing the TRZ. Increased permeability scatter and reduced porosity ranges characterize
both the TRZ and the rim and flank trends compared to the central platform (A). In the TRZ, reduced porosity is caused
mainly by bitumen, whereas the permeability scatter is caused by a change in pore connectivity from the combination
of bitumen plugging and solution enlargement. In the rim and flank, the porosity range is reduced further (B), because
of the presence of tightly cemented microbial boundstone facies. The permeability scatter is partly a function of solution-
enlarged pores, but also because of the increased importance of vuggy-moldic pore types and microfracture porosity.
86 / Collins et al.

FIGURE 28. Histograms from plug data comparing matrix porosity ranges in the rim and flank by facies (A) and by
well (B). Boundstones and breccias have similar distributions, rudstones have slightly better porosity (based on higher
frequency in the 3 – 5% bins), whereas fine-grained platform-derived facies have a lower porosity range. Overall, however,
porosity variability between facies is instead weak compared to the variations by well, suggesting a strong diagenetic
overprint.

abundant overall and appear to be more poorly con-
nected, that rim and flank matrix permeability is
comparable to that of the transitional reservoir zone.
This may be partly an artifact of the increased abun-
dance of small fractures, producing a greater amount
of plug damage. However, dissolution is observed
along some small fractures in plugs, so whereas the
number of plugs with high permeability may be in-
flated, the range is probably real. It should also be
remembered that these are comparative statistical
relationships that only apply where matrix porosity
exists and cannot be extrapolated to generalizations
about the overall distribution of porosity.
Porosity in upper- and middle-slope microbial
boundstones and breccias was substantially reduced
by sediment fill and prebitumen cementation of
early pore types. These include fenestrae, microbial
growth and framework cavities, and other primary
voids, as well as early fractures and dissolution voids.
In middle- to lower-slope and apron facies, it was
expected that the presence of skeletal rudstones or
floatbreccias with grainy matrix might improve
reservoir quality in some areas because of possible
preservation of primary porosity, compared to areas
containing tightly packed breccias or breccias with
finer grained or micritic matrix. However, like the
upper slope, matrix reservoir quality in most middle-
to lower-slope facies is affected by bitumen, corro-
sion, and prebitumen cement, in particular syntaxial
calcite (of probable early to moderate burial origin)
and clear equant calcite (including 908 calcite of
burial origin). Early marine cements prevalent in the
upper-slope facies are less noticeable. Burial cemen-
tation and dissolution processes may be part of the
reason why there is no good evidence that breccias,
in general, have better reservoir quality compared to
other facies, or that better reservoir quality is asso-
ciated with a particular type of breccia.
Matrix and Nonmatrix Reservoir Distribution
Nonmatrix features consisting of fractures, large
vugs, and lost circulation zones are an important
component of the rim and flank reservoirs. These
elements are absent or greatly reduced by compar-
ison in the central platform area. The distribution of
these features and their relationships to matrix prop-
erties are the basis of rim and flank reservoir char-
acterization. As described previously, open fractures,
vugs, and lost circulation zones appear to be linked
to areas of bitumen cementation and solution en-
largement (see Figure 25). Whereas bitumen is asso-
ciated with dissolution along fractures and in the
surrounding matrix, it is also responsible for plug-
ging up some fractures and vugs. Furthermore, bitu-
men and dissolution affect matrix reservoir quality to
varying degrees in fractured intervals, and matrix dis-
solution and bitumen can sometimes be more exten-
sively developed in areas of lower fracture density.
Understanding these variations in spatial and genetic
relationships is a key to improved reservoir-quality
prediction.
Bitumen-associated matrix dissolution includes
microporosity, micrite corrosion, vugs, and enlarged
primary voids. In detrital middle- to lower-slope and
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 87
apron breccias, residual matrix porosity controlled
by primary texture and prebitumen cementation
appears to have regulated the extent to which matrix
corrosion is developed locally. Incompletely cemented
grainy facies lend themselves to more extensive cor-
rosion, and small-scale variations in these parameters
sometimes result in apparent bedding or facies control,
obscuring a relationship to larger scale fracture den-
sity. In upper-slope boundstones and breccias, mi-
croporosity occurs mainly in primary microbial fab-
rics and entrained micritic skeletal grains. Because
little or no matrix is present in these rocks, a general
relationship to open and partly open fractures is more
clearly evident. The amount of matrix corrosion and
dissolution that develops depends on the type of
boundstone (which, in turn, determines the density
of entrained grains and microbial fabrics) and on the
amount of prebitumen cement (commonly abundant
because of high volumes of marine cement). Whereas
this combination can produce a complex-looking dis-
tribution of corrosion and bitumen relative to primary
textures in boundstones and breccias, a clear spatial
connection exists between the matrix and nonmatrix
components of bitumen and associated dissolution,
and variations in matrix components are determined
more by prebitumen cementation patterns than by
fracture density or facies.
Early Karst Features
Early karst features are commonly cited as impor-
tant influences on porosity and permeability distri-
bution in carbonate reservoirs. The overall effect may
be positive or negative (Esteban and Wilson, 1993;
Wagner et al., 1995). Early karst features can vary
considerably in spatial distribution and relationship
to stratigraphy (for example, Choquette and James,
1987; Lucia, 1995; Mylroie and Carew, 1995), al-
though for isolated platforms, dissolution is com-
monly concentrated near the platform margin (for
example, Cander, 1995; Tinker et al., 1995). Early
karsts may also influence later diagenesis, serving as
targets for selective re-excavation or forming signifi-
cant impediments to later fluid flow (Ford, 1995).
Large-scale features interpreted as early karst are
present at Tengiz; however, observed relationships
between these features, the distribution of later dia-
genetic features (bitumen cement and corrosion),
and overall porosity and permeability suggest that
early karsts are largely incidental to reservoir quality.
Potential early karst features include local subaer-
ial exposure effects associated with sequence bound-
aries, brecciated intervals, and large, sediment-filled
cavities (Figure 29). Almost all of these features are
completely filled with sediment and, to a lesser ex-
tent, cement. Later bitumen and corrosive diagenesis
are present in these intervals and, in some cases, are
even concentrated within or around the karst or dis-
solution features (Figure 29A). However, when com-
pared to randomly selected cored intervals or to cored
intervals of equal thickness where similar features are
absent, the intervals containing the karst features
tend to have, on average, diminished amounts of bi-
tumen and dissolution, and average reservoir quality
in these zones does not appear to be consistently
better or worse than elsewhere.
Lost Circulation Zones
Lost circulation during drilling in the rim and flank
can be shown to occur in association with increased
open-fracture density or with solution enlargement
of fractures in instances where FMI control, or in
some cases, core control, exists. Recent drilling with-
out catastrophic losses and with more complete core
recoveries in intervals dominated by vuggy porosity
further suggests that perhaps the distribution of frac-
tures or solution-enlarged fractures are the most sig-
nificant controls on lost circulation in general. Because
solution-enlarged fractures are commonly associated
with areas of dissolution, corrosion, and bitumen, an
important question is whether fracture distribution
was the primary control on where solution enlarge-
ment and lost circulation occurred, or whether the
distribution of corrosion and dissolution, constrained
by other factors, determined where fracture enlarge-
ment and lost circulation occurred. Ultimately, the
distribution of lost circulation zones in the rim and
flank may indicate specific pathways through the
reservoir taken by corrosive fluids or map out areas of
elevated fracture density.
As previously outlined, different fracture sets in
the Tengiz rim and flank can be distinguished by
orientation relative to core, by types of fill, and by
the amount of dissolution. Possible origins are varied,
including syndepositional instability, burial stresses
created by differential compaction of the rim and
platform areas (and, to a lesser extent, the rim and
flank areas), differential stresses resulting from over-
burden loading during salt diapirism, and stresses
associated with various tectonic events affecting the
Precaspian Basin (D. G. Carpenter, 2005, personal com-
munication). Fracture density controlled by the me-
chanical properties of different facies is one possible
influence on the current distribution of lost circulation
zones. Both core and FMI data show that microbial
88 / Collins et al.

FIGURE 29. Core examples with sediment-filled features of possible karst origin. (A) Cavities filled with laminated,
greenish-colored shale and carbonate formed in outer-platform crinoid skeletal packstone-rudstone. Cavity walls (yellow
dashed lines) are surrounded by a zone of matrix corrosion and bitumen (white lines), but matrix reservoir quality overall
is marginal. (B) Cavity developed in outer-platform to upper-slope boundstone filled with laminated green-gray marl and
boundstone breccia. Because of the unique setting of the T-4635 well (see Figure 15A), this feature could be either a karst
feature or a cavity that opened during rim detachment and subsequently filled in a deeper water setting. (C – E) Examples
of smaller, geopetally filled fenestral cavities. The sediment fill sequence, which includes green and red clay (C and D),
banded (marine) cement and cream-colored micrite (D), and dark-colored sediment (D and E), suggests a complex history
of cavity formation and fill. All examples are associated with relatively poor matrix reservoir quality, however.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 89
boundstones and tightly packed breccias have a
greater fracture density compared to rudstones, pack-
stones, and matrix-supported breccias, suggesting
that lost circulation zones might be more common
in upper- to middle-slope facies. However, based on
well occurrences, it can be demonstrated that lost
circulation zones are more common in the upper half
to one-third and slightly more common near the base
of the rim and flank section, regardless of position in
the field (Figure 30). Such a distribution is inconsis-
tent with a primary facies control on lost circulation
and, like many other observations, points to a dia-
genetic overprint.
Current well spacing is still insufficient to clearly
demonstrate geographical distribution patterns for
lost circulation zones, a situation made even worse
because most rim and flank wells drilled to date
have encountered lost circulation somewhere in the
reservoir section. More generally, facies in which
fracture permeability was dominant compared to
residual matrix permeability may be more prone to
lost circulation because later dissolution was con-
centrated along fractures. Facies with higher residual
matrix permeability (for example, poorly cemented
FIGURE 30. Distribution of lost circu-
lation zones (LCZ) in the rim and flank.
Most wells in the rim and flank have
lost circulation somewhere in the
reservoir section. The charts represent
48 LCZ in 13 wells (T-463, T-4346,
T-4635, T-4556, T-5056, T-5059, T-5435,
T-5454, T-5857, T-6261, T-6457,
T-7252, and T-7453) that have been
quantified by sustained loss rate,
depth, and thickness. The distribution
by distance in meters from the top
of the section does not show a clear
trend (left), but when the distribution
is normalized to indicate relative
proportion in the rim interval regard-
less of well position on the rim
profile (right), the pattern supports
anecdotal evidence that LCZ are
more common in the upper part of
the section, regardless of facies pres-
ent there.
grainy facies and loosely packed
breccias) may have allowed corro-
sive fluids to spread away from
fractures and, thus, may ultimately
prove less prone to lost circulation.
This may be a reason why, in the
transitional reservoir zone where per-
meable grainy facies are most abun-
dant, a relationship between lost circulation zones,
bitumen, and corrosion is less obvious. Combined
with reduced fracture density in the transitional res-
ervoir zone, this factor may also be why lost circu-
lation zones are less frequent in the outer platform
region. Whatever diagenetic processes prove ultimately
responsible for dissolution and lost circulation, their
effects have produced a reservoir with large-scale in-
terwell connectivity across the transitional reservoir
zone, rim, and flank, despite a major stratal surface
separating them and a complex facies distribution
within each.
SUMMARY AND CONCLUSIONS
Based on currently available data, the Tengiz rim
and flank are interpreted to have accumulated in
two stages. The early stage was dominated by insta-
bility and relatively continuous large-scale failure and
mass wasting, resulting in deposition of debris aprons
around the base of the buildup consisting of coarse
boundstone breccias and grainy platform-derived
sediments. In the second stage, a wedge consisting
of shallow- and outer-platform facies, upper-slope
90 / Collins et al.
microbial boundstone, middle-slope boundstone brec-
cia, and lower-slope breccia and platform-derived
sediment prograded out over the early debris apron.
This stage was more successful along the eastern
margin of the buildup because of favorable platform
profile geometry, or elevated microbial productivity,
or reduced tectonic influence in those areas, or some
combination of those factors. Along the western mar-
gin of the buildup, progradation was only locally or
partially successful, and rim failure predominated
throughout the Serpukhovian and Bashkirian.
This study does not provide a process-based, fully
predictive model of the distribution of reservoir prop-
erties in the Tengiz outer platform, rim, and flank. It
raises as many questions as it answers, and some
critical uncertainties remain. Some conclusions have
alternatives and will likely require modification or
change as drilling continues. Future surprises are
probable, and subsequent generations of models will
no doubt be different from the present one in some
manner. However, the following observations indi-
cate factors that are likely to have a significant impact
on reservoir quality and prediction in the rim and
flank:
1) Reservoir quality is dominantly controlled by frac-
tures and large-scale dissolution and is a function
of earlier porosity destruction because of prebitu-
men cementation versus later dissolution associ-
ated with bitumen formation.
2) Porosity destruction caused by sediment infill
and prebitumen cementation was significant and
affected primary voids, early dissolution voids,
early fractures, and early karst features. In par-
ticular, early fractures and large-scale dissolution
features or karsts are largely filled and do not
have a major impact on reservoir quality.
3) Corrosive dissolution and bitumen are a late-
diagenetic overprint and are responsible for sig-
nificant porosity modification in a wide variety
of features including later fractures, microbial
fabrics, micritic skeletal grains, breccia clasts, cer-
tain cements, and stylolites.
4) Facies control of fracture density results in both
open and closed types being more common in
mechanically rigid facies, such as microbial bound-
stone, cemented grainstone, and clast-supported
breccia, and less abundant in facies such as rud-
stone, packstone, and matrix-supported breccia.
5) Lost circulation zones appear to be more com-
monly associated with solution-enlarged fractures
than with vuggy intervals and are therefore likely
controlled by either fracture distribution or the
distribution of late dissolution fairways.
6) Enhanced matrix porosity caused by late disso-
lution and bitumen cement is commonly better
developed where solution-enlarged fractures are
present; however, extensive matrix corrosion
may be associated with intervals that are less
likely to produce severe lost circulation.
7) Reservoir-quality variability exists at large and
small scales, resulting in considerable well-to-well
differences in regions that are similar in terms of
facies, stratigraphy, and depositional history. This
suggests that facies is a weak secondary control on
reservoir quality.
ACKNOWLEDGMENTS
This study involved the work of a large group of
geoscientists and specialists. The authors recognize
the following individuals for insightful contributions
to this manuscript: Kevin Nahm (Chevron), who
believed early on that mechanically isotropic micro-
bial boundstones could be mapped with seismic
analysis; Brent Francis (ExxonMobil), who progressed
this idea through development of the compaction
model and who provided seismic interpretations for
the debris apron concept; and Tom Heidrick (retired
from Chevron), who engaged in valuable discussions
regarding the minimal effects of early karst and the
potential for later burial fractures. Wayne Narr
(Chevron), Jim DeGraff (ExxonMobil), and Dan
Carpenter (ExxonMobil) are acknowledged for build-
ing on this perception through rigorous description
and analyses of fracture characteristics and distribu-
tion. Perhaps most importantly, Michael Clark (Chev-
ron) is thanked for initiating an intense coring and
core analysis program that formed the basis for most
of our geologic studies and for his strong support of the
sequence-stratigraphic framework.
We also gratefully thank the following individuals
from ExxonMobil who made significant contributions
of supporting data or data analysis: Gareth Jones and
Yitian Xiao (numerical simulation of Tengiz diagenet-
ic processes); Sean Guidry (fluid-inclusion analyses);
Paul Hicks (temperature histories from basin model-
ing); Shin-Ju Ye and Peter Hillock (facies analysis of
FMI logs). Early versions of the manuscript were
improved by helpful reviews from Niall Toomey,
Kelly Bergman, and Gareth Jones. Finally, we thank
TengizChevroil and its shareholder companies for
permission to publish this study.
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 91

APPENDIX 1: RIM AND FLANK FACIES (MODIFIED FROM WEBER ET AL., 2003)

Lithofacies and Description Environment

Microfacies
Crinoid-brachiopod Medium-dark gray; large crinoids and/or Deep subtidal outer platform, possibly
skeletal rudstone fragmented thick-shelled brachiopods below wave base; abundant in central
to packstone dominate; sometimes with microbial to upper parts of late Visean and
fabrics and cements Serpukhovian outer-platform cycles
Outer platform breccias Grainstone floatbreccia to packbreccia: Deep outer platform, possibly lowstand;
and conglomerates light-medium brown; subangular to basal parts of Serpukhovian outer-platform
subrounded clasts of platform-derived cycles
grainstone and packstone; skeletal
wackestone to grainstone matrix,
sometimes with microbial fabrics and
cements
Grainstone packbreccia: light brown, gray, Shallow outer platform; lag or possibly
and black; angular clasts of grainstone and karst? Tops of Serpukhovian outer-platform
rudstone with mixed lithofacies types cycles
Coral rudstone to conglomerate: medium Shallow outer-platform; basal parts of
brown to gray; colonial coral fragments; Serpukhovian outer-platform cycles;
coarse-grained skeletal matrix with possible patch-reef detritus
abundant crinoids
Brachiopod rudstone Light gray-brown to medium brown-gray; Shallow outer-platform; basal parts of
(coquina) to boundstone brachiopod shell fragments and whole Serpukhovian outer-platform cycles;
shells; some articulated, in feeding or mud-mounds, mound detritus, and
growth position; abundant micritic matrix, intermound deposits
some microbial fabrics and calcite cement;
minor coarse skeletal debris, crinoids
Microbial boundstone Type C (skeletal-micritic boundstone): Serpukhovian deep, subtidal, outer-platform
light-medium brown, massive to peloidal to upper-slope platform break
with abundant fenestrae; crinoids,
brachiopods, fenestellid bryozoans, algae,
forams, and other skeletal grains bound
together by clotted microbial fabric and
multiple cement generations (as much as
50% cement)
Type B (laminar-cemented boundstone): Serpukhovian upper slope
light brown; irregular concentric or wavy-
aminated fabric consisting of microbial
filaments and early cement (botryoidal,
fibrous and radiaxial; inclusion rich; as
much as 75%); minor skeletal grains,
including fenestellid bryozoa, crinoids,
ostracods, thin-walled pelecypods and
forams
Type A (peloidal-micritic boundstone): Serpukhovian upper to middle slope
light gray-light brown; generally massive
or featureless, with minor growth structures
and abundant irregular fenestrae or tubules;
clotted peloidal fabric and early cement
(as much as 25%); sparse fauna, including
fenestellid bryozoans, ostracods, and thin-
walled pelecypods
92 / Collins et al.

APPENDIX 1: (CONT.)

Lithofacies and Description Environment

Microfacies
Boundstone breccia Welded to mosaic breccia: light-medium Serpukhovian upper to middle slope and
brown and gray; subangular centimeter- to debris apron
meter-size clasts of upper-slope boundstone
types; tight, clast-supported fabric, sometimes
with stylolitic clast boundaries (welded
breccia), sometimes with blocky calcite
cement (mosaic breccia); minor platform-
derived matrix (skeletal grains, large crinoids);
occasional platform-derived grainstone or
packstone lithoclasts
Float- to packbreccia: light-dark brown and Serpukhovian middle to lower slope and
gray; subangular to subrounded, millimeter- debris apron
to meter-size clasts of upper slope-derived
boundstone types; grainstone-packstone
matrix with platform-derived skeletal grains;
large crinoids and fragmented brachiopod
shells; occasional platform-derived grainstone
or packstone lithoclasts
Crinoid-lithoclast Brown-light gray; poorly to moderately Serpukhovian middle to lower slope and
rudstone-packstone sorted, crinoid-rich rudstone (platform- debris apron
derived); fragmented brachiopod shells;
sand-size skeletal grains and millimeter- to
centimeter-size lithoclasts (upper slope and
platform derived)
Skeletal-lithoclastic Medium-dark gray; thin to thick bedded, Serpukhovian lower slope and debris apron
rudstone-grainstone- wavy-bedded to laminated, graded; locally
floatstone siliceous to slightly dolomitic; platform-
derived grains (forams, crinoids, algae,
brachiopods, crinoids); occasional
millimeter- to centimeter-size boundstone
or slope-derived clasts, volcanic siltstone
or sandstone clasts
Allochthonous grainstone Light brown-medium gray; moderately Lower slope and debris apron; Serpukhovian
sorted ooid or skeletal grainstone; massive and Bashkirian
to well bedded
Lime wackestone- Brown, gray, and black; argillaceous, Lower slope, debris apron and basinal;
mudstone siliceous, and commonly dolomitic; thin Visean, Serpukhovian, and Bashkirian
laminated to thin bedded; sponge spicules,
minor skeletal grains, and sparse
radiolarians; commonly interbedded with
volcanic silt and sand; potential source rock
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 93

APPENDIX 2: FLUID-INCLUSION RESULTS

Well Depth (m) Platform Sequence Phase Type # incl. Thomog (Water) (8C)
T-4635 4466.25 Serpukhovian rim Equant CC Primary 2 85 – 90
T-4635 4466.25 Serpukhovian rim Equant CC Primary 2 90 – 95
T-4635 4482.63 Serpukhovian rim Equant CC Primary 1 88
T-4635 4482.63 Serpukhovian rim Equant CC Primary 3 80 – 85
T-4635 4482.63 Serpukhovian rim Equant CC Primary 2 85 – 90
T-4635 5055.25 Famennian Calcite Unspecified 5 105 – 110
T-4635 5055.25 Famennian Calcite Unspecified 6 105 – 115
T-4635 5055.25 Famennian Calcite Unspecified >10 110 – 120
T-4635 5055.25 Famennian Calcite Unspecified 4 90 – 95
T-47 5392.5 Famennian Quartz vein Unspecified No assemblage 115 – 125
T-5056 4194.2 Serpukhovian rim Equant CC Primary 1 80 – 85
T-5056 4194.2 Serpukhovian rim Equant CC Primary 5 85 – 90
T-5056 4200.47 Serpukhovian rim Fluorite Primary 4 110 – 125
T-5056 4200.47 Serpukhovian rim Fluorite Primary 4 110 – 125
T-5056 5453.55 Famennian Zoned CC Primary 2 100
T-5056 5453.55 Famennian Zoned CC Primary 3 95 – 100
T-5056 5453.55 Famennian Equant CC Unspecified 5 85 – 95
T-5056 5476.22 Famennian Calcite Primary 4 95 – 105
T-5857 4816.93 Serpukhovian rim Prismatic CC Primary 2 80 – 85
T-5857 4816.93 Serpukhovian rim Prismatic CC Primary 3 85 – 90
T-6337 4894.37 Visean D Calcite Primary 2 85 – 90
T-6337 4945.64 Visean D Equant CC Primary 4 85 – 90
T-6337 4945.64 Visean D Equant CC Primary 4 85 – 90
T-6337 4945.64 Visean D Equant CC Primary 3 85 – 90
T-8 3972 Bashkirian Coarse CC Unspecified 4 100 – 107
T-8 4044.5 Serpukhovian Zoned CC Primary 4 80 – 95
T-8 4044.5 Serpukhovian Equant CC Primary 2 85 – 95
T-8 4056.8 Serpukhovian Equant CC Primary 2 85 – 90

Data for hydrocarbon-water or vapor-filled inclusions:

Phase refers to cement types described in the text (see Figure 24). ‘‘Calcite’’ indicates the cement phase preceding bitumen formation and, in most
cases, refers to equant calcite. Prismatic calcite represents the cores of zoned cement crystals (see Figure 20), and zoned calcite represents subsequent
inclusion-rich growth zones (either calcite or dolomite). Coarse calcite is a postbitumen cement phase. An inclusion type of ‘‘unspecified’’ indicates an
assemblage or train whose orientation was not specified during analysis, but was not designated as cutting across growth planes and may be
primary. The column ‘‘# incl.’’ indicates the number of individual inclusions in a train or assemblage.

REFERENCES CITED platform, Kazakhstan: Revista Italiana di Paleontologia e

Bahamonde, J. R., C. Vera, and J. R. Colmenero, 2000, A
steep-fronted Carboniferous carbonate platform:
Clinoformal geometry and lithofacies (Picos de Euro-
pa, NW Spain): Sedimentology, v. 47, p. 645 – 664.
Blendinger, W., 2001, Triassic carbonate buildup flanks in
the Dolomites, northern Italy: Breccias, boulder fabric
and the importance of early diagenesis: Sedimentol-
ogy, v. 48, no. 5, p. 919 – 933.
Brenckle, P. L., and N. V. Milkina, 2003, Foraminiferal
timing of carbonate deposition on the Late Devonian
(Famennian)–middle Pennsylvanian (Bashkirian) Tengiz
Stratigrafia, v. 109, no. 2, p. 131– 158.
Cander, H., 1995, Interplay of water-rock interaction ef-
ficiency, unconformities, and fluid flow in a carbon-
ate aquifer: Floridan aquifer system, in D. A. Budd,
A. H. Saller, and P. M. Harris, eds., Unconformities
and porosity in carbonate strata: AAPG Memoir 63,
p. 103 – 124.
Choquette, P. W., and N. P. James, 1987, Introduction, in
N. P. James and P. W. Choquette, eds., Paleokarst:
New York, Springer-Verlag, p. 1 – 21.
Della Porta, G., J. A. M. Kenter, J. R. Bahamonde, A.
Immenhauser, and E. Villa, 2003, Microbial boundstone
94 / Collins et al.
dominated carbonate slope (Upper Carboniferous, N
Spain): Microfacies, lithofacies distribution and stratal
geometry: Facies, v. 49, p. 175–208.
Esteban, M., and J. L. Wilson, 1993, Introduction to karst
systems and paleokarst reservoirs, in R. D. Fritz, J. L.
Wilson, and D. A. Yurewicz, eds., Paleokarst related
hydrocarbon reservoirs: SEPM Core Worshop 18, p. 1 –
9.
Ford, D. C., 1995, Paleokarsts as a target for modern
karstification: Carbonates and Evaporites, v. 10, no. 2,
p. 138 – 147.
Heydari, E., 2000, Porosity loss, fluid flow and mass
transfer in limestone reservoirs: Application to the
Upper Jurassic Smackover Formation, Mississippi:
AAPG Bulletin, v. 84, no. 1, p. 100 – 118.
Hill, C. A., 1987, Geology of Carlsbad Cavern and other
caves in the Guadalupe Mountains: New Mexico
Bureau of Mines and Mineral Resources, New Mexico
and Texas Bulletin 117, 150 p.
Hill, C. A., 1990, Sulfuric acid speleogenesis of Carlsbad
Cavern and its relationship to hydrocarbons, Dela-
ware basin, New Mexico and Texas: AAPG Bulletin,
v. 74, no. 11, p. 1685 – 1694.
Hill, C. A., 1995, H2S-related porosity and sulfuric acid oil-
field karst, in D. A. Budd, A. H. Saller, and P. M. Harris,
eds., Unconformities and porosity in carbonate strata:
AAPG Memoir 63, p. 301 – 313.
Jones, G. D., F. F. Whitaker, P. L. Smart, and W. E. Sanford,
2004, Numerical analysis of seawater circulation in
carbonate platforms: II. The dynamic interaction
between geothermal and brine reflux circulation:
American Journal of Science, v. 304, p. 250 – 284.
Kenter, J. A. M., P. M. Harris, and G. Della Porta, 2005,
Steep microbial boundstone-dominated platform
margins — Examples and implications: Sedimentary
Geology, v. 178, p. 5 – 30.
Kenter, J. A. M., P. M. Harris, J. F. Collins, L. J. Weber, G.
Kuanysheva, and D. J. Fischer, 2006, Late Visean to Bash-
kirian platform cyclicity in the central Tengiz buildup,
Precaspian Basin, Kazakhstan: Depositional evolution
and reservoir development, in P. M. Harris and L. J.
Weber, eds., Giant hydrocarbon reservoirs of the
world: From rocks to reservoir chacracterization and
modeling: AAPG Memoir 88/SEPM Special Publica-
tion, p. 7 – 54.
Kerans, C., and P. M. Harris, 1993, Outer shelf and shelf
crest, in D. G. Bebout and C. Kerans, eds., Guide to the
Permian reef geology trail, McKittrick Canyon, Guadalupe
Mountains National Park, west Texas, Guidebook 26:
Austin, Bureau of Economic Geology, p. 32 – 43.
Kerans, C., N. F. Hurley, and P. E. Playford, 1986, Marine
diagenesis in Devonian reef complexes of the Can-
ning basin, Western Australia, in J. H. Schroeder and
B. H. Purser, eds., Reef diagenesis: Berlin, Springer-
Verlag, p. 357 – 380.
Kirkland, B. L., J. A. D. Dickson, R. A. Wood, and L. S. Land,
1998, Microbialite and microstratigraphy: Encrusta-
tions in the middle and upper Capitan Formation,
Guadaupe Mountains, Texas and New Mexico, U.S.A.:
Journal of Sedimentary Research, v. 68, p. 956 – 969.
Kohout, F. A., H. R. Henry, and J. E. Banks, 1977,
Hydrology related to geothermal conditions of the
Floridan Plateau, in K. L. Smith and G. M. Griffin,
eds., The geothermal nature of the Floridan Plateau:
Florida Department of Natural Resources Bureau
Geology Special Publication 21, p. 1 – 34.
Lisovsky, N. N., G. N. Gogonenkov, and Y. A. Petzoukha,
1992, The Tengiz oil field in the Pre-Caspian Basin of
Kazakhstan (former USSR) — Supergiant of the 1980s,
in M. T. Halbouty, ed., Giant oil and gas fields of the
decade 1978 – 1988: AAPG Memoir 54, p. 101 – 122.
Lucia, F. J., 1995, Lower Paleozoic cavern development,
collapse, and dolomitization, Franklin Mountains, El
Paso, Texas, in D. A. Budd, A. H. Saller, and P. M.
Harris, eds., Unconformities and porosity in carbon-
ate strata: AAPG Memoir 63, p. 279 – 300.
Lucia, F. J., 1999, Carbonate reservoir characterization:
New York, Springer-Verlag, 222 p.
Machel, H. G., 1987, Some aspects of diagenetic sulphate-
hydrocarbon redox reactions, in J. D. Marshall, ed.,
Diagenesis of sedimentary sequences: Geological
Society (London) Special Publication 36, p. 15 – 38.
Machel, H. G., 1989, Relationships between sulphate
reduction and oxidation of organic compounds to
carbonate diagenesis, hydrocarbon accumulations,
salt domes, and metal sulphide deposits: Carbonates
and Evaporites, v. 4, no. 2, p. 137 – 151.
Machel, H. G., and J. H. Anderson, 1989, Pervasive
subsurface dolomitization of the Nisku Formation in
central Alberta: Journal of Sedimentary Petrology,
v. 59, no. 6, p. 891 – 911.
Melim, L. A., and P. A. Scholle, 1995, The forereef facies of
the Permian Capitan Formation: The role of supply
versus sea-level changes: Journal of Sedimentary Re-
search, v. B65, no. 1, p. 107 – 118.
Morrow, D. W., 1998, Regional subsurface dolomitization:
Models and constraints: Geoscience Canada, v. 25,
p. 57 – 70.
Mylroie, J. E., and J. L. Carew, 1995, Karst development on
carbonate islands, in D. A. Budd, A. H. Saller, and P. M.
Harris, eds., Unconformities and porosity in carbon-
ate strata: AAPG Memoir 63, p. 55 – 76.
Playford, P. E., 1984, Platform-margin and marginal-slope
relationships in Devonian reef complexes of the
Canning basin, in P. G. Purcell, ed., The Canning
basin, Western Australia: Geological Society of Aus-
tralia and Petroleum Exploration Society of Australia,
Canning basin Symposium, Perth 1984, p. 189 – 234.
Ross, C. A., and J. R. P. Ross, 1987, Late Paleozoic sea levels
and depositional sequences: Cushman Foundation for
Foraminiferal Research Special Publication 24, p. 137 –
149.
Ross, C. A., and J. R. P. Ross, 1988, Late Paleozoic
transgressive-regressive deposition, in C. K. Wilgus,
B. S. Hastings, C. Kendall, H. W. Posamentier, C. A.
Ross, and J. C. Van Wagoner, eds., Sea-level changes:
 The Outer Platform, Rim, and Flank of the Tengiz Buildup / 95
An integrated approach: SEPM Special Publication 42,
p. 227 – 247.
Saller, A. H., 1984, Petrologic and geochemical constraints
on the origin of subsurface dolomite, Enewetak Atoll:
An example of dolomitization by normal sea water:
Geology, v. 12, p. 217 – 220.
Sanford, W. E., F. F. Whitaker, P. L. Smart, and G. D. Jones,
1998, Numercial analysis of seawater circulation in
carbonate platforms: I. Geothermal circulation: Amer-
ican Journal of Science, v. 298, p. 801 – 828.
Stephens, N. P., and D. Y. Sumner, 2003, Famennian mi-
crobial reef facies, Napier and Oscar Ranges, Canning
basin, Western Australia: Sedimentology, v. 50, p. 1283–
1302.
Stoessell, R. K., 1992, Effects of sulfate reduction on
CaCO3 dissolution and precipitation in mixing-zone
fluids: Journal of Sedimentary Petrology, v. 62, no. 5,
p. 873 – 880.
Tinker, S., 1998, Shelf-to-basin facies distributions and
sequence stratigraphy of a steep-rimmed carbonate
margin: Capitan depositional system, McKittrick
Canyon, New Mexico and Texas: Journal of Sedimen-
tary Research, v. 68, p. 1146 – 1174.
Tinker, S. W., J. R. Ehrets, and M. D. Brondos, 1995, Multiple
karst events related to stratigraphic cyclicity: San Andres
Formation, Yates field, West Texas, in D. A. Budd, A. H.
Saller, and P. M. Harris, eds., Unconformities and poros-
ity in carbonate strata: AAPG Memoir 63, p. 213– 237.
Wagner, P. D., D. R. Tasker, and G. P. Wahlman, 1995,
Reservoir degradation and compartmentalization be-
low subaerial unconformities: Limestone examples
from west Texas, China, and Oman, in D. A. Budd,
A. H. Saller, and P. M. Harris, eds., Unconformities
and porosity in carbonate strata: AAPG Memoir 63,
p. 301 – 313.
Weber, L. J., B. P. Francis, P. M. Harris, and M. Clark, 2003,
Stratigraphy, lithofacies, and reservoir distribution,
Tengiz field, Kazakhstan, in W. M. Ahr, P. M. Harris,
W. A. Morgan, and I. D. Somerville, eds., Permo-
Carboniferous carbonate platforms and reefs: SEPM
Special Publication 78 and AAPG Memoir 83, p. 351 –
394.
Wendte, J., H. Qing, J. Dravis, S. L. O. Moore, L. D. Stasiuk,
and G. Ward, 1998, High temperature saline (thermo-
flux) dolomitization of Swan Hills platform and bank
carbonates, Wild River area, west-central Alberta: Bul-
letin of Canadian Petroleum Geology, v. 46, p. 210 –
266.
 3
Lindsay, R. F., D. L. Cantrell, G. W. Hughes, T. H. Keith, H. W. Mueller III,
and S. D. Russell, 2006, Ghawar Arab-D reservoir: Widespread porosity
in shoaling-upward carbonate cycles, Saudi Arabia, in P. M. Harris
and L. J. Weber, eds., Giant hydrocarbon reservoirs of the world: From
rocks to reservoir characterization and modeling: AAPG Memoir 88/
SEPM Special Publication, p. 97 – 137.

Ghawar Arab-D Reservoir:

Widespread Porosity in
Shoaling-upward Carbonate
Cycles, Saudi Arabia
Robert F. Lindsay, Dave L. Cantrell, Geraint W. Hughes, Thomas H. Keith,
Harry W. Mueller III, and S. Duffy Russell
Saudi Aramco, Dhahran, Saudi Arabia

ABSTRACT

G
hawar field is the world’s largest, most prolific field, producing 30–318 API oil
from the Arab-D carbonate reservoir. The field is more than 250 km (155 mi)
long, as much as 30 km (18.5 mi) wide, and has more than 300 m (1000 ft) of
structural closure. The Arab-D reservoir, limestone with some dolostone horizons,
stratigraphically comprises the D member of the Arab Formation and the upper part of
the Jubaila Formation. Based on ammonite and benthonic foraminiferal evidence, the
reservoir formations are Upper Jurassic, Kimmeridgian, in age. The reservoir has an
average thickness of more than 60 m (200 ft), an average porosity of more than 15%,
and a permeability up to several darcys. The upper half of the reservoir is dominated by
exceptionally high reservoir quality; the lower half contains interbeds of high and
relatively lower reservoir quality. Early correlation of well logs and cores, before the
advent of sequence stratigraphy, subdivided the reservoir from top to base into zones 1,
2, 3, and 4. Zones 2 and 3 have been subsequently subdivided into zones 2a and 2b and
zones 3a and 3b, with a more detailed zonation scheme for reservoir management.
The reservoir is composed of two composite sequences. One composite se-
quence is the Arab-D Member of the Arab Formation, with the upper boundary at
the top of Arab-D carbonate and below the C-D evaporite, with the sequence
boundary locally marked by pods of collapse breccia. The second composite
sequence forms the upper part of the Jubaila Formation, for which the sequence
boundary between the Arab-D Member and Jubaila Formation is located in zone
2b and is marked by a flood of slightly deeper water cycles over the more grain-
dominated cycles in upper zone 3a and lower zone 2b. Several high-frequency
sequences (HFSs) have been identified, each comprising several cycle sets (parase-
quence sets). Each cycle set is composed of approximately five individual car-
bonate cycles (parasequences), and each cycle is composed of one to three beds.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215875M88576

97
98 / Lindsay et al.

These carbonates were deposited approximately 58 south of the equator on a

broad, arid, storm-dominated carbonate ramp. From upslope to downslope, the
ramp consisted of the following subenvironments: (1) inner ramp; (2) ramp-crest
shoal; (3) proximal middle ramp; (4) distal middle ramp; and (5) outer ramp. The
inner ramp was a lagoon with localized intertidal islands composed of grainstones
and packstones and a highly diverse, shallow-marine benthonic foraminiferal mi-
crofauna. The distal or seaward part of this regime consists of packstones character-
ized by dasyclad and encrusting algae. The ramp-crest shoal is composed of skeletal
and oolitic grainstone, mud-lean packstone, and some mud-rich packstone. Skele-
tal sands of micritized foraminiferal tests and broken skeletal detritus also include
larger fragments of transgressive and storm-derived stromatoporoids and corals.
The proximal middle ramp is composed of domed and encrusting stromatoporoid-
coral mounds and intermound sheltered areas dominated by branched stromatopo-
roids. The distal middle ramp, deposited below fair-weather-wave base, is composed
of micritic to very fine-grained sediment capped by Thalassinoides firmgrounds.
These firmgrounds are overlain by storm-derived rudstone and floatstone of inner-
ramp, ramp-crest, and proximal middle-ramp bioclasts. The outer ramp is com-
posed of deeper shallow-marine deposits of micritic to very fine-grained sediment
capped by Thalassinoides firmgrounds. In this setting, smaller, benthonic forami-
nifera are common along with tetraxon and triaxon sponge spicules and coccoliths.
From highest to lowest reservoir quality, the lithofacies or rock types con-
sist of (1) skeletal-oolitic grainstone, mud-lean packstone, and some mud-rich
packstone; (2) stromatoporoid-red and green algae-coral rudstone and float-
stone; (3) Cladocoropsis rudstone and floatstone; (4) dolomite, porous and locally
extremely permeable to nonporous; (5) bivalve-coated grain-intraclast rudstone
and floatstone; and (6) micritic to very fine-grained deposits.
Limestone porosity is a mixture of the following common pore types: interpar-
ticle (dominant), moldic (common), intraparticle (common), and microporosity.
Less common is porosity associated with Thalassinoides burrows, with vertically
oriented tunnels filled by grain-rich sediment. Shelter porosity is uncommon. Do-
lostone porosity, less common than the major limestone pore types, is a mixture of
moldic, intercrystalline, and (least common) intracrystalline porosity. Fractures
(least common) do not contribute much porosity but contribute permeability.
Diagenetic effects common within Arab-D reservoir carbonates include sev-
eral dissolution events, recrystallization, and physical compaction. Cementation,
episodes of dolomitization, and chemical compaction-stylolitization, although
locally important, were less abundant events.
The vertical seal for the reservoir is the overlying Arab C-D anhydrite. It is
more than 30 m (100 ft) thick and is composed of varvelike laminae to very thin
beds of anhydrite (thicker) and carbonate or organic matter (thinner) deposited in
a salina. The salina periodically shallowed upward into peritidal and intertidal
settings. A few porous-permeable carbonate stringers were deposited when rel-
ative sea level rise flooded the evaporitic shelf and temporarily restarted the sub-
tidal carbonate factory, whereas relative sea level fall reestablished the subtidal
brine factory and precipitated more evaporites.

INTRODUCTION derived from thermally mature, organic-rich carbonate

source rock of the Jurassic-age Hanifa and/or Tuwaiq
The Arab-D reservoir of Saudi Arabia is the most pro- Mountain Formations (Ayres et al., 1982; Droste, 1990)
lific oil-producing interval in the world (Bates, 1973; that subsequently migrated into highly porous and
Beydoun, 1988; 1998; 1991; Durham, 2005). Oils were permeable carbonate reservoir rocks, Arab-D Member
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 99
of Arab Formation and upper Jubaila Formation, in
large structural traps (Figure 1). Overlying highly effi-
cient evaporite seals, the Arab C-D anhydrite and, ulti-
mately, the Hith anhydrite, prevented further vertical
migration and ensured the containment of oil in these
large structural traps.
Early correlation of well logs, before the advent of
sequence stratigraphy, subdivided the Arab-D reservoir
in Ghawar field from top to base into zones 1, 2, 3, and
4. Zones 2 and 3 have been subsequently subdivided
into zones 2a and 2b and zones 3a and 3b, respectively,
with a more detailed zonation scheme for reservoir
management. Arab-D Member of the Arab Formation
consists of zone 1, zone 2a, and upper zone 2b, and
the remainder of the Arab-D reservoir, consisting
of lower zone 2b, zones 3a and 3b, and zone 4, is part
of the upper Jubaila Formation (Meyer et al., 2000).
DISCOVERY OF OIL IN ARABIA
The first wildcat well drilled in Saudi Arabia was
on Dammam dome, a surface structure with four-way
closure visible from the island of Bahrain, some 40 km
(25 mi) away. Saudi Aramco Dammam 4 discovered oil
in Arab Formation carbonates on March 4, 1938.
Other surface structures were identified as addi-
tional wildcat drilling sites in eastern Saudi Arabia.
However, some surface structures were drilled (e.g.,
Ma’agala and El Alat), only to find they were false struc-
tures created by differential solution of lower Eocene
FIGURE 1. Generalized Late
Jurassic stratigraphy and lithol-
ogies in Saudi Arabia, with Arab
and Hith reservoirs named.
Modified from Powers (1968)
and Meyer et al. (1996).
Rus evaporite beds. To ad-
dress this problem, a shallow
drilling program (generally
less than 300 m [1000 ft])
collected stratigraphic and
structural information be-
neath the pre-Neogene un-
conformity, drilling to the
brown crystalline limestone
of Eocene age beneath the
Rus evaporite. Shallow dril-
ling identified true subsur-
face structures from false
structures and outlined an
anticlinal trend that became
known as Abqaiq, southwest of Dammam dome.
Saudi Aramco Abqaiq 1 was drilled in 1940 and dis-
covered oil in porous Arab Formation carbonates.
At the same time, field mapping near the Rub’Al
Khali desert Empty Quarter, 320 km (200 mi) southwest
of Dammam dome, identified Wadi Sahaba, a west-
to east-oriented wadi that abruptly made a turn to
the south before continuing on to the east. Dips taken
throughout the area revealed a large, broad, low-relief
dome that was named the ‘‘Haradh feature.’’ This was
the first indication of the southern tip of the large En
Nala anticline, the structural feature on which the shal-
lowest crest of the Ghawar field was ultimately found.
As the shallow drilling program continued, it re-
vealed a structural closure at Ain Dar, west of Abqaiq
and, ultimately, through 1941, connected that clo-
sure with a series of closures all the way to the Haradh
closure along the En Nala anticline. In mid-1948,
Saudi Aramco Ain Dar 1 established production from
the D member of the Arab Formation and the Jubaila
Formation and raised the possibility that all of the
structurally high areas along the En Nala anticline
could be productive. This possibility was confirmed
in late 1948 by the successful completion of Saudi
Aramco Haradh 1, some 200 km (120 mi) farther south.
Between Ain Dar and Haradh, the next closed structure
to be drilled was in an area called ‘‘Ghawar,’’ referring to
an area below an escarpment that dominated the area.
The wildcat, named ‘‘Ithmaniya’’ well 1, now called
‘‘Uthmaniyah well 1,’’ after a small djebel east of the
100 / Lindsay et al.
staked location, was yet another discovery in the Arab-D
and upper Jubaila. Another structural closure east of Ain
Dar, now called Shedgum, produced an additional dis-
covery in the Arab-D and upper Jubaila in 1952. South of
Uthmaniyah, the Saudi Aramco Huiyah 1 wildcat, later
changed to Hawiyah, was drilled and became yet an-
other discovery in the Arab-D and upper Jubaila.
Between 1954 and 1957, successful wildcat and
production wells left little doubt that all of the dis-
covery areas, Ain Dar, Haradh, Uthmaniyah, Shed-
gum, and Hawiyah, were part of one large oil field. To
simplify logistics, each area name was used for well
names, such as, Ain Dar well 1, Haradh well 1, etc.,
whereas the entire field became known as Ghawar.
Basic tools used to discover Ghawar were fairly sim-
ple and involved the Brunton compass, alidade, and
plane table for surface mapping; barometer to determine
altitude; a shallow drilling program; and good geologic
insight. Both the historical outline and the techniques
used are expanded by Stegner (1971), Barger (2000), Keith
(2005), and T. A. Pledge (2006, personal communication).
STRUCTURAL SETTING
The main oil-producing area of Saudi Arabia, in-
cluding Abqaiq and Ghawar fields, is located on the
northeastern part of the central Arabian arch (Figure 2).
FIGURE 2. Generalized geo-
logic map of the Arabian Pen-
insula and the position of
the central Arabian arch. Mod-
ified from Al-Hinai et al.
(1997) and U.S. Geological
Survey (1963).
To the west is the Arabian
shield, a vast complex of Pre-
cambrian igneous and meta-
morphic rocks with some youn-
ger igneous rocks (Pollastro
et al., 1999). Bordering the
shield to the east are interior
escarpments where long ar-
cuate belts of Paleozoic, Me-
sozoic, and lower Tertiary rocks
crop out and dip basinward
(eastward) at about 18 or less
(Powers, 1968). In eastern
Arabia, almost flat-lying Ter-
tiary and younger sediments
of the interior or Arabian plat-
form effectively cover and mask these older sedi-
ments (Powers et al., 1966).
The north-northeast-trending En Nala anticline,
the structural trap for Abqaiq and Ghawar fields, is
dominated by mostly north – south to north-north-
east – south-southwest-trending anticlinal trends
and flexures that reflect deep-seated basement faults
(Figure 3) (Ayres, et al., 1982; Al-Husseini, 2000).
Drilling and seismic evidence (Wender et al., 1998;
Konert et al., 2001) supports the presence, at depth,
of large faults bounding basement blocks under
the Ghawar structure that extend from Precambrian
metasedimentary basement up through Paleozoic sedi-
mentary rocks and die out in the Mesozoic section
(Konert et al., 2001). Broad flexing of post-Paleozoic
sedimentary rock is the dominant structural style in
eastern Arabia (Powers, 1968; Konert et al., 2001)
and is exemplified by the En Nala anticline (Powers
et al., 1966), the structural trap for Ghawar and Abqaiq
fields. Although there is some evidence of subtle growth
of the En Nala structure at least as early as the Perm-
ian, growth was most pronounced in the Late Creta-
ceous, with regional compressive-transpressive stresses
being imposed on eastern Arabia in response to the
closing of the Neo-Tethys Ocean (Beydoun, 1991; Nich-
olson, 2000, 2002). The most recent pulses of structural
growth occurred during the late Eocene–Oligocene
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 101

FIGURE 3. Major tectonic fea-

tures of the Arabian plate and
Iran. Ghawar field is the large
green oil field in the center that
trends north-northeast. Modi-
fied from Konert et al. (2001).

sion-transtension associated
with the continuing vertical
growth of these anticlines. In
core, oblique to vertically ori-
ented stylolites are evidence of
postlithification compression-
transpression associated with
the vertical fractures.

STRATIGRAPHIC
NOMENCLATURE
AND AGE

The uppermost Jubaila For-

mation, or Ju2 Member (Enay,
1987) as defined in outcrop
where the Jurassic section was
first described and named,
is correlative with the middle
and lower part of the Arab-D
reservoir in the subsurface
(Powers et al., 1966; Powers,
1968) (Figures 1, 4). In out-
crop, the top of the Jubaila
Formation is placed immedi-
ately above the highest oc-
curring stromatoporoids, here
interpreted to mean encrust-
ing and domed stromatopor-
oids, almost invariably not
in original growth position,
and at the base of the over-
lying rippled calcarenite (D.
and during the late Miocene–Pliocene, as the Arabian Vaslet, 2002, personal communication to G.W.
plate welded onto the Eurasian plate along the Zagros Hughes) (Figure 4). The upper part of the Arab-D
suture or crush zone (Beydoun, 1991; Glennie, 2000) reservoir is composed of the D member of the Arab
(Figure 3). Minor faulting, fracturing, and jointing are Formation. The Arab-D reservoir is the thickest and
found in the younger formations, especially along oldest of the four reservoirs in the Arab Formation
crests of large flexures where two vertical joint sets are and is overlain by the C-D anhydrite, the upper unit
recognized, including a dominant north-northeast to of the Arab-D Member. Ages for the Arab-D – Jubaila
northeast trend and a subsidiary north-northwest to succession are well constrained, with the Jubaila
northwest joint set (Hancock et al., 1984). Image Formation considered to be early Kimmeridgian
logs suggest that there are also east – west-trending (Arkell, 1952), whereas the overlying Arab Formation
fractures (K. A. T. MacPherson, 2005, personal com- is thought to be late Kimmeridgian to early Titho-
munication). These features may result from exten- nian (LeNindre et al., 1990; De Matos and Hulstrand,
102 / Lindsay et al.
1995; Hughes, 1997, 2004). The D member is restricted
to the Kimmeridgian (Sharland et al., 2001). Although
numerous benthonic foraminifera with stratigraphic
ranges consistent with a Kimmeridgian age have been
described from the Arab-D reservoir, none establishes
a more precise date.
PALEOGEOGRAPHIC SETTING AND REGIONAL
DEPOSITIONAL ENVIRONMENTS
During the Late Jurassic, a vast, shallow carbonate
ramp extended from the central Arabian Peninsula,
west of the present site of Riyadh, as far east as the
present Zagros Mountains of Iran, as far north as
central Iraq and south into Oman (Figure 5). A thick
succession of carbonate and evaporite sediments was
deposited on this shelf (Figure 1). Differential intra-
plate subsidence, partly enhanced by peripheral aggra-
dation of shallow-marine carbonates, led to the devel-
opment of intrashelf basins, including the Gotnia,
Arabian, and Rub’Al Khali basins (Al-Husseini, 1997;
Ziegler, 2001), during deposition of the Callovian
Tuwaiq Mountain and Oxfordian Hanifa formations.
Organic-rich carbonate source rocks were deposited
in these intrashelf basins during part of the Oxfordian
(Ayres et al., 1982; Droste, 1990; Carrigan et al., 1994).
These intrashelf basins tended to be maintained
through the remainder of the Jurassic, influencing the
deposition of the overlying Jubaila, Arab, and Hith
formations (Figure 6) and bounding broad, relatively
stable shallow shelf or platform areas on which the
Arab reservoir carbonates were deposited. The pe-
ripheral highs around the intrashelf basins are in-
terpreted to have been the loci for the initiation of
FIGURE 4. Jubaila Formation west of
Riyadh in Wadi Leban. Ju2 member
is the upper, resistive beds. Ju1 member
is the lower, recessive beds. Ju2 cliff
former is 15 m (50 ft) thick.
stromatoporoid and coral banks and
overlying ramp-crest skeletal-oolitic
grain shoals, which subsequently pro-
graded over much of the area of the
intrashelf basins.
Overall, the paleoclimate was prob-
ably hot and arid, similar to the pres-
ent climate on the Arabian Peninsula
(Handford et al., 2002). The general
distribution of evaporites in the Hith
Formation, which formed in response
to a restriction of the shelf, and especially of intrashelf
basins, from open-marine circulation, was also influ-
enced by these intrashelf basins; evaporites are typi-
cally thickest and most halite prone in the intrashelf
basins.
ARAB-D LITHOFACIES AND
DEPOSITIONAL ENVIRONMENTS
The interpretations presented herein are based
primarily on observations from core in the Shedgum
area of Ghawar field (Figures 7, 8 [the foldout located
in the back of this publication]).
Previous workers (Mitchell et al., 1988) have de-
veloped a classification scheme to organize Arab-D
rocks into genetically meaningful packages. This clas-
sification divides Arab-D rocks into six depositional
lithofacies, which include one anhydrite and five car-
bonate lithofacies. Carbonate lithofacies are distin-
guished on the basis of their typical depositional com-
ponents and include (1) skeletal-oolitic limestones
and dolomites (SO); (2) Cladocoropsis limestones and
dolomites (CLADO); (3) stromatoporoid-red (green)
algae-coral limestones (SRAC) (In this lithofacies, a
common organic component is Thaumatoporella par-
vovesiculifera (Hughes, 1996), a microbial encruster that
De Castro (1991, 2002) suggests is a chlorophycean,
or green alga. The encrusting form gives it the appear-
ance, both megascopically and under hand lens or bin-
ocular microscope, of a red alga and was originally
described as a red alga. The ‘‘SRAC’’ acronym is suffi-
ciently widely present in the literature that we continue
to use it despite the fact that few red algal fragments
actually exist in the lithofacies.); (4) bivalve-coated
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 103

FIGURE 5. Late Jurassic paleogeography of the Arabian Peninsula and surrounding area. Average directions of Arab-D
and Jubaila progradation are shown by arrows. Blue band west of Riyadh is the Tuwaiq escarpment. Modified from
Handford et al. (2002), Al-Husseini (1997), Ayres et al. (1982), R. B. Koepnick and L. E. Waite (1991, personal communication),
Murris (1980), and Scotese (1998).

grain-intraclast limestones (BCGI); and (5) micritic
limestones and dolomites (MIC). Because dolomitiza-
tion frequently destroys evidence of original depo-
sitional lithofacies, the diagenetic lithofacies dolomite
(DOLO) was added (Mitchell et al., 1988) (Figure 9).
Although some later workers have sought to revise this
classification (Meyer and Price, 1993; Meyer et al., 1996;
Handford et al., 2002), these later modifications were
not widely accepted and have since fallen out of use.
These lithofacies, with the exception of the dia-
genetic dolomite lithofacies (DOLO), were deposited
approximately 58 south of the equator on a broad,
arid, storm-dominated carbonate ramp. The ramp
consisted of the following subenvironments from
upramp to downramp: (1) inner ramp; (2) ramp-crest
shoal; (3) proximal middle ramp; (4) distal middle
ramp; and (5) outer ramp (Figure 10). Each of these
environments is characterized by one or two of the
depositional lithofacies. In addition, there was a sa-
lina environment that developed after the deposition
of the sediments of the Arab-D reservoir, depositing
the dominantly anhydrite interval that is the vertical
seal of the reservoir. Because a close relationship exists
between the depositional environments and the litho-
facies, both will be discussed together below.
Anhydrite Lithofacies; Salina Environment
Anhydrite (nonporous) with some interbedded car-
bonate stringers (some of which, especially where
dolomitized, are porous) makes up most of the C-D
anhydrite and overlies the Arab-D reservoir (Figure 8,
the foldout located in the back of this publication).
Typical anhydrite fabrics include nodular, bedded
nodular, nodular mosaic, and massive (classification
of Maiklem et al., 1969); abundant, vertically aligned
(elongate) and uncommon mosaic anhydrite are also
observed (Figures 11, 12). In thin section, anhydrite
locally exhibits a felted replacive texture of very small,
oriented, needlelike anhydrite crystals and clear, coars-
er, rectangular laths of diagenetic anhydrite. On the
crest of the En Nala anticline in Ghawar field, capil-
lary pressure has been strong enough to force light-end
hydrocarbons, which look like wispy films, between
evaporite crystals and nodules.
104 / Lindsay et al.

FIGURE 6. Regional depositional environments of the Arab and Hith formations. Most of the time, deposition of high-
energy grainstones and packstones dominated the broad carbonate shelf. This was succeeded by later widespread
deposition of Hith evaporites. Modified from Ziegler (2001).
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 105
FIGURE 7. Shedgum Arab-D type log (open-hole log
gamma-ray density-neutron), Ghawar field. Six cored
intervals are shown to the right, with a total of 91 m (299 ft)
cored and depth adjusted to match open-hole logs. Core
coverage is through the complete Arab-D reservoir.
Deposition of the anhydrite occurred after a major
fall of relative sea level that subaerially exposed the
Arab-D carbonate, marking the top of the Arab-D car-
bonate. Small sink holes, filled with collapse breccia,
have been seen in outcrop (Tuwaiq escarpment) and
in Ghawar field in core and on image logs (Figure 13)
where anhydrite overlies and fills in the karstic fea-
tures. Following subaerial exposure, the initial rise of
sea level was slow, and the regional carbonate shelf
and associated interior basins, such as the Arabian
intrashelf basin, became the site of evaporite deposi-
tion (probably a combination of mostly gypsum and
some anhydrite) (Figures 11, 12). Historically, nodu-
lar mosaic to massive anhydrite, typical of the C-D
anhydrite, has commonly been interpreted as form-
ing within a sabkha environment (Kinsman, 1964;
Shearman, 1978). However, vertically aligned crys-
tals, thought to have been deposited originally as
subaqueous gypsum palmate crystals, and bedded an-
hydrite nodules in the C-D anhydrite suggest that
most of the anhydrite probably formed subaqueously
in a salina to peritidal to intertidal setting (McGuire
et al., 1993; Handford et al., 2002) (Figures 11, 12).
Recent detailed descriptions of the anhydrite in
two cored wells show that the sediments were de-
posited in cycles, with each evaporitic cycle starting
with subaqueous salina deposition and growth of
palmate crystals. Slight shallowing to a peritidal po-
sition between the salina proper and the intertidal
resulted in palmate crystals growing at an oblique
angle. Subsequent intertidal environments produced
flat to obliquely oriented gypsum crystals with lami-
nated carbonate mud. Salina and peritidal settings
are the most common, with the intertidal setting as
the least common. The laminae to very thin beds
that are the most common bedding in the evapo-
rite appear to be varvelike and form evaporite cou-
plets that contain thicker evaporite bases, interpret-
ed as summertime deposition, and carbonate- and
organic-rich laminae caps, interpreted as wintertime
deposition.
Uncommon, large transgressive events freshened
the highly saline setting to normal-marine salinity,
which allowed the subtidal carbonate factory to start
up and deposit a cycle of carbonate referred to as a
carbonate stringer. Sometimes, these only formed
very thin to thin beds, but occasionally deposited a
carbonate cycle 1.5–2 m (5–6 ft) thick, with the top
portion porous and grain rich. During the following
overall relative sea level fall, restriction and evapora-
tion recommenced, and the subtidal evaporite factory
again started to deposit subaqueous evaporites.
106 / Lindsay et al.
FIGURE 9. Summary of characteristics and distribution of carbonate lithofacies in the Arab-D reservoir, Ghawar field.
Modified from Mitchell et al. (1988).
Inner Ramp
The inner ramp, landward and upslope from the
landwardmost extent of sediment agitation by sig-
nificant wave activity and behind the shoal crest,
was a lagoon with localized intertidal islands com-
posed of wackestone, packstone, and some grain-
stone (Figure 10). The sediments of the inner-ramp
lagoonal setting are only present at the top of the
Arab-D reservoir (zone 1) and are approximately 3 m
(10 ft) thick (Figure 8, the foldout located in the back
of this publication). This depositional setting con-
tains a restricted fauna, suggesting a lack of normal-
marine circulation. Normal-marine fauna, such as the
delicate stick-shaped stromatoporoid Cladocoropsis,
have been described locally within this setting and
may owe their presence to either storm transport or to
a normal-marine passageway, through the ramp-crest
shoal, that connected part of the inner ramp with the
proximal part of the middle ramp.
Skeletal-oolitic Lithofacies (SO); Ramp-crest
Shoal Environment
Skeletal-oolitic (SO) limestone and dolomitic lime-
stone is present in the upper part of the Arab-D reser-
voir (zone 2a) and another interval deeper in the res-
ervoir (zone 2b) (Figure 9). This lithofacies is typically
well sorted and contains grainstone, mud-lean pack-
stone, and some mud-rich packstone (Figures 8 [the
foldout located in the back of this publication] and
14). Overall, the lithofacies represent the best sorted
and, at the megascopic visual level, the most uniform-
appearing carbonate rock type of the Arab-D. Major
FIGURE 10. Idealized depositional model of the Arab-D reservoir carbonate ramp in Ghawar field. This model depicts deposition of one carbonate cycle from the
inner ramp, ramp-crest shoal, proximal, and distal middle ramp to outer ramp. Blue is a transgressive mud-rich deposit capped by a firmground that developed
during maximum flooding and was burrowed by Thalassinoides. Brown with diamond shapes is distal middle-ramp debris-flow deposits of rudstone and floatstone.
Green is proximal middle-ramp biostromes and mounds of stromatoporoids and corals. Blue with fork shapes is proximal middle-ramp Cladocoropsis banks
deposited in sheltered areas between and upslope of biostromes and mounds. Red is the cross-bedded ramp-crest grainstone shoal. Purple is inner-ramp lagoon-
intertidal islands. Those facies and rock types that are abundant in each reservoir zone are underlined by the titles and associated arrows zone 1, zone 2, and zone 3.
Distribution of the common to abundant faunal and floral elements (both laterally at any one time and stratigraphically through the reservoir zones) is shown
below the figure.
Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 107
108 / Lindsay et al.
FIGURE 11. Arab C-D evaporite from 2071 – 2073 m (6796 – 6804 ft) in top of core 1 in the Shedgum Arab-D core, Ghawar
field. Evaporite crystals were originally gypsum that grew on the floor of a salina. Note the varvelike cyclicity and
enterolithic folds.
grain types include foraminifera, dasycladacean algae,
micritized grains and foraminifera, ooids, and bivalves
(Figures 14 – 16). Porosity, from most abundant to
least abundant, includes interparticle, moldic, intra-
particle, microporosity, and, locally common within
dolostones, intercrystal pores (Figures 14–16). Cross-
bedding and horizontal laminations are common
(Figures 8 [the foldout located in the back of this
publication] and 14). Other common features include
firmgrounds and hardgrounds (Figure 8, the foldout
located in the back of this publication), burrows, bor-
ings, and coarsening-upward beds. Common types of
diagenesis include dissolution and recrystallization,
fine isopachous bladed calcite and microspar cementa-
tion, and dolomitization. Although scattered dolomite
crystals are commonly found, there does not appear to
be any complete dolomitization of this lithofacies.
The skeletal-oolitic lithofacies were deposited in
the ramp-crest shoal environment, consisting of a se-
ries of ooid-peloid-intraclast sand shoals that were
locally and briefly subaerially exposed as islands in a
high-energy very shallow subtidal to intertidal setting
in fair-weather-wave base. It was continuously wave
washed and occasionally storm swept by the highest
day-to-day energy to have affected the carbonate ramp
(Figure 10). Fair-weather-wave base is defined as a zone
where waves first touched the sea floor and continued
to wash the sea floor with high energy, until friction
reduced wave energy, and sediment movement ceased
(zone Y of Irwin, 1965). The grain-rich setting is cross-
bedded and slightly cemented by fine, isopachous, ma-
rine, and microspar cement. Skeletal detritus includes
storm-derived fragments and larger to nearly whole
Cladocoropsis, stromatoporoids, and corals from farther
seaward. Microbial encrustations are fairly common.
The ramp-crest shoal depositional setting and as-
sociated skeletal-oolitic lithofacies form the upper
Arab-D reservoir (most of zone 2a) (Figures 8 [the fold-
out located in the back of this publication] and 10),
as well as the uppermost part of the lower sequence
(uppermost Jubaila Formation).
Cladocoropsis Lithofacies (CLADO); Sheltered
Proximal Middle-ramp Environment
Cladocoropsis (CLADO) limestone and dolostone
contain greater than 10% of the branching stromato-
poroid Cladocoropsis mirabilis (Champetier and Four-
cade, 1966) and occur in the upper and middle Arab-D
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 109
reservoir (zones 2a and upper 2b) (Figures 8 [the
foldout located in the back of this publication] and
17). Cladocoropsis rudstones and floatstones contain
a matrix of grainstone, mud-lean packstone, and
mud-rich packstone. They are not as well sorted as
those of the skeletal-oolitic lithofacies and are there-
fore more heterogeneous (Figure 17). Cladocoropsis
fragments, dasycladacean algae, foraminifera, and mi-
critized grains and forams are common (Figure 17).
Pore types, from most common to least common, are
interparticle, intraparticle, moldic, and microporous,
with all types abundant (Figure 17), and when do-
lomitized, intercrystal and moldic porosity are the
dominant pore types. Sedimentary structures are sub-
tle, but close examination of cores shows that cross-
bedding and horizontal laminations are present in
highstand parts of cycles (parasequences), whereas
soft-body bioturbation and firmgrounds and hard-
grounds are more common in transgressive parts of
cycles (Figure 8, the foldout located in the back of this
publication). Diagenesis is commonly limited to dolo-
mitization, dissolution, recrystallization, and uncom-
mon equant calcite cementation. When completely
FIGURE 12. Arab C-D evaporite
from 6804 to 6809 ft (2073 to
2075 m) in top of core 1 in the
Shedgum Arab-D core, Gha-
war field. Evaporite crystals
were originally palmate gyp-
sum that grew on the floor of a
salina. Note varvelike cyclicity.
Carbonate at the base has
been dolomitized. Cores are
2.5 in. (6.35 cm) wide.
dolomitized, Cladocoropsis
fragments typically undergo
dissolution to form moldic
porosity and, if originally in
contact with each other, de-
velop super-K (high-per-
meability) intervals.
The Cladocoropsis (CLADO)
lithofacies, along with the
stromatoporoid-red algae-
coral (SRAC) lithofacies, de-
scribed below, was deposited
in a proximal middle-ramp
environment, with deposi-
tion just within fair-weather-
wave base, where constant
daily circulation occurred.
This environment was char-
acterized by grain-rich, domed, and encrusting stro-
matoporoid and coral biostromes and mounds.
Intermound sheltered areas and sheltered areas im-
mediately upslope of the biostromes and mounds
were populated by Cladocoropsis (Figures 8 [the fold-
out located in the back of this publication] and 10)
(Leinfelder et al., 2005). This interpretation is based, in
part, on comparison with domed and branched co-
rals in modern environments ( James, 1983), where
domed forms occur in biostrome and mound envi-
ronments with moderate wave energy and low rates
of sedimentation, whereas branched forms, such as
Cladocoropsis, occupy lower energy sites with higher
rates of sedimentation. Cladocoropsis is commonly
accompanied by the foraminifera Kurnubia palasti-
niensis and Nautiloculina oolithica (Figures 15, 16). In
the shallower marine moderate-energy SRAC litho-
facies, the microbiota is dominated by a dasyclad
alga attributed to Clypeina sulcata; the green alga
T. parvovesiculifera is common and is joined by a
background population of the forams Kurnubia sp.,
Nautiloculina sp., Redmondoides lugeoni, and some
biserial agglutinated forms (Figures 15, 16). Microbial
110 / Lindsay et al.

FIGURE 13. A. Examples of breccia-filling sinkholes in the Arab-D in

Ghawar field and in outcrop. The upper photograph is a small sinkhole
in the Arab Formation filled with fitted-fabric and chaotic collapse breccia
in Wadi Birk, 150 km (93 mi) south of Riyadh. Red baseball cap is at the
base of the breccia, and each geologist stands at the side of the breccia
with rock hammer (left) and hand (right) on the breccia-country rock
contact. B. Formation MicroImager image log in a horizontal well in the
south part of Ghawar, with rounded chaotic breccia just beneath the
Arab-D composite sequence boundary and overlying Arab C-D evaporite
(left). (Scale is in feet.) C. Photograph from the central part of Ghawar in
a vertical well. Rounded chaotic breccia has red and brown (terra rosa)
geopetal between breccia clasts, along with plant rootlets. Overlying Arab
C-D evaporite contains palmate gypsum crystal ghosts. (Photo is 6.4 cm
[2.5 in.] wide.)
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 111
FIGURE 14. Ramp crest shoal skeletal-oolitic lithofacies
composed of cross-bedded, intraclast-peloid-coated grain-
ooid grainstone with abundant interparticle porosity
and some moldic, intraparticle, and microporosity. Ce-
mentation was by very fine-bladed calcite and microspar
cement. Note minor physical and chemical compaction
in the upper and more prevalent compaction in lower
photomicrographs. Photomicrograph widths 1.1 mm (up-
per) and 2.25 mm (lower), Shedgum area.
encrustations are also common (Figure 8, the foldout
located in the back of this publication).
Considerable facies variability exists within and
between the biostromes and mounds versus shel-
tered parts of this depositional environment. These
two lithofacies contain diverse, normal-marine biota,
which include common to abundant foraminifera, al-
gae, and microbial encrustations, along with marine
cement, which are evidence of deposition under normal-
marine conditions. Fairly common cross-bedding also
argues for a moderate to potentially high-energy set-
ting. Hermatypic corals imply deposition in the photic
zone, whereas the presence of such stenotopic organ-
isms (organisms that have a limited tolerance for
changes in environmental conditions) as coral indi-
cates normal-marine conditions.
Cladocoropsis have also been reworked with no in-
place, upright-standing organisms being identified
and very few actual intact branches being preserved.
Fieldwork in the Tuwaiq Mountain Formation out-
crops in the Tuwaiq escarpment, west of Riyadh, iden-
tified Shuqria, an earlier relative of Cladocoropsis, to
have populated sheltered intermound areas between
stromatoporoid and coral biostromes and mounds.
Cladocoropsis probably populated a similar depositional
setting. Storm events reworked Cladocoropsis and have
strewn them within the proximal middle ramp, and
more powerful storms reworked fragments back into
the ramp crest (Figure 10).
Stromatoporoid-Red Algae-Coral Lithofacies
(SRAC); Biostrome-mound Proximal
Middle-ramp Environment
Stromatoporoid-red algae-coral (SRAC) limestones
and dolomitic limestones occur in the middle Arab-D
reservoir (zone 2b) and represent the most heterog-
eneous rock types (Figures 8 [the foldout located in
the back of this publication] and 18). Locally, this
lithofacies is present in the lower Arab-D reservoir as
allochthonous debris-flow deposits (Figures 8 [the fold-
out located in the back of this publication] and 10).
This lithofacies is, in general, extremely poorly sorted
and contains both pores and grains (Figure 18) that
exhibit a wide range of sizes and shapes. Common
grain types include domed, digitate, and encrusting
stromatoporoids, corals, foraminifera, micritized grains,
and microbial encrustations (Figure 18). Scleractinian
corals found in the Arab-D reservoir were originally
composed of metastable aragonite and were suscep-
tible to dissolution during diagenesis (Figure 18) and
form large, irregular molds. Stromatoporoids and cor-
als form floatstone to rudstone fabrics (Figure 18),
with a matrix of grainstone, mud-lean packstone, and
mud-rich packstone. Moldic and intraparticle pores
are most common in this lithofacies, with interparticle
pores subordinate and micropores common. Typical
diagenetic modifications include dissolution, bioero-
sion, recrystallization, dolomitization, anhydrite em-
placement, fine isopachous bladed calcite and equant
calcite cementation, and microspar cementation.
High-energy storm events buffeted biostromes and
mounds and reworked them locally, with stronger
storms reworking bioclasts and grains downslope
as debris flows into the distal middle-ramp setting
(Figure 10) and commonly, but less abundantly, trans-
porting fragments upslope into the ramp-crest and
inner-ramp environments. The former effect is seen in
outcrops of the Jubaila Formation in the Tuwaiq escarp-
ment west of Riyadh where stromatoporoid and coral
biostromes and mounds are virtually all reworked. Only
one in-place stromatoporoid-coral biostrome, 1 m (3 ft)
thick and 30 m (100 ft) in width (Figure 19), has been
identified in Wadi Hilwah (Hughes, 2004). All other
112 / Lindsay et al.
outcrops show stromatoporoids and corals to have
been at least locally reworked in the proximal middle
ramp, or they have been reworked farther downdip
into the distal middle ramp (Figure 10).
Bivalve-coated Grain-intraclast Lithofacies (BGCI);
Distal Middle-ramp Environment
Bivalve-coated grain-intraclast (BCGI) limestone
and dolomitic limestone occur mostly in the lower
Arab-D reservoir (zones 3a, 3b, and 4) as rudstones
and floatstones, with grainstone, mud-lean packstone,
mud-rich packstone and uncommon wackestone ma-
trix (Figures 8 [the foldout located in the back of this
publication] and 20). Bivalves, coated grains, intra-
clasts, micritized grains, and foraminifera are domi-
nant grain types. Porosity is dominantly interparticle
FIGURE 15. Shedgum area
Arab-D and Jubaila significant
biocomponents. Width of view
is in millimeters. 1 = brachio-
pod valves (2 mm); 2 = cerithid
gastropod (4 mm); 3 – 4 = Tro-
cholina palastiniensis (1 mm);
5–6 = Quinqueloculina sp. (1 mm);
7 – 8 = Pfenderina salernitana
(1 mm); 9 – 10 = Mangashtia
viennoti (1 mm); 11 = Alveosepta
jacardi (1 mm); 12 – 13 = Red-
mondoides lugeoni (1 mm); 14 =
Nautiloculina oolithica (1 mm);
15 = Kurnubia palastiniensis
(1 mm); 16 – 17 = Clypeina sulcata
(1 and 2 mm); 18 = Heteroporella
jaffrezoi (1 mm); 19 = coral
(8 mm); 20 = Cladocoropsis mira-
bilis (4 mm); 21 = Thaumato-
porella parvovesiculifera (2 mm);
22 = tetraxon sponge spicule;
23 –24 = Lenticulina sp. (1 mm).
(Figure 20) and moldic, com-
monly from dissolution of
metastable bivalves, but in-
cludes a significant admixture
of microporosity and intrapar-
ticle pores. Diagenetic modifi-
cations noted in this lithofacies
include dissolution, recrystal-
lization, dolomitization, stylo-
litization, and equant calcite
cementation.
The BCGI lithofacies,
along with the micritic litho-
facies described below, are
the typical sediments of the
distal middle ramp, where deposition occurred be-
neath and seaward of fair-weather-wave base. The dom-
inant sediment is micritic to very fine-grained sediment
that is capped by firmgrounds (Figure 20) character-
ized by Thalassinoides burrow systems (Figure 10) and
overlain by storm-derived rudstone and floatstone of
inner-ramp, ramp-crest, and proximal middle-ramp
bioclasts, most commonly, the BCGI lithofacies but
locally including stromatoporoids and corals (SRAC)
(Figure 20).
Burrowing organisms (infauna) were common, and
either these created vertically oriented Thalassinoides
burrows in micritic to very fine-grained muds, or soft-
body burrowers have homogenized the sediment.
Preserved sedimentary structures are uncommon,
other than those created by bioturbation. These factors
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 113

FIGURE 16. Biozonation and key biocomponents of the Arab-D Member of the Arab Formation and Jubaila Formation in
the Shedgum Arab-D core, Ghawar field.

FIGURE 17. Proximal middle-ramp Cladocoropsis rudstone
with a packstone matrix containing interparticle, moldic,
intraparticle, and microporosity. All Cladocoropsis frag-
ments have been reworked by storms. Note the small mi-
cropellets in the matrix (upper). Matrix is composed of
intraclasts, micritized forams, micropellets, dasyclad algae
(Clypeina sulcata), and stromatoporoid fragments (lower).
Photomicrograph widths are 1.1 mm (upper) and 7.4 mm
(lower), Shedgum core.
FIGURE 18. Proximal middle-ramp stromatoporoid-coral
rudstone, containing a mud-rich packstone matrix and
interparticle, moldic, intraparticle, and microporosity.
Stromatoporoids have been reworked by storms and bio-
eroded by bivalves, still visible. Coral mold, with casts of
coralites, is beneath 223 (lower left). Note the micropellets
that acted like microball bearings during storm reworking.
Both photomicrograph widths are 1.1 mm, Shedgum Arab-
D core on display.
114 / Lindsay et al.
suggest that the sediment accumulated in quiet, low-
energy conditions beneath fair-weather-wave base
away from strong wave or current action.
Firmgrounds formed on the flooding surfaces of
individual carbonate cycles. Firmgrounds are typified
by vertical Thalassinoides burrows, along with side
gallery living areas with an average vertical depth of
penetration of 1 m (3 ft) and a range of vertical pene-
tration from only a few tenths of a meter (<1 ft) to as
much as 2.4 m (8 ft) (Figure 21). Thalassinoides bur-
rowers easily mined downward through soft muddy
sediment, but ceased excavation once a grain-rich sub-
strate, containing larger grains and skeletal detritus,
was intersected.
Bivalve-coated grain-intraclast (BCGI) limestone
and, to a lesser extent, stromatoporoids and corals
of the SRAC lithofacies were deposited as rudstone
and floatstone by storm-induced high-energy debris
flows that transported debris and matrix downslope
into a deeper, shallow-marine distal middle-ramp
FIGURE 19. Upper photo is Arab-D –
Jubaila contact in Wadi Hilwah. Geolo-
gist (encircled) is standing by the only
in-place coral-stromatoporoid bios-
trome found in outcrop. All others have
been reworked by storms. Arab-D –
Jubaila contact is about 15 ft (4.5 m)
above the geologist. Top of Arab-D is at
the hill top. Lower photo is Arab-D –
Jubaila contact in Wadi Nisah. Arab-D
is the thin, platy outcrops above geol-
ogists, whereas Jubaila is the thicker
beds beneath their feet. Note the large
storm ripple in Jubaila.
setting (Figure 10). Microbial encrus-
tations, coated grains, and ooids in-
dicate that original, coarse-sediment
formation was farther upslope in
the photic zone and fair-weather-
wave base, in shallow-marine wa-
ter of sufficient turbulence to move
grains around and achieve relative-
ly even coatings of grains. Muddy
intraclasts, ripped up from areas of
mud deposition down the ramp, is
another indicator of brief, high-
energy storm events.
The matrix associated with rud-
stone and floatstone tends to be mud-
rich packstone, with mud-lean pack-
stone and grainstone less common.
The muddy matrix contains very
small micropellets, many of which may have been
created by the fruiting bodies of microbial encrusta-
tions (R. R. Leinfelder, 2003, personal communica-
tion). The entrained mud-rich matrix and micropellets
acted like microball bearings to mediate transport
of storm-generated debris flows much farther
downslope. The relative proportion of mud incorpo-
rated into debris flows was responsible for two types
of deposits. One is a cohesive debris flow with a
mud-rich matrix (Figure 20), and the other is a non-
cohesive debris flow with a less mud-rich matrix
(Figure 20). Additional mud matrix assisted debris
flows in transporting sediment farther down the
ramp slope.
Weaker storm events were responsible for rework-
ing bivalve-coated grain-intraclast (BCGI), whereas
larger, more powerful storms were required to dis-
lodge and transport stromatoporoids and corals be-
cause they formed biostromes and mounds and were
better anchored than were BCGI lithofacies.
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 115

the surrounding lime mudstone matrix is only partially

dolomitized.
As described above, micritic (MIC) limestones,
bivalve-coated grain-intraclast (BCGI) limestones, and,
to a lesser extent, stromatoporoid-red algae-coral (SRAC)
limestones, are typically interbedded in the distal
middle-ramp environment. In the outer-ramp environ-
ment, beyond the reach of storm-generated debris flows
(Figure 10), lime mudstones (micritic lithofacies) were
the only sediments deposited. Each cyclic package is
capped by a firmground (Figures 8 [the foldout located
in the back of this publication] and 10) that are typified
by Thalassinoides burrows that penetrated the muddy
carbonate layer. Although a lagoon paleoenvironment
has been suggested for this lithofacies (Enay, 1987),
the presence of the forams Lenticulina spp., Nodosaria
spp., and Pseudomarsonella spp., along with monaxon
and tetraxon sponge spicules, coccoliths, and juvenile
costate brachiopods, is more consistent with open,
normal-marine conditions below fair-weather-wave
base, in perhaps 25 – 30 m (80– 100 ft) of water depth.
Modern assemblages, including Lenticulina and Nodo-
saria, are confined to water depths greater than 100 m
FIGURE 20. Distal middle-ramp bivalve-coated grain-
intraclast rudstone. The upper part is the cohesive debris
flow (mud rich) and the lower part is the noncohesive
debris flow (grain rich). Debris flows were created by storms
and rest atop firmgrounds burrowed by Thalassinoides.
Photomicrograph widths are 7.4 mm (upper) and 2.2 mm
(lower), Shedgum Arab-D core.

Micritic Lithofacies (MIC); Distal Middle-ramp

and Outer-ramp Environments
Micritic (MIC) limestone and dolostone (Figures 8 [the
foldout located in the back of this publication] and 21)
are common and tend to dominate the lower half of
the Arab-D reservoir (zones 3a, 3b, and 4), less common
in the middle Arab-D reservoir (zone 2b), and least
common to nonexistent in the upper Arab-D reservoir
(zone 2a and 1). In contrast to the other depositional
lithofacies, micritic limestones are mostly mud sup-
ported. Grains present in this lithofacies include mi-
critized grains, bivalves, ostracods, foraminifera, and
FIGURE 21. Outer ramp mudstone, bioturbated by
intraclasts. Although porosity is generally low, typical Thalassinoides burrows and by soft-body burrowers, that
pore types include microporosity, moldic, and intra- contains only low amounts of microporosity. In the upper
particle pores. In addition, interparticle and occasional photomicrograph, note a few dolomite crystals and one
intercrystal porosity is preserved in the grain-rich sedi- quartz silt grain (center) in the pelleted matrix. In the
lower photomicrograph is the contact of a Thalassinoides
ment that filled vertically oriented Thalassinoides bur-
burrow tube that was filled with fine sediment and partially
row tubes (Figure 21). Diagenetic alterations include dolomitized and the mud-rich matrix containing a few
dolomitization, dissolution, recrystallization, and stylo- windblown quartz silt grains. Photomicrograph widths are
litization. Burrows may be highly dolomitized, whereas 1.1 mm (upper) and 7.4 mm (lower), Shedgum Arab-D core.
116 / Lindsay et al.

typically consists of anhedral to euhedral crystals

(30– 500 mm in size) that can be sucrosic or mosaic,
depending on whether intercrystal porosity is pres-
ent. Relict burrows and firmground surfaces are lo-
cally visible (Figure 23). Intercrystal porosity is domi-
nant, with moldic porosity common (Figures 22, 23)
and fracture porosity less common. Thin layers of su-
crosic dolomite locally have sufficiently high poros-
ity that they have anomalously high permeability re-
ferred to as super-K. Uncommon intracrystal porosity
developed when dolomite crystals were partially dis-
solved to form dolo-donuts (Figure 23). Dolomite crys-
tals commonly have cloudy, inclusion-rich centers and
clear, limpid rims (Figures 22, 23). Uncommonly, three-
to four-generation dolomite crystals have been identi-
fied that contain inclusion-rich centers, surrounded
by a clear, limpid layer, with an additional inclusion-
rich layer, followed by a clear, limpid, outermost layer.
Dolomitization of Cladocoropsis lithofacies has lo-
cally resulted in the dissolution of the Cladocoropsis
fragments and the development of moldic porosity.
FIGURE 22. Proximal middle-ramp dolomitized, cross-
bedded, Cladocoropsis, peloid rudstone (upper) and dolo-
mitized, cross-bedded mud-lean packstone (lower), con-
taining moldic and intercrystal porosity. The upper
photomicrograph contains saddle dolomite with ghosts
of peloids. Photomicrograph widths are 2.25 mm (upper)
and 1.1 mm (lower), Shedgum Arab-D core.

(330 ft), but the Mesozoic equivalents are known to

have occupied depths shallower than this.
The outer ramp was probably affected by storms.
Using the Persian Gulf as a modern analog, yearly
storms have been found to stir bottom sediment to a
depth of 50 m (164 ft), whereas a 100-yr storm can stir
bottom sediment to 110 m (360 ft) (Saudi Aramco,
oceanographic department database). However, bio-
turbation from soft-body burrowing organisms and
Thalassinoides probably churned the muddier sedi-
ment and erased most traces of storm events. No
rudstone or floatstone deposits were reworked and
carried by debris flows this far down the ramp from
the inner ramp, ramp crest, and proximal middle ramp
(Figure 10), so distinction between burrow fill and
FIGURE 23. Dolomitized, cross-bedded, Cladocoropsis,
the surrounding matrix is subtle except in those in- skeletal, peloid rudstone, with a mud-rich packstone
stances where differential partial dolomitization re- matrix and containing intercrystal and moldic porosity
sulted from slight depositional differences. and very high permeability (super-K). Long laths are
anhydrite crystals. The lower part is dolomitized, peloid
Dolomite Lithofacies (DOLO) mud-lean packstone at a potential subaerial exposure
surface, with plant rootlets and/or Skolithos – Arenicolites
Dolomite (DOLO) (Figures 8 [the foldout located
burrow traces. The pore space is lined with insoluble
in the back of this publication], 22, 23) generally oc- residue and one dolo-donut intraparticle pore. Photomi-
curs as thin to thick (0.3–4.6-m, 1–14-ft), sheetlike beds crograph widths are 2.24 mm (upper) and 1.1 mm
throughout most of the Arab-D reservoir. Dolomite (lower), Shedgum Arab-D core.
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 117
If individual moldic pores are touching or if the matrix
dolomite has sufficient intercrystal porosity, they de-
velop thin intervals with low porosity but very high
permeability, also referred to as super-K (Figure 23).
The least common dolomite occurrence is saddle
(baroque) dolomite with crystals 500 mm up to a few
thousand micrometers in size (Figure 22). This type
of dolomite has been interpreted to result from hy-
drothermal emplacement along widely spaced frac-
ture systems (Cantrell et al., 2004). Distribution is
areally restricted, but may extend vertically through
much of the reservoir.
ARAB-D RESERVOIR VERTICAL SUCCESSION
AND SEQUENCE STRATIGRAPHY
In general, from bottom to top, the Arab-D reser-
voir consists of (1) outer-ramp, mud-rich carbonate
cycles (parasequences); (2) distal middle-ramp, mud-
rich, and rudstone- and floatstone-capped cycles;
(3) proximal middle-ramp biostromes and mounds of
stromatoporoids and corals and sheltered Cladocoropsis
bank cyclic deposition; (4) ramp-crest shoal, high-
energy cyclic deposition; and (5) inner-ramp lagoon
and localized intertidal islands that form a few cycles
at the top of the section (Figures 8 [the foldout located
in the back of this publication] and 10).
This shoaling-upward succession can be subdi-
vided into 11 informal biozones, with the depths for
the Shedgum Arab-D well shown (Figure 16). In this
figure, the term ‘‘outer ramp’’ has to be understood in the
context of normal, tropical carbonate deposition to in-
dicate water depths of perhaps as much as 30 m (100 ft).
In each of the biozones, there is a characteristic asso-
ciation of biotic elements indicative of the water depth
and associated paleoenvironment shown (Figure 16).
Outer-ramp and distal middle-ramp cyclic depo-
sition was seaward and beneath fair-weather-wave
base, but within or at least influenced by transport
from sediments that were within storm-wave base.
Proximal middle-ramp and ramp-crest shoal depo-
sition was within fair-weather-wave base, and inner-
ramp deposition was landward of wave base. This
succession of carbonate deposition on a carbonate
ramp recorded an overall shallowing-upward history
related to a long-term base-level fall (Mitchell et al.,
1988; Meyer and Price, 1993; Handford et al., 2002).
This overall shallowing-upward, long-term base-
level fall and loss of accommodation space was punc-
tuated by a series of smaller relative sea level rises and
falls (Figure 8, the foldout located in the back of this
publication). These rises and falls of sea level de-
posited the Arab-D reservoir as a series of submeter-
to meter-scale and thicker carbonate cycles (para-
sequences), with each carbonate cycle composed of
one to three beds. Approximately five individual car-
bonate cycles combine to form a cycle set (parase-
quence set), with two to six cycle sets forming a HFS
and four to five HFSs forming a composite sequence.
Two composite sequences form the Arab-D reservoir.
One composite sequence forms the Arab-D Member
of the Arab Formation; the other forms the upper
part of the Jubaila Formation (Figure 24). It is tempt-
ing to relate these different cycle scales to Milanko-
vitch cycles (Milankovitch, 1930, 1941).
Carbonate Cycles (Parasequences)
On the individual carbonate cycle (parasequence)
scale, the Arab-D reservoir is composed of a trans-
gressive mudstone, wackestone, or packstone base,
deposited below fair-weather-wave base, and over-
lain by highstand cross-bedded grainstone, mud-lean
packstone, or mud-rich packstone, deposited with-
in fair-weather-wave base and/or storm-wave base
(Figures 8 [the foldout located in the back of this
publication], 10, 24).
A total of 120 cycles (parasequences) have been
identified in the Shedgum Arab-D core presented
here (Figures 8 [the foldout located in the back of this
publication] and 24). A Fischer plot of these cycles
has been constructed and annotated to show which
cycles were deposited in outer-ramp, distal to proxi-
mal middle-ramp, ramp-crest, and inner-ramp depo-
sitional settings (Figure 24). Individual carbonate cycles
(parasequences), from bottom to top, in the Arab-D
reservoir consist of the cycles discussed below.
Outer-ramp Cycles
Outer-ramp carbonate cycles are submeter to meter
scale and occasionally thicker and consist of mudstone,
wackestone, and local mud-rich packstone that have
an uppermost firmground with Thalassinoides burrows,
with no highstand grain-rich cap (Figures 8 [the fold-
out located in the back of this publication], 10, 24),
although a thin layer of highstand-equivalent sediment
may be preserved at some localities. Thalassinoides-
burrowed firmgrounds represent the maximum flood-
ing surface (MFS) of an individual cycle.
Distal Middle-ramp Cycles
Distal middle-ramp carbonate cycles are sub-
meter to meter scale and thicker and contain trans-
gressive mud-rich bases, similar to the outer-ramp
setting, with the maximum flooding surface (MFS)
118 / Lindsay et al.
FIGURE 24. Fischer plot of the Shedgum Arab-D cored interval. Each carbonate cycle (parasequence) is color coded to
show were it was deposited on the carbonate ramp. The small dimple on each cycle represents maximum flooding.
of the cycle at the top of the mud-rich base, where
a firmground with Thalassinoides burrows formed.
Thalassinoides burrows in Jubaila and Arab Formation
outcrops are spaced 0.1 m (4 in.) from each other.
Highstand progradation caused by relative sea level
fall brought the proximal middle ramp, ramp crest
shoal, and inner ramp close enough so that storms
could rework inner-ramp, ramp-crest, and proximal
middle-ramp mud, grains, and skeletal bioclasts
downslope as debris flows that covered the firm-
ground surface with grainstone or rudstone and float-
stone (Figures 8 [the foldout located in the back of
this publication], 10, 24). In some cases, more than
one debris flow (evidenced by two or more stacked,
fining-upward packages) covered the firmground,
attesting to multiple storms passing through the area.
Farthest downslope, in the distalmost parts of the mid-
dle ramp, highstand rudstone and floatstone caps
are thin and are locally reduced to coarse packstones.
In contrast, upslope and closer to the transition from
proximal to distal parts of the middle ramp, high-
stand rudstones-floatstones thicken, and transgres-
sive cycle bases thin to a point where they are of equal
thickness (Figure 10).
Proximal Middle-ramp Cycles
The proximal middle ramp contains a thinner
transgressive base, with the maximum flooding sur-
face forming a firmground or the top of a bioturbated
interval and a thicker highstand cap composed of bio-
stromes and mounds of stromatoporoids and corals
and Cladocoropsis banks. Cladocoropsis banks were de-
posited in sheltered areas between biostromes and
mounds and in sheltered areas immediately upslope
of the biostromes and mounds, in slightly deeper parts
of the proximal middle ramp (Figures 8 [the foldout
located in the back of this publication], 10, 24). Stro-
matoporoid and coral biostromes and mounds were
reworked by storms, although stromatoporoids and
corals have been found to be still in place (Figure 19).
Cladocoropsis was easily reworked by storms, and none
have been found in their original growth position.
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 119
Ramp-crest Shoal Cycles
The ramp-crest shoal contains thinner transgres-
sive bases that are only locally preserved, with the
remainder eroded away or not deposited as wave base
stirred the floor of the ramp crest and continuously
washed grains with high energy and deposited a thin
to thick cross-bedded highstand cycle cap (Figures 8
[the foldout located in the back of this publication],
10, 24). In this setting, transgressive bases of HFSs
may contain Cladocoropsis, stromatoporoids, and co-
rals. Storms during the highstand also reworked frag-
ments of Cladocoropsis, stromatoporoids, and corals
into cross-bedded, intraclast-peloid-ooid grainstone
and mud-lean packstone (Figure 10).
Inner-ramp Cycles
A lagoon formed behind the ramp crest and included
uncommon, small intertidal islands (Figures 8 [the
foldout located in the back of this publication], 10,
24). Transgressive bases contain bioturbated, mud-
rich to grain-rich strata with Thalassinoides-burrowed
firmgrounds overlain by highstand-associated grain-
rich caps. These form the last few cycles at the top of the
Arab-D reservoir. Intertidal islands, thought to range
in size from a few well spacings to less than one well
spacing (one well space = 1 km [0.6 mi]) in aerial ex-
tent, have been identified in a few cored wells. Gen-
erally, the last cycle of deposition of Arab-D carbonate
was a transgressive lagoon cycle with a Thalassinoides-
burrowed firmground that lacked a highstand cap. We
interpret this facies to be the upslope deposit of a
series of cycles deposited during a forced regression,
terminating the deposition of the Arab-D.
Cycle Sets (Parasequence Sets)
A total of 28 cycle sets (parasequence sets) has been
identified in the Shedgum Arab-D core (Figures 8 [the
foldout located in the back of this publication] and
24). These cycle sets are generally made up of approxi-
mately five individual cycles (parasequences), ranging
from two to six, which record an increase in accom-
modation space that progressively declined with sub-
sequent deposition. In the Arab-D reservoir, the last
few shallowing-upward cycles may thicken, which
implies that the earlier deposited cycles did not
shallow up to sea level and only partly filled accom-
modation space (Figure 24). An overall increasing ac-
commodation space trend that is matched by sedi-
ment fill at the beginning of a cycle set resulted in
shallowing to near sea level, resulting in the last few
cycles being deposited in higher energy within fair-
weather-wave base. In the lower part of the reservoir,
in a deeper shallow-marine setting, shallowing upward
only extended into the influence of storm-wave base
(Figure 10).
High-frequency Sequences
In the Shedgum Arab-D core on display, eight HFSs
have been identified; the top of a possible ninth HFS
occurs at the base of the cored interval (Figures 8 [the
foldout located in the back of this publication] and
24). Because the entire reservoir interval was depos-
ited in an overall shallowing-upward depositional sys-
tem with progressive loss of accommodation space,
each successive HFS has a somewhat different and,
therefore, unique, lithofacies succession from those
above and below it.
The general stacking pattern of HFSs is a transgres-
sive systems tract (TST) of mud-rich cycles and a high-
stand systems tract (HST) of grain-rich cycles, with the
maximum flooding surface (MFS) at the top of the
mud-rich or muddier cycle (Figures 8 [the foldout
located in the back of this publication], 10, 24). In the
lower part of the Arab-D reservoir (zone 3), HFSs are
46 – 52 ft thick, with thick transgressive systems tracts
(TSTs) and thin highstand system tracts (HSTs). Toward
the top of the lower part of the reservoir, transgressive
systems tracts (TSTs) become thinner, and highstand
system tracts (HSTs) thicken in an expected trend. In
the upper part of the Arab-D reservoir (zones 2 and 1),
HFSs progressively thin upsection from 46, 40, 29, 19,
and finally, to 9 ft (14, 12, 9, 6, and finally, to 2.7 m)
(Figure 8, the foldout located in the back of this publi-
cation). This thinning of HFSs demonstrates an overall
shallowing-upward history and base-level fall with the
progressive loss of accommodation space. Transgres-
sive systems tracts (TSTs) are initially thick and drama-
tically thin, whereas HSTs become the thicker and
dominant overall deposition in the ramp crest (Figure 8,
the foldout located in the back of this publication).
In the lower part of the reservoir, transgressive sys-
tems tracts (TSTs) tend to be composed of outer-ramp
and distal middle-ramp, mud-rich cycles (Figure 10).
At the base of the cored interval, HSTs are composed
of thin, storm-generated rudstone and floatstone with
a mud-rich matrix (Figure 8, the foldout located in the
back of this publication). Upsection, transgressive sys-
tems tracts (TSTs) become less mud rich and more
grain rich and are composed of distal and proximal
middle-ramp cycles (Figures 8 [the foldout located
in the back of this publication], 10, 24). Highstand
systems tracts are thicker, grain rich, and cross-bedded
as they shallow upward into fair-weather-wave base
and develop a ramp-crest shoal (Figure 8, the foldout
120 / Lindsay et al.
located in the back of this publication). Two paleosol
horizons (described below) have been identified in
the lower part of the reservoir (zone 3), with no pa-
leosols identified in the upper part of the reservoir
(zones 1 and 2) (Figure 23). In the ramp-crest shoal,
the HSTs become thick and dominated deposition in
the upper part of the reservoir (Figure 8, the foldout
is located in the back of this publication).
Base-level fall led to the deposition of several
seaward-stepping ramp-crest shoal complexes with
low-angle clinoform geometries. Downlapping clino-
form progradation was both northward and south-
ward from an area in the vicinity of the Shedgum
Arab-D core. Perhaps, several HFSs prograded and
downlapped to fill accommodation space through-
out the length of Ghawar field.
Composite Sequences
Two composite sequences have been identified. The
upper composite sequence forms the Arab-D Member
of the Arab Formation and is herein referred to as the
Arab-D composite sequence and covers a vertical in-
terval from 6809.2 to 6945.75 ft (2075.4 to 2117.06 m)
(zone 1, zone 2a, and the upper half of zone 2b)
(Figures 8 [the foldout located in the back of this
publication] and 24). The top of the composite se-
quence is located at the top of the Arab-D reservoir,
where the contact between the uppermost Arab-D
carbonate and the overlying Arab C-D evaporite forms
the sequence boundary. Along the composite sequence
boundary, uncommon, small, shallow, karst-generated
sink holes (not present in this well) filled with a few feet
of collapse breccia are present. These can be seen in
core, in outcrop, and on FMI Fullbore Formation Mi-
croImager logs (Figure 13). In outcrop, small teepee
structures characterize the top of the Arab-D compos-
ite sequence boundary in Wadi Hilwah (Figure 19).
The Arab-D composite sequence in the Shedgum
Arab-D core consists of 55 cycles (parasequences), 13 cycle
sets (parasequence sets), and 5 HFSs (Figures 8 [the
foldout located in the back of this publication] and 24).
If the minimal amount of time for each cycle to be
deposited is 20,000 yr (assuming association with
20,000-yr astronomical precession cycle), then the
Arab-D Member may represent approximately 1.1 m.y.
of time. Most of the Arab-D Member deposition was
in the ramp-crest shoal, which is composed of approxi-
mately 22 cycles, 5 cycle sets (parasequence sets), and
2 HFSs, with a hardground separating the ramp crest
from the inner ramp (Figure 24). The inner ramp con-
sists of five cycles that compose one cycle set that
forms a 9-ft (2.7-m)-thick, potential, HFS (Figure 24).
The remaining 28 cycles, incorporated into 7 cycle sets
making up 2 HFSs and part of a third, are composed
of proximal middle-ramp cycles, with only one distal
middle-ramp cycle and no outer ramp cycles (Figure 24).
The lower composite sequence is the upper part of
the Jubaila Formation (6945.75-ft [2117.06-m] core
depth to below the base of the core — lower half of
zone 2b and all of zone 3) (Figures 8 [the foldout lo-
cated in the back of this publication] and 24). This
composite sequence is referred to as the Jubaila com-
posite sequence. The top of the composite sequence was
briefly subaerially exposed and contains uncommon
meniscus and flowstone (gravity) cements. Two HFSs
in other cored wells in the southern part of Ghawar
shallowed upward and developed short-lived paleosols.
One paleosol is a Psamment paleosol, with plant root-
lets that were in a shoreface to localized island setting
and the other deeper in the section contained man-
grovelike rootlets in a paleosol at the zone 3a–3b contact
(paleosols were identified by G. Retallack, 2003, 2004,
personal communication). The lower half of the Jubaila
composite sequence also contains a minor amount
of windblown quartz silt (<1%) discussed below.
The Jubaila composite sequence in the example
core consists of 65 cycles, 15 cycle sets, and 4 HFSs
(Figure 24). The ramp-crest shoal is composed of only
three cycles within one cycle set in the uppermost
HFS (Figures 8 [the foldout located in the back of this
publication] and 24). No inner-ramp lagoon and in-
tertidal island setting was deposited in this vicinity.
However, sediments containing the shallow-marine
benthonic foraminifera Trocholina alpina (Figures 15, 16)
were reworked from a nearby lagoon and mixed into
the ramp crest at this level. Outer-ramp and distal
middle-ramp depositional settings represent most of
the deposition, with 55 cycles in 10 cycle sets that
are in 2 HFSs (Figures 8 [the foldout located in the
back of this publication] and 24). The proximal mid-
dle ramp is represented by 12 cycles and 3 cycle sets
in the uppermost HFS (Figures 8 [the foldout located
in the back of this publication] and 24).
Meyer et al. (1996, 2000) and Al-Dhubeeb (2005)
used biostratigraphy to correlate the contact between
the Arab-D Member of the Arab Formation and the
underlying Jubaila Formation in outcrop sections in
the Tuwaiq escarpment into the subsurface through
the Khurais and Ghawar fields. Hughes (2004) de-
fined nine paleonvironmentally influenced biozones
(D1 to D9) in Ghawar field (Figure 25). Al-Dhubeeb
(2005) has demonstrated that three of those biozones
(D3 to D5) are not represented in Khurais field, and
that three slightly younger zones (D5 to D7) are not
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 121
represented in the Tuwaiq escarpment outcrop west
of Riyadh (Figure 25). In outcrop, the Arab-D Member
is composed of inner-ramp lagoonal deposits of thin,
platy, mud-rich beds, which rest upon reworked prox-
imal middle-ramp stromatoporoid and coral rudstone
and floatstone debris flows of the Jubaila Formation
(Figure 19) (Meyer, 2000). In outcrop, the composite
sequence boundary separating the overlying Arab-D
Member and the underlying Jubaila Formation is lo-
cated at a datum where stromatoporoids and corals
disappear upsection and are overlain by thin, platy,
and muddy inner-ramp lagoon beds (Figure 19) (Meyer
et al., 1996; D. Vaslet, 2002, personal communica-
tion). In Ghawar field, with the complete biostrati-
graphic succession and complete inner-ramp to outer-
ramp succession present, identifying the composite
sequence boundary is more difficult. However, the
same general criteria, upward decrease in the abun-
dance and consistent presence of stromatoporoids
and corals, coupled with an increase in abundance of
the branched stromatoporoid Cladocoropsis, can be
used to identify this boundary.
Although the top of the Jubaila composite se-
quence is difficult to identify, a few additional pieces
of information have helped identify the composite
sequence boundary. First, at one outcrop locality in
the Tuwaiq escarpment, 1 m (3 ft) of sandstone was
found resting upon the composite sequence bound-
ary that separates the Jubaila Formation from the
FIGURE 25. Biostratigraphic
zonation of the Arab-D Mem-
ber and upper Jubaila Forma-
tion from outcrops in the
Tuwaiq escarpment into the
subsurface through Khurais
and Ghawar fields. Three
zones belonging to the Arab-D
Member are missing in out-
crop. In Khurais field, three
zones are also missing (upper
and middle in the Arab-D and
the lower from uppermost
Jubaila). In Ghawar field, all
zonations are present. The
distance from the Tuwaiq es-
carpment to Ghawar field is
300 km (186 mi). From
A. G. Al-Dhubeeb (2005, per-
sonal communication).
overlying Arab-D Member
(Steineke et al., 1958). In the
subsurface in Ghawar field,
thin sections from cored
wells contain quartz silt in the upper part of zone 2b
(Figure 8, the foldout located in the back of this
publication). The quartz silt is interpreted to have
been reworked by eolian processes off the subaerially
exposed western shelf (Tuwaiq escarpment area) Ju-
baila composite sequence boundary and blown into
the basin (Figure 26). In the Arabian intrashelf basin,
the ultimate resting site for the windblown silt is
within the transgressive systems tract (TST) of the
first HFS of the Arab-D composite sequence. This HFS
is interpreted to onlap onto the Jubaila composite
sequence boundary regionally, but did not toplap it
(Figures 8 [the foldout located in the back of this
publication] and 27). The second HFS of the Arab-D
composite sequence only received windblown silt in
the lower part of the transgressive systems tract (TST)
(Figure 8, the foldout located in the back of this
publication). This second HFS is interpreted to have
also initially onlapped onto the Jubaila composite
sequence in the vicinity of the present-day Tuwaiq
escarpment and, during maximum flooding (MFS),
toplapped the composite sequence (Figure 27). It has
been estimated that approximately 47 ft (14 m) of
topographic relief existed between the subject well in
the Arabian intrashelf basin and outcrops in the
Tuwaiq escarpment (Figure 27). The topographic re-
lief implied by this onlap-toplap series of events is in-
terpreted to have been 14 m (47 ft) because the first
two HFSs of the Arab-D Member (some 14 m [47 ft]
122 / Lindsay et al.
thick in the Shedgum Arab-D core) onlapped and even-
tually toplapped onto the Jubaila composite sequence
boundary in the Tuwaiq escarpment (Figure 27).
Based on this interpretation, the onlapping, HFSs
imply that the western shelf in the Tuwaiq escarp-
ment was subaerially exposed. The interpretation of
subaerial exposure is further supported by several fea-
tures. First, deep incisions have been found in the
uppermost Jubaila shelf margin in an outcrop in the
vicinity of the Diplomatic Quarter (DQ) on the south-
west side of Riyadh. The last and deepest incision
forms the top of the Jubaila Formation and the top of
the Jubaila composite sequence (Figure 28). These
observations could easily have been produced under
FIGURE 27. Conceptual model of basal Arab-D Member onlap and toplap onto the Jubaila composite sequence
boundary, using the Shedgum Arab-D core (Ghawar field) in the Arabian intrashelf basin compared to outcrops in the
Tuwaiq escarpment, west of Riyadh, a distance of 310 km (189 mi).
FIGURE 26. Windblown quartz
and dolomite silt set in a do-
lomitized matrix that contains
intercrystal and moldic porosi-
ty. Quartz silt was blown off
the subaerially exposed Jubaila
Formation composite sequence
boundary (Tuwaiq escarpment)
into the Arabian intrashelf ba-
sin. Black crystals near the top
of the photomicrograph are
pyrite cubes. Photomicrograph
width is 0.55 mm (6903.7 ft),
Shedgum Arab-D core.
highstand conditions of the
Jubaila composite sequence
as overall accommodation
space was lost. Second, Powers
et al. (1966) described a mud-
rich interval in outcrop that
could be regionally mapped
throughout the subsurface.
This interval may represent maximum flooding
(MFS) of the Arabian intrashelf basin during Arab-D
Member deposition and may be the dolomitized in-
terval separating zones 2a and 2b or, more likely, the
next lower, somewhat persistent, dolomitized in-
terval (upper zone 2b) (Figure 8, the foldout located
in the back of this publication). Dolomitization in
Ghawar is generally in mud-rich transgressive sys-
tems tracts (TSTs) (Figure 8, the foldout located in
the back of this publication) (although significant
dolomitization locally occurs in grainy intervals,
most notably in the CLADO facies, where disso-
lution of the Cladocoropsis colonies may produce
superpermeability).
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 123
In outcrop, the upper part of the Jubaila Formation
contains a rich assemblage of reworked stromato-
poroids and corals with only a few whole pieces of
Cladocoropsis present (Meyer, 2000). In the subsurface
in Ghawar field, a rich assemblage of stromatoporoids
and corals is also present, whose upsection give way to
Cladocoropsis bank assemblages that are not repre-
sented in outcrop (Figure 8, the foldout located in the
back of this publication). Stromatoporoid and coral
biostromes and mounds are most common in the up-
per Jubaila Formation, whereas Cladocoropsis banks
are most common in the lower part of the Arab-D
Member. Exclusion of extensive Cladocoropsis in the
Arab-D outcrop can be also explained by unsuitable,
possibly too shallow environmental conditions, such
as inner-ramp deposition.
In the subsurface at Ghawar field, a combination
of information can be used to identify the top of the
Jubaila composite sequence boundary: (1) the strat-
igraphic level above which stromatoporoids and cor-
als are less abundant; (2) the stratigraphic level above
which Cladocoropsis is more abundant; (3) the interval
through which windblown quartz silt is present in the
overlying onlapping, HFSs; and (4) the stratigraphic
level above which overlying transgressive mud-rich
intervals have been dolomitized.
ARAB-D LATERAL FACIES VARIATIONS
Although impossible to interpret from a single core
from the Shedgum area, significant lateral variability
in the distribution of lithofacies does occur in the field
(Figure 29). Overall, a general pattern of thinning of the
carbonate section and thickening of overlying evapo-
rite occurs from north to south across Ghawar. This
FIGURE 28. Photographs of
the DQ section along the Makkah
Highway, west side of Riyadh,
where the Jubaila composite se-
quence boundary, three quarters
up the road cut, has been
downcut into by erosion and
channelized. Storms have re-
worked corals and stromatopo-
roids along the boundary and
into the channels. About 3 – 5 m
(10 – 15 ft) of Arab-D overlies
the Jubaila at this section. Ver-
tical height of the bottom pho-
to is 4.6 m (15 ft).
thickness change has been
alternatively interpreted as a
facies change from carbonate
to evaporite at the top of the Arab-D reservoir (Mitchell
et al., 1988) or as a transition from a paleotopographic
high in northern Ghawar into an intrashelf basin to the
south (Handford et al., 2002).
In addition, smaller scale heterogeneity occurs
internally within the Arab-D reservoir and varies in
a systematic way both temporally and geographically
in the field. Lateral variability is least pronounced
in the lower Arab-D reservoir, where meter-scale,
coarsening-upward cycles consist of transgressive mi-
critic wackestone and mudstone overlain by storm-
transported, bivalve-coated grain-intraclast (BCGI)
rudstone and floatstone and occasionally by stroma-
toporoid-red algae-coral (SRAC) deposits. They repre-
sent either episodic storm deposition or cyclic varia-
tion in sea level. The facies contained therein appear
to be relatively continuous and create a highly strati-
fied reservoir interval (Mitchell et al., 1988) that are
interpreted to very gently downlap. In contrast, the
middle and upper Arab-D reservoir is characterized
by a much greater diversity of facies that tend to
lack lateral continuity. Locally, thick accumulations
of stromatoporoid-red algal-coral rudstone and float-
stone with an internal matrix of mud-rich to mud-lean
packstone and grainstone are present and have been
interpreted as local carbonate buildups or mounds
(Mitchell et al., 1988; Handford et al., 2002) or as a
stromatoporoid bank (Hughes, 1996; Meyer et al.,
1996). Skeletal-oolitic (SO) grain-rich shoals cap the
upper Arab-D interval and are thought to have a
shingled geometry, probably derived from the south-
ward progradation in the southern 80% of Ghawar
and northward progradation in the northern 20%.
Highstand systems tracts (HSTs) contain the great-
est lateral variability and exhibit potential downlap
124 / Lindsay et al.

surfaces, whereas transgressive systems tracts (TSTs)

contain the best and most extensive lateral continu-
ity. Variations in continuity and development of
downlapping shingled or clinoform surfaces may be
related to relative sea level rise and fall that are not
necessarily representative of the Late Jurassic green-
house setting. Instead, the Late Jurassic appears to be
intermediate between icehouse and greenhouse con-
ditions (J. F. Read, 2004, personal communication). If
truly greenhouse, tidal-flat caps should be common,

FIGURE 29. South to north cross section of Arab-D reservoir lithofacies. Modified from Mitchell et al. (1988) and Cantrell (2004).
and if icehouse, numerous exposure surfaces should
be common (Read, 2004), although perhaps obscured
if little meteoric water was available during subaerial
exposure. What is described with respect to the Arab-D
reservoir is considered to represent intermediate or
transitional conditions between the greenhouse and
icehouse extremes.

DIAGENESIS

Diagenesis has modified the original sedimentary

textures and fabrics of the Arab-D reservoir and is a
significant factor in shaping the final reservoir quality
of these rocks. Major diagenetic processes active in
the Arab-D reservoir include dolomitization, dissolu-
tion and formation of microporosity, cementation,
and compaction (Mitchell et al., 1988). The most sig-
nificant of these processes are dolomitization, dis-
solution, and microporosity formation because they
created the most variation in reservoir quality.

Dolomitization
Dolomitization has long been recognized as a sig-
nificant component of the Arab-D reservoir in Ghawar
field (Powers, 1962; Mitchell et al., 1988; Meyer et al.,
1996) and contributes to vertical and lateral hetero-
geneity to the reservoir in terms of textural, miner-
alogical, and pore-type alteration (Cantrell et al., 2001,
2004; Swart et al., 2005). At least three dolomite
phases exist in the Arab-D, all of which are petro-
graphically, geochemically, and stratigraphically dis-
tinguishable (Table 1).
Fabric-preserving dolomite (Cantrell, 2004) con-
sists of very fine crystals of dolomite (10 – 50 mm), in
which the original limestone fabric is well preserved.
Fabric-preserving dolomite typically occurs in the up-
permost Arab D (zone 1) as thin, sheetlike, or strati-
form layers that are always intimately associated with
overlying laminated and bedded anhydrite. Reservoir
quality is generally fair to poor, with porosity typi-
cally less than 10% and permeability ranging from a
few millidarcys to a few tens of millidarcys. Because of
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 125
Table 1. Summary of Arab-D reservoir dolomite types.*
Fabric-preserving Dolomite
Style of dolomitization Fabric-preserving
Petrography Cloudy crystals, 10 – 50 mm
in size
Cathodoluminescence Nonluminescent
Blue-light flourescence Flouresces strongly
Distribution Stratiform
Average d18O ratio 2.08% PDB
Average 87Sr/86Sr ratio 0.70681
Time formed Very early, almost
penecontemporaneously(?)
Water compositions Hypersaline
*From Cantrell (2004).
its relatively heavy oxygen isotope values, intimate
association with the overlying anhydrite, and fabric-
preserving nature, fabric-preserving dolomite is in-
terpreted to have formed early in the diagenetic his-
tory of the sediment by dense, highly evaporated,
magnesium-rich brines percolating down into the
underlying sediment from the overlying salina. The
early dolomitization interpretation is in agreement
with many previous studies of fabric-preserving or
mimetic dolomites (Dawans and Swart, 1988; Pleydell
et al., 1990; Land, 1991; Kimbell, 1993), and it is gen-
erally recognized that fabric-preserving texture will
not form if the precursor limestone has already sta-
bilized to low-Mg calcite (Sibley, 1982, 1991). It should
be noted that there is dolomite lower in the section
(zones 3a and 3b) where some aspects of the original
sediment may be interpreted (e.g., grain size and shape,
cross-bed inclination, etc.), but where fine details
are not preserved. These dolomites are included with
the non-fabric-preserving dolomites (NFPs) discussed
below.
Non-fabric-preserving dolomite consists of medium-
size crystals (50 to >150 mm), nonsaddle (nonba-
roque) dolomite, in which the original limestone fab-
ric has been lost, and in many cases, all fabric evidence
has been completely obliterated (Table 1) (Cantrell,
2004). Typically, NFP dolomite occurs as stratiform
layers scattered throughout the reservoir. It gener-
ally has very poor reservoir quality. Intercrystal or
vuggy porosity is typically less than 10%, and perme-
ability is less than 1 md. Locally, intercrystal porosity
NFP Dolomite Baroque
Nonfabric preserving, Obliterative, with saddle-
typically obliterates all shaped crystals that display
traces of the original undulose extinction
limestone fabric
Crystals may be clear or Crystals are typically
internally zoned, 50 to somewhat cloudy,
more than 150 mm in size 100 – 700 mm in size
Nonluminescent Patchy luminescence
Flouresces weakly Flouresces weakly
Stratiform Nonstratiform
2.66% PDB 7.37% PDB
0.70682 0.70712
Early Later
Hypersaline Hot, mineral-laden fluids
from depth
may exceed 20%, and permeability may approach sev-
eral darcys. NFP dolomite contains heavy oxygen iso-
tope values (Table 1). Because of its general geochemical
similarity to fabric-preserving dolomites, it is inter-
preted to have formed from hypersaline fluids derived
from the overlying salina anhydrite. In a few uncom-
mon examples, some NFP dolomite contains light oxy-
gen isotope values, has good reservoir quality (similar
to above), and is thought to represent a transitional
form, with the third dolomite type known as saddle or
baroque dolomite. Strontium isotopic ratios (Table 1)
suggest that both the fabric-preserving and NFP dolo-
mite formed early or shortly after deposition of the
original sediment or during early burial (Cantrell, 2004).
Saddle or baroque dolomite is a coarse-crystal do-
lomite with saddle-shaped crystals displaying undu-
lose extinction in thin section. It is uncommon in the
reservoir and appears to be limited to wells that con-
tain abnormally thick sections of dolomite. In ex-
treme cases, saddle dolomite is vertically pervasive
and crosscuts stratigraphy. Geochemically, saddle do-
lomite is distinctive with high iron and very low
(light) oxygen isotopic compositions (Table 1) and is
interpreted to have formed from high-temperature
fluids during burial diagenesis. Reservoir quality is
fair to poor. The interpretation that saddle dolomite
formed relatively late in the diagenetic history of a
rock from high-temperature fluids during deeper
burial diagenesis is consistent with many previous
studies (Folk and Assereto, 1974; Radke and Mathis,
1980) and is in agreement with the fluid-inclusion
126 / Lindsay et al.
results from the Arab-D (mean homogenization tem-
perature = 97.48C, Table 1) (Cantrell, 2004).
In addition, quantification and mapping of the
dolomite content of the Arab-D across Ghawar re-
vealed that dolomite does not occur randomly or
uniformly across the field. Fabric-preserving and most
NFPs typically occur as bed-parallel, thin, lenticular
or sheetlike units that range in thickness from 0.3
to 4.6 m (1 to 15 ft). These stratiform dolomite bod-
ies commonly occur within fairly well-defined strati-
graphic intervals in the Arab-D (Figure 30A). As such,
these dolomites may vertically stratify the reservoir
and result in local impediments to vertical flow. Note,
however, that thin intervals (locally as little as 0.3 m
[1 ft] thick) of high-porosity and high-permeability
NFP dolomite may contribute a high proportion of
the flow from an individual well, creating a high-flow
or super-K zone.
Detailed core descriptions throughout the field
have begun to show a preference for dolomitization
to be more pervasive in the transgressive systems tract
(TST) and at the maximum flooding surface (MFS),
with less dolomite associated with the HST. The ap-
parent reason for this phenomenon is that the TST
FIGURE 30. Cross sections con-
trast the lateral persistence and
continuity of stratigraphic dolo-
mite (section A) with vertically
pervasive, nonstratigraphic dolo-
mite (section B). Modified from
Cantrell et al. (2004).
is more mud rich. Descending
brines sourced from the over-
lying C-D evaporite will per-
colate through the more grain-
rich HST and be dramatically
slowed to preferentially dolo-
mitize the TST.
In contrast, the more un-
common saddle dolomite and
some isotopically light forms
of NFP dolomite (more coarse-
ly crystalline than most NFP
dolomite) are nonstratiform in
nature, in that they typically
occur as abnormally thick do-
lomitized intervals that cross-
cut stratigraphic reservoir
zones (Figure 30B). Mapping
reveals that this type of dolo-
mite occurs as linear northeast–southwest-oriented
fracture and fault trends across Ghawar (Figure 31).
Dolomite along these trends typically makes up 40–
60% (and locally as much as 100%) of the entire Arab-D
reservoir interval, whereas offtrend dolomite typically
makes up less than 20% of the reservoir. These fracture-
and fault-related saddle dolomite trends appear to be
very limited areally (1 – 2 km [0.6 – 1.2 mi] in width)
but extend obliquely across the field for distances
greater than 60 km (37 mi). Because of their narrow
width (which is on the order of one to two well spac-
ings), these trends are interpreted to be vertical in
orientation and to potentially create horizontally
separated compartments in the reservoir and would
generally act as vertically oriented, elongate, low-
permeability barriers that would drastically effect
horizontal flow in the reservoir. When porous, these
trends of saddle (baroque) dolomite can contain high
permeability and act as sites of high fluid flow, or
super-K, in the reservoir (the occurrence of which is
currently unpredictable).
A model of dolomitization (Figure 32) is proposed
to explain the occurrence of both stratiform and
nonstratiform types of dolomite. Stratiform dolomite
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 127

originated via several mech-

anisms that occurred early in
the diagenetic history of the
sediment and are related to
downward percolation of hy-
persaline brines as the over-
lying C-D evaporite was be-
ing deposited (Figure 32). In
contrast, nonstratiform, pre-
dominantly saddle dolomite
formed from hot, mineral-
laden fluids that ascended
upsection and emplaced in-
to the reservoir from depth
along a series of faults and/
or fractures and preferen-
tially dolomitized the sec-
tion adjacent to faults and
fractures (Figure 32). In some
instances, the fractures were
first solution widened prior
to emplacement of the sad-
dle dolomite. Saddle dolo-
mite has been observed in
faulted blocks of Jubaila –
Arab-D carbonates along the
north flank of Wadi Nisah,
which provides some evi-
dence for fault-related do-
lomitization of the saddle
dolomite.

Implications
With the exception of
saddle dolomite, which is
thought to be associated with
the movement of fluids
along fracture and fault zones
during later burial, the oc-
currence of dolomite in the
Arab-D reservoir reflects spe-
cific physical and chemical
conditions that occurred dur-
ing deposition to early buri-
al of the Arab-D. As a result,
placing these dolomites in
the framework of the overall
depositional and stratigraph-
FIGURE 31. The left is a map of percent dolomite in the Arab-D reservoir, Ghawar field. ic architecture of the Arab-D
The right is a map of percent dolomite in zone 2b, Arab-D reservoir, Ghawar field. is essential to understand the
Modified from Cantrell et al. (2004). proper temporal and spatial
128 / Lindsay et al.

and nearest to the source

of descending brines, where
brine concentrations were
highest and reaction rates
were rapid. In addition, al-
though NFP dolomite occurs
in both Arab-D composite se-
quences, its distribution is
not uniform. No significant
dolomitization of any kind
occurs in the skeletal-oolitic
(SO) grainstone-dominated
upper Arab-D (zone 2A), but
NFP dolomite is common in
the middle part and the up-
per part of the lower Arab-D
reservoir (zones 2B and 3A).
This NFP dolomite occurs in
association with stromato-
poroid-red algal-coral (SRAC)
and Cladocoropsis (CLADO)
rudstone and floatstone (pack-
stone and grainstone matrix)
in the basal part zone 2B and
with mudstone and wacke-
stone interbedded with bi-
valve-coated grain-intraclast
(BCGI) rudstone and float-
stone (grainstone, mud-lean
and mud-rich packstone ma-
trix) in zone 3A. Little or no
dolomite occurs in zone 3B
or below in the northern
half of Ghawar, but some
zone 3B mudstone is dolomi-
tized in the southern half.
Although fabric-preserving
and NFP dolomites are inter-
preted to have similar ori-
gins, their different dolomite
FIGURE 32. Dolomitization model for the Arab-D reservoir in Ghawar field. Modified fabrics reflect their different
from Cantrell et al. (2004). stratigraphic contexts and
potentially slightly different
times of formation. Fabric-
context in which dolostones formed and to provide a preserving dolomite formed early (penecontempora-
predictive framework for the occurrence of similar do- neously), prior to extensive stabilization of the sedi-
lostones in other intervals. ment to low-Mg calcite, whereas NFP formed later
It is clear that the partitioning of dolomite types (but still early, as suggested by the Sr isotope data),
occurs in the Arab-D, with NFP dolomite occurring in subsequent to mineralogical stabilization of the sedi-
both composite sequences of the Arab-D carbonate, ment, but after initial compaction.
whereas fabric-preserving dolomite occurs only in Overall, the vertical distribution of dolomite re-
zone 1, at the top of the upper composite sequence flects both the vertical succession of the Arab-D as well
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 129
as the hydrodynamics of the top-down dolomitiza-
tion mechanism proposed for both types of dolomite.
Dense, hypersaline brines that migrated downward
from the overlying salina initially moved into zone 1
shortly after deposition and produced fabric-preserving
dolomite in unstabilized sediment. As deposition of
salina evaporites continued, brines percolated down-
ward and moved into the older Arab-D reservoir com-
posite sequences below. Initially, these fluids moved
unrestricted through well-sorted skeletal-oolitic (SO)
grainstones in the Arab-D (zone 2A) but slowed sig-
nificantly as they passed through the highly heteroge-
neous, poorly sorted, and more mud-prone sediment-
rock at the zone 2A – 2B contact and in zones 2B
and 3A. There, increased residence time of the fluids
resulted in increased probability of dolomitization
of these mud-prone sediments. In addition, the finer
grain size in the muddier parts (TST) of these sequences
promoted dolomitization by providing a greater num-
ber of nucleation sites for dolomitization (Sibley and
Gregg, 1987) than would the less mud-prone, more
grain-rich sediment-rock above. Ultimately, these
fluids continued moving downsection (into lower
zone 3A and zone 3B) and eventually equilibrated
with the local host sediment-rock-water system and
were no longer capable of dolomitizing. As a result,
the amount of dolomite generally decreases down-
section below the middle Arab-D reservoir.
Dolomitization thus induced heterogeneity in a
variety of ways that are generally predictable in the
Arab-D reservoir. In a stratigraphic sense, dolomite
occurrence is predictable based on the following
criteria: (1) proximity to a thick, laterally extensive
evaporite unit (C-D evaporite), the deposition of which
provided a source of downward-migrating hypersa-
line brines; (2) a major sequence boundary that sepa-
rated the overlying evaporite from underlying car-
bonate; and (3) cyclic carbonates that contain poorly
sorted, mud-prone sediments that are prone to retain
FIGURE 33. Arab-D reservoir
rock classification for Ghawar
field. Modified from Cantrell
and Hagerty (2003).
dolomitizing fluids. The upper
composite sequence bound-
ary at the top of the Arab-D
served to help partition fabric-
preserving from NFP dolo-
mite; mud-rich transgressive
systems tracts in HFSs also
help localize dolomite. The
presence of certain sedimentary fabrics, such as poorly
sorted and mud-prone rocks in these cycles, appears
to act as the primary control over the occurrence of
nonsaddle dolomite. Provided that these three con-
trolling factors are present, it is possible to under-
stand and predict patterns of dolomitization in the
Arab-D, as well as in other similar carbonate-evaporite
cycles that formed elsewhere. Vertically oriented high-
temperature saddle (baroque) dolomite, in fracture-
and fault-related structural trends, across the Arab-D
reservoir have the potential to influence horizontal
flow in the reservoir.
RESERVOIR ROCK CLASSIFICATION
An integrated petrographic and petrophysical
study of Arab-D carbonates in the Ghawar field
has provided a new carbonate reservoir rock clas-
sification (Figure 33) (Cantrell and Hagerty, 2003).
Each reservoir rock type recognized in this classi-
fication has a distinct pore network as defined by
porosity-permeability relationships and pore-throat
radii connectivity expressed as pore-throat radii size
distribution and J-function curves. This classification
divides Arab-D carbonates into seven limestone and
four dolomite rock types (Figure 33). The amount of
matrix (lime mud) and pore types are the primary de-
fining parameters for limestones. Reservoir rock type
for dolomite is based on crystal texture. Note that
numerous pigeon holes in the classification scheme
(Figure 33) have been left blank, indicating that those
criteria are of no significance in defining that par-
ticular rock type.
The seven limestone reservoir rock types are based
on the values of five petrographic parameters, ap-
plied in the following order (that is, starting from
the top of the table in Figure 33 and working toward
the bottom): (1) abundance of cement; (2) abun-
dance of matrix (lime mud); (3) sorting (Krumbein and
130 / Lindsay et al.
Pettijohn, 1988); (4) dominant pore type; and (5) size of
the largest molds. Of the five, the amount of matrix is
the most important parameter, although the amount
of cement is the first criterion for subdividing the
limestone rock types (Figure 33). In general terms, six
of these seven types fall into two broad families, A
and B, each of which can then be subdivided into
three members (types I, II, and III) according to their
matrix content. The first family, A, is a fairly coarse-
grained, poorly sorted rock with relatively large molds.
The second family, B, is a generally fine- to medium-
grained, well-sorted rock with few or small molds. The
seventh rock type contains more than 10% cement,
which modifies the pore-size distribution enough to
warrant a separate reservoir rock type. Each of the
reservoir rock types exhibits a distinctive pore-throat
size distribution and, in turn, a Leverett J-function.
The seven types are also characterized by distinctive
porosity-permeability relationships (Figures 34, 35).
FIGURE 34. Typical core-plug photo-
graph examples of the major Arab-D
reservoir limestone rock types. Divi-
sions at top of photographs are in
millimeters. Modified from Cantrell
(2004).
The four dolomite reservoir rock
types are classified according to do-
lomite crystal texture, although strati-
graphic position and abundance of
porosity can also be effective in their
classification. The four textures are
fabric preserving (Vfp), sucrosic (Vs),
intermediate (Vi) and mosaic (Vm).
As noted previously, the Vfp dolo-
mite is only found in zone 1 of the
Arab-D, where it is the major dolo-
mite type. Vs dolomite occurs in
dolomites with more than 12% po-
rosity; Vm is less than 5% porosity,
and Vi is between 5 and 12% porosity.
Vfp dolomites have pore systems sim-
ilar to their precursor limestone, but
the pore systems of the other dolo-
mite types are unique.
Implications
The distribution of these rock
types is not random in the reservoir
and reflects the stratigraphy and
vertical succession of rock types previously described
in the Arab-D reservoir (Figure 8, the foldout located
in the back of this publication). As noted previously,
zone 2A is dominated by relatively uniform, well-
sorted, skeletal-oolitic grainstone that falls into res-
ervoir rock type IB, whereas more heterogeneous
Cladocoropsis (CLADO) and stromatoporoid-red algal-
coral (SRAC) rudstone and floatstone that contain a
matrix of packstone and grainstone occur in zone 2B
and fall into reservoir rock type IA or IIA. Zone 3A
tends to be dominated by heterogeneous and rela-
tively mud-prone, bivalve-coated grain-intraclast
(BCGI) rudstone and floatstone that contain a ma-
trix of packstone that tend to fall in reservoir rock
type IIB, whereas zone 3B typically contains abun-
dant micritic wackestone to mudstone that would
be classified in reservoir rock type IIIB. The occurrence
of these reservoir rock types is thus predictable in
the reservoir.
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 131
This new classification scheme is significant, in
that it focuses on observed heterogeneities in the
rocks and groups them according to impact on flow
properties. Thus, this classification is quite different
from other, more widely used carbonate classifica-
tions. Most other classifications have relied on either
textural (Dunham, 1962; Powers, 1962; Lucia et al.,
2001) or lithofacies information (Mitchell et al., 1988;
Meyer and Price, 1993). Figure 36 compares two stan-
dard classifications with the scheme used in this study.
Although collection of texture and lithofacies infor-
mation can provide the information required to de-
velop the depositional and stratigraphic framework
of the reservoir and its gross reservoir quality, it does
not provide sufficient information to define pore
network types in the reservoir. These textural and
lithofacies classifications only indirectly relate to
pore types and, in most instances, are unsuccessful
in subdividing the reservoir rocks into groupings that
are significant relative to their flow characteristics.
The recognition and codification of the present clas-
sification, producing reservoir rock groupings that
have significantly similar flow characteristics, in-
vites development of a coordinated scheme for the
prediction of occurrence of these groupings in the
reservoir.
FIGURE 35. Porosity-perme-
ability crossplots of individual
limestone reservoir rock types.
Outliers are labeled according
to their nonuniformity or tex-
tural characteristics (Bu = bur-
rows; Cg = coarse grained;
Fr = microfractures; Tx = tex-
tural variations). Green are for
zones 1 and 2, and red are
for zones 3 and 4. Modified
from Cantrell (2004).
RESERVOIR
CHARACTERIZATION
USING IMAGE LOGS
Background
Understanding the hetero-
geneity of porosity and per-
meability is critical to evalu-
ating production behavior of
Ghawar field (Figure 37). De-
velopment scenarios have
been implemented that em-
ploy long-radius, multilater-
al horizontal wellbores, termed ‘‘maximum reservoir
contact’’ (MRC) wells, which target the uppermost
Arab-D reservoir while avoiding deeper potential high-
permeability zones. Because nearly all of these wells
are not cored, few available data characterize reser-
voir quality in multilateral wells, with the exception
of image and conventional logs.
Several challenges exist to characterizing reservoir
porosity and permeability variations in the Ghawar
field. First, in some instances, a complex secondary
diagenetic overprint (such as dolomitization) has altered
the primary texture. Quantifying petrophysical prop-
erties of the reservoir and correlation of wire-line-log
signatures to depositional facies become challenging.
In these instances, porosity and permeability are con-
trolled much more by diagenesis than by deposition-
al facies, although depositional facies influenced the
degree of diagenetic overprint. Second, conventional
wire-line logs lack the fundamental vertical and azi-
muthal resolution to provide an adequate characteri-
zation of the reservoir heterogeneity. Significant tex-
tural and stratigraphic variations between adjacent
thin layers hold the key to understanding the sequence
stratigraphy. Third, standard core plugs, taken at ver-
tical intervals of 6 in. (15 cm) or more, may not be
adequate to characterize the reservoir quality.
132 / Lindsay et al.
FIGURE 36. Distribution of limestone samples by standard
carbonate rock classifications and reservoir rock types.
Dunham textures: GRN = grainstone; MLP = mud-lean
packstone; PCK = packstone; WCK = wackestone; MUD =
mudstone. Depositional facies: SO = skeletal-oolitic; MIC =
micritic; BCGI = bivalve-coated grain-intraclast; SRAC =
stromatoporoid-red algae-coral; CLAD = Cladocoropsis.
Modified from Cantrell (2004).
Methodology
Carbonate rocks are characterized by complex po-
rosity systems (Lucia, 1995; Kazatchenko and Mousa-
tov, 2002). Small-scale heterogeneity of porosity and
permeability can only be resolved in wellbore logs with
image log data (Newberry et al., 1996). Microresistivity
image logs have ideal, small-scale vertical resolution as
well as azimuthal borehole coverage for detailed res-
ervoir heterogeneity analysis. Core data are necessary
to calibrate all the log results, especially in heteroge-
neous rocks. In wells without borehole images, open-
hole lithology logs are required to determine reservoir
rock types for heterogeneity analysis.
Image logs that detect relative microresistivity
changes (e.g., pores invaded by conductive drilling
mud that contrast with resistive rock matrix) in the
wellbore may aid in the identification of rock tex-
ture. Therefore, image logs have been used for various
quantitative analyses in carbonate reservoir charac-
terization, such as partitioning of matrix and vuggy
porosity through borehole porosity image analysis
(Newberry et al., 1996; Chitale et al., 2004), vug con-
nectivity through image texture analysis leading to
permeability prediction (Russell et al., 2002, 2004),
and extrapolation of image log-derived permeability
data into wells without image logs using sonic-derived
vuggy porosity measurements (Xu et al., 2006).
Porosity and permeability in vuggy carbonate res-
ervoirs is quantified through an optimized workflow
of borehole image and conventional log processing
that is calibrated to core data. Permeability from cores,
which commonly has a poor correlation with con-
ventional log-derived permeability, is found to have
an exponential correlation to the vug porosity com-
ponent of the total porosity computed from borehole
image porosity analysis.
Results
Primary depositional texture exerts a strong in-
fluence on reservoir quality in the Arab-D because
of the effect of lime mud on permeability. Mud-
dominated textures, such as mudstone and wacke-
stone, have generally low porosity and permeability,
with the exception of diagenetically altered dolo-
stone that has relatively high permeability in in-
termediate porosity. Packstone and floatstone have
higher porosity and permeability, and grain-supported
textures, such as grainstone and rudstone, have the
best reservoir quality. Traditional log-based porosity-
permeability transforms do not adequately predict
permeability in different textures with similar total
porosity values because the vuggy porosity com-
ponent of the total porosity is not measured on con-
ventional wire-line logs, which provide porosity
values averaged more than 0.6 – 1.2 m (2 – 4 ft) ver-
tically. Stromatoporoid floatstone and rudstone
contain significant vuggy porosity in image logs,
and the clearly multimodal porosity spectrum mea-
sured azimuthally in image logs contributes to anom-
alously high permeability that matches production
data. Sucrosic dolostone, commonly characterized by
large intercrystalline vugs, accounts for extremely
high flow from very thin intervals (<0.3 m [<1 ft]
thick) because of permeability that is measured from
image log analyses at two to three orders of magnitude
higher than conventional log transforms. This dramatic
increase of permeability with vug porosity partition-
ing without total porosity change can result in the
inverse correlation between porosity and permeabil-
ity (Russell et al., 2002). Conventional wire-line-log
porosity-permeability transforms can neither predict
nor resolve such heterogeneity of permeability. Fi-
nally, well-sorted, highly permeable oolitic, peloidal
grainstones of the upper Arab-D are commonly very
thin (less than 0.3 m [1 ft]) and account for high flow
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 133

FIGURE 37. Cross-bedded peloid grainstone in zone 2A. Image logs resolve thin beds of grainstone (black lines on left)
and commonly show hints of cross-bedding (as indicated by the enlarged view of the red outlined interval shown
below).

rates (Figure 37). These grainstones are well sorted
and exhibit a high degree of interconnectedness of
intergranular pores. Image-derived permeability data
strongly correlate with core-derived permeability in
these facies. Image logs clearly resolve the thin grain-
stone layers, and calibration of image facies with high-
resolution core description data provides a method-
ology for extrapolating grainstone layers into uncored
wells using image logs. A methodology is shown for
predicting flow behavior from image logs alone using
image-derived permeability. After calibration in verti-
cal cored wells, the classification of image textures and
facies is the basis for estimating grainstone, stroma-
toporoid, and dolostone facies, cyclicity, and their
associated permeabilities in image logs from hori-
zontal and uncored wells.
CONCLUSIONS
1) The Arab-D reservoir in the Ghawar field was de-
posited on a tropical carbonate ramp that ex-
tended from outer ramp (below fair-weather-wave
base) to a depth of approximately 30 m (100 ft)
water depth characterized by micritic cycles with
Thalassinoides burrows. Updip, the distal middle
ramp is characterized by micritic cycles capped
by Thalassinoides firmgrounds with overlying float-
stone to rudstone of dominantly bivalve-coated
grain-intraclast sediment. The proximal middle
ramp is characterized by domed and encrusting
stromatoporoid and coral biostromes and mounds
with delicately branched Cladocoropsis stromato-
poroids in sheltered areas. The ramp-crest shoal
is characterized by skeletal-oolitic grainstone to
packstone, and the inner-ramp lagoon is char-
acterized by more micritic sediment with local
intertidal islands.
2) The reservoir is made up of two composite se-
quences. The lower composite sequence is the
upper part of the Jubaila Formation. The se-
quence boundary occurs at the top of the ramp-
crest shoal, which is overlain by middle-ramp
lithofacies. The upper composite sequence is the
D member of the Arab Formation, the sequence
134 / Lindsay et al.
boundary at its top marked by local karst-collapse
breccias and flooded over by subaqueous evapo-
rites interpreted as the TST of the following com-
posite sequence.
3) Major diagenetic features of the Arab-D reservoir
include dolomitization, dissolution, cementation,
and compaction. Dolomitization was mostly
non-fabric-preserving and stratiform, occurring
in more micritic TST sediments and probably
formed by downward seepage of Mg-rich brines
during deposition of the overlying C-D anhydrite.
Locally, saddle dolomite occurs in bodies that
crosscut stratigraphy, perhaps resulting from
late fluids moving up local fracture zones. Dis-
solution is widespread, especially in the grainier
parts of the reservoir, but the cause is less obvious
given the evaporitic nature of the overlying hy-
drographic system.
4) Porosity includes interparticle, moldic, intrapar-
ticle, and microporosity. Permeability is mostly
dominated by interparticle pores, which are typi-
cally only lightly cemented. Although moldic pores
are abundant, they typically only contribute sig-
nificantly to permeability in dolomite, where the
delicately branched stromatoporoid Cladocoropsis
has been leached and the matrix has been re-
placed by tightly intergrown dolomite. Dolomite
is porous to nonporous (tight), with porosity a
mixture of moldic, intercrystal, and intracrystal
pores.
5) Image logs, in conjunction with other logs, show
significant promise for the identification of litho-
facies, especially coarse-grained lithofacies, in the
Arab-D reservoir. Image log data clearly resolve
thin layers of connected vugs in sucrosic dolo-
stone and stromatoporoid rudstone and float-
stone that contribute to high permeability and
flow rates. Permeability from image log analyses
correlates strongly with core permeability, yet
image-derived permeabilities are two to three or-
ders of magnitude higher than conventional log
transforms in intervals with high vuggy poros-
ity. Predicted image-derived flow behavior cor-
relates strongly with actual flowmeters.
ACKNOWLEDGMENTS
The authors thank the management of Saudi
Aramco for granting permission to publish this con-
tribution. In particular, the authors thank Abdulla
Al-Naim, Aboud Al-Afifi, Sa’id Al-Hajri, Adnan Al-
Sharif, and George Grover for their support and review
of the manuscript. We thank Nassir Alnaji, David
Bacchus, and the cartography section of Saudi Aramco
for their help with the creation of the digital and
scanned images.
REFERENCES CITED
Al-Dhubeeb, A. G., 2005, Biofacies as a tool for calibrat-
ing the Jurassic Jubaila – Arab formational contact
from outcrop in Riyadh area to subsurface in eastern
Saudi Arabia: Master’s thesis, King Fahd University
of Petroleum and Minerals, Dhahran, Saudi Arabia,
219 p.
Al-Hinai, K. G., A. E. Dabbagh, W. C. Gardner, M. Khan,
and S. Saner, 1997, Shuttle imaging radar views of
some geological features in the Arabian Peninsula:
GeoArabia, v. 2, p. 165 – 178.
Al-Husseini, M. I., 1997, Jurassic sequence stratigraphy of
the western and southern Arabian Gulf: GeoArabia,
v. 2, p. 361 – 382.
Al-Husseini, M. I., 2000, Origin of the Arabian plate struc-
tures: Amar collision and Najd rift: GeoArabia, v. 5,
no. 4, p. 527 – 542.
Arkell, W. J., 1952, Jurassic ammonites from Jebel Tuwaiq,
central Arabia: Royal Society of London Philosophical
Transactions, Series B, v. 236, p. 241 – 313.
Ayres, M. G., M. Bilal, R. W. Jones, L. W. Slentz, M. Tartir,
and A. O. Wilson, 1982, Hydrocarbon habitat in main
producing areas, Saudi Arabia: AAPG Bulletin, v. 66,
p. 1 – 9.
Barger, T. C., 2000, Out in the blue: Vista, Selwa Press,
320 p.
Bates, B. S., 1973, Oscar for an oilfield: Saudi Aramco World,
v. 24, no. 6, p. 14 – 15.
Beydoun, Z. R., 1988, The Middle East: Regional geology
and petroleum resources: Beaconsfield, United King-
dom, Scientific Press, 292 p.
Beydoun, Z. R., 1991, Arabian plate hydrocarbon geology
and potential — A plate tectonic approach: AAPG
Studies in Geology 33, 77 p.
Beydoun, Z. R., 1998, Arabian plate oil and gas: Why so
rich and so prolific?: Episodes, v. 21, p. 74 – 81.
Cantrell, D. L., 2004, Carbonate heterogeneity during
global greenhouse time: Examples from the Jurassic of
the Middle East: Ph.D. thesis, University of Manchester,
Manchester, England, United Kingdom, 368 p.
Cantrell, D. L., and R. M. Hagerty, 2003, Reservoir rock
classification, Arab-D reservoir, Ghawar field, Saudi
Arabia: GeoArabia, v. 8, p. 435 – 462.
Cantrell, D. L., P. K. Swart, C. R. Handford, C. G St. C.
Kendall, and H. Westphall, 2001, Geology and produc-
tion significance of dolomite, Arab-D reservoir, Ghawar
field, Saudi Arabia: GeoArabia, v. 6, p. 45 – 60.
Cantrell, D. L., P. K. Swart, and R. M. Hagerty, 2004,
Genesis and characterization of dolomite, Arab-D res-
ervoir, Ghawar field, Saudi Arabia: GeoArabia, v. 9,
p. 11 – 36.
Carrigan, W. J., G. A. Cole, E. L. Colling, and P. J. Jones, 1994,
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 135
Geochemistry of the Upper Jurassic Tuwaiq Mountain
and Hanifa formations petroleum source rocks of east-
ern Saudi Arabia: in B. J. Katz, ed., Petroleum source
rocks, casebooks in earth sciences series: Berlin, Springer-
Verlag, p. 67–87.
Champetier, Y., and E. Fourcade, 1966, A propos de
Cladocoropsis mirabilis Felix dans le Jurassique superieur
de Sud-Est de l’Espagne: Estudios Geologicos, v. 22,
p. 101 – 111.
Chitale, D. V., J. Quirein, T. Perkins, G. B. Lambert, and
J. C. Cooper, 2004, Application of a new borehole
imager and technique to characterize secondary po-
rosity and net-to-gross in vugular and fractured car-
bonate reservoirs in Permian basin: 45th Society of
Professional Well Log Analysts Annual Logging Sym-
posium, Noordwijk, June 6 – 9, 2004.
Dawans, J. M., and P . K. Swart, 1988, Textural and geochem-
ical alternations in late Cenozoic Bahamian dolomites:
Sedimentology, v. 35, p. 385 – 404.
De Castro, P., 1991, On the cortical layer of Thaumato-
porellaceans (green algae) (abs.): 5th International
Symposium on Fossil Algae, Capri, April 7 – 12, p. 16 –
17.
De Castro, P., 2002, Thaumatoporella parvovesiculifera
(Raineri): Typification, age and historical background
(Senonian, Sorrento Peninsula, southern Italy): Bollet-
tino della Societa Paleotologica Italiana, v. 41, no. 2 – 3,
p. 121 – 129.
De Matos, J. E., and R. F. Hulstrand, 1995, Regional
characteristics and depositional sequences of the
Oxfordian and Kimmeridgian, Abu Dhabi, in M. I. Al-
Husseini, ed., Middle East Geosciences Conference,
GEO’94: Bahrain, Gulf PetroLink, v. 1, p. 346 – 356.
Droste, H. H. J., 1990, Depositional cycles and source rock
development in an epeiric intra-platform basin, the
Hanifia Formation of the Arabian Peninsula: Sedimen-
tary Geology, v. 69, p. 281 – 296.
Dunham, R. J., 1962, Classification of carbonate rocks
according to depositional texture, in W. E. Ham, ed.,
Classification of carbonate rocks: AAPG Memoir 1,
p. 108 – 121.
Durham, L. S., 2005, Saudi Arabia’s Ghawar field: The ele-
phant of all elephants: AAPG Explorer, January 2005,
p. 4 and 7.
Enay R., 1987, Le Jurassique d’Arabie Saoudite Centrale:
Geobios Memoir Special 9, 314 p.
Folk, R. L., and R. Assereto, 1974, Giant aragonite rays and
baroque white dolomite in tepee-fillings, Triassic of
Lombardy, Italy (abs.): SEPM Annual Meeting Pro-
grams with Abstracts, v. 164, p. 34 – 35.
Glennie, K. W., 2000, Cretaceous tectonic evolution of
Arabia’s eastern plate margin: A tale of two oceans: in
A. S. Alsharhan and R. W. Scott, eds., Middle East
models of Jurassic/Cretaceous carbonate systems:
SEPM Special Publication 69, p. 9 – 20.
Hancock, P. L., A. Al-Kadhi, and N. A. Sha’at, 1984, Re-
gional joint sets in the Arabian platform as indicators
of intraplate processes: Tectonics, v. 3, p. 27 – 43.
Handford, C. R., D. L. Cantrell, and T. H. Keith, 2002,
Regional facies relationships and sequence stratig-
raphy of a super-giant reservoir (Arab-D Member),
Saudi Arabia, in J. M. Armentrout, ed., Sequence strati-
graphic models for exploration and production: Evolv-
ing methodology, emerging models and application
histories, 22nd Annual Bob F. Perkins Research Confer-
ence: Gulf Coast Section SEPM, p. 539 – 564.
Hughes, G. W., 1996, A new bioevent stratigraphy of late
Jurassic Arab-D carbonates in Saudi Arabia: GeoArabia,
v. 1, p. 417 – 434.
Hughes, G. W., 1997, The Great Pearl Bank barrier of the
Persian Gulf as a possible Shu’aiba analogue: Geo-
Arabia, v. 2, p. 279 – 304.
Hughes, G. W., 2004, Middle to Upper Jurassic Saudi
Arabian carbonate petroleum reservoirs: Biostratigra-
phy, micropaleontology and paleoenvironments: Geo-
Arabia, v. 9, no. 3, p. 79– 114.
Irwin, M. L., 1965, General theory of epeiric clear water
sedimentation: AAPG Bulletin, v. 49, p. 445 – 459.
James, N. P., 1983, Reef environment, in P. A. Scholle,
D. G. Bebout, and C. H. Moore, eds., Carbonate de-
positional environments: AAPG Memoir 33, p. 345 –
440.
Kazatchenko, E., and A. Mousatov, 2002, Primary and
secondary porosity estimation of carbonate formation
using total porosity and the formation factor: Pre-
sented at 2002 Society of Petroleum Engineers Annual
Technical Conference and Exhibition, San Antonio,
Texas, SPE Paper 77787, 6 p.
Keith, T. H., 2005, Finding the super-giant: The discovery
of Ghawar field: GeoFrontier, Dhahran Geoscience
Society, v. 2, no. 1, p. 29 – 44.
Kimbell, T. N., 1993, Sedimentology and diagenesis of late
Pleistocene fore-reef calcarenites, Barbados, West In-
dies: A geochemical and petrographic investigation of
mixing zone diagenesis: Ph.D. dissertation, University
of Texas, Dallas, 296 p.
Kinsman, D. J. J., 1964, Recent carbonate sedimentation
near Abu Dhabi, Trucial Coast, Persian Gulf: Ph.D.
thesis, University of London, London, 302 p.
Konert, G., A. M. Al-Afifi, S. A. Al-Hajri, and H. J. Droste,
2001, Paleozoic stratigraphy and hydrocarbon hab-
itat of the Arabian plate: GeoArabia, v. 6, p. 407 –
442.
Krumbein, W. C., and F. J. Pettijohn, 1988, Manual of
sedimentary petrography: SEPM Reprint Series 13,
549 p.
Land, L. S., 1991, Dolomitization of the Hope Gate
Formation (north Jamaica) by seawater: Reassessment
of mixing-zone dolomite, in H. P. Taylor, J. R. O’Neill,
and I. R. Kaplan, eds., Stable isotope geochemistry: A
tribute to Samuel Epstein: Geochemical Society
(London) Special Publication 3, p. 121 – 133.
Leinfelder, R. R., F. Schlagintweit, W. Werner, O. Ebli,
M. Nose, D. U. Schmid, and G. W. Hughes, 2005,
Significance of stromatoporoids in Jurassic reefs and
carbonate platforms — Concepts and implications:
136 / Lindsay et al.
Facies: International Journal of Paleontology, Sedi-
mentology and Geology, v. 51, p. 287 – 325.
LeNindre, Y.-M., J. Manivit, H. Manivit, and D. Vaslet,
1990, Stratigraphie sequentielle du Jurassique et du
Cretace en Arabie Saoudite: Bulletin de la Societe
Geologique de France, series 8, 6, p. 1025 – 1034.
Lucia, F. J., 1995, Rock-fabric/petrophysical classification
of carbonate pore space for reservoir characterization:
AAPG Bulletin, v. 79, p. 1275 – 1300.
Lucia, F. J., J. W. Jennings, M. Rahnis, and F. O. Meyer,
2001, Permeability and rock fabric from wireline logs,
Arab-D reservoir, Ghawar field, Saudi Arabia: GeoArabia,
v. 6, p. 619–646.
Maiklem, W. R., D. G. Bebout, and R. P. Glaister, 1969, Clas-
sification of anhydrite; a practical approach: Bulletin
of Canadian Petroleum Geology, v. 17, no. 2, p. 194 –
233.
McGuire, M. D., G. Kompanik, M. Al-Shammery, M. Al-
Amoudi, R. B. Koepnick, J. R. Markello, M. L. Stockton,
and L. E. Waite, 1993, Importance of sequence strati-
graphic concepts in development of reservoir architec-
ture in upper Jurassic grainstones, Hadriya and Hanifa
reservoirs, Saudi Arabia: Proceedings of 8th Middle East
Oil Show, Society of Petroleum Engineers Paper 25578,
p. 489 – 499.
Meyer, F. O., 2000, Carbonate sheet slump from the
Jubaila Formation, Saudi Arabia: Slope implications:
GeoArabia, v. 5, no. 1, p. 144 – 145.
Meyer, F. O., and R. C. Price, 1993, A new Arab-D deposi-
tional model, Ghawar field, Saudi Arabia: 8th Middle
East Oil Show and Conference Proceedings, Bahrain,
p. 465 – 474.
Meyer, F. O., R. C. Price, I. A. Al-Ghamdi, I. M. Al-Goba,
S. M. Al-Raimi, and J. C. Cole, 1996, Sequential stra-
tigraphy of outcropping strata equivalent to Arab-D
reservoir, Wadi Nisah, Saudi Arabia: GeoArabia, v. 1,
p. 435 – 456.
Meyer, F. O., G. W. Hughes, and I. Al-Ghamdi, 2000, Ju-
baila Formation, Tuwaiq Mountain escarpment, Saudi
Arabia: Window to lower Arab-D reservoir faunal as-
semblages and bedding geometry (abs.): GeoArabia,
v. 5, no. 1, p. 143.
Milankovitch, M., 1930, Mathematische klimalehre and
astronomische theorie der klimaschwankungen, in
W. Koppen and R. Geiger, eds., Handbuch der Klima-
tologie, I (A): Borntraeger, Berlin, Gebruder, 176 p.
Milankovitch, M., 1941, Kanon der erdbestrahlung und
seine anwendung auf das eiszeitenproblem: Akad
Royale Serbe, v. 133, 633 p.
Mitchell, J. C., P. J. Lehmann, D. L. Cantrell, I. A. Al-Jallal,
and M. A. R. Al-Thagafy, 1988, Lithofacies, diagenesis
and depositional sequence; Arab-D Member, Ghawar
field, Saudi Arabia, in A. J. Lomando and P. M. Harris,
eds., Giant oil and gas fields— A core workshop: SEPM
Core Workshop, v. 12, p. 459 – 514.
Murris, R. J., 1980, Middle East stratigraphic evolution and
oil habitat: AAPG Bulletin, v. 64, p. 597 – 618.
Newberry, B. M., L. M. Grace, and D. D. Stief, 1996, Analy-
sis of carbonate dual porosity systems from borehole
electrical images: Society of Petroleum Engineers
Paper 35158, 7 p.
Nicholson, P. G., 2000, Compressional, fault-related folds
and Saudi Arabia’s major hydrocarbon fields (abs.):
GeoArabia, v. 5, p. 152 – 153.
Nicholson, P. G., 2002, A 700 million year tectonic frame-
work for hydrocarbon exploration and production in
Saudi Arabia (abs.): GeoArabia, v. 7, p. 284.
Pleydell, S. M., B. Jones, F. J. Longstaffe, and H. Baadsgaard,
1990, Dolomitization of the Oligocene – Miocene Bluff
Formation on Grand Cayman, British West Indies:
Canadian Journal Earth Sciences, v. 27, p. 1098 –
1110.
Pollastro, R. M., A. S. Karshbaum, and R. J. Vigor, 1999,
Maps showing geology, oil and gas fields and geologic
provinces of the Arabian Peninsula: U.S. Geological
Survey Open-file Report, 97-470B, version 2.0, one
CD-ROM.
Powers, R. W., 1962, Arabian Upper Jurassic carbonate
reservoir rocks, in W. E. Ham, ed., Classification of car-
bonate rocks: AAPG Memoir 1, p. 122– 192.
Powers, R. W., 1968, Saudi Arabia: Lexique Stratigraphique
International, 3: Paris, Centre National de la Recherche
Scientifique, 171 p.
Powers, R. W., L. R. Ramirez, C. D. Redmond, and E. L.
Elberg, 1966, Sedimentary geology of Saudi Arabia: Ge-
ology of the Arabian Peninsula: U.S. Geological Survey
Professional Paper 560-D, 150 p.
Radke, B. M., and R. L. Mathis, 1980, On the formation
and occurrence of saddle dolomite: Journal of Sedi-
mentary Petrology, v. 50, p. 1149 – 1168.
Read, J. F., 2004, Carbonate sequence stratigraphy, short
course notes: Saudi Aramco, Dhahran, Saudi Arabia,
unpaginated.
Russell, S. D., M. Akbar, B. Vissapragada, and G. Walkden,
2002, Rock types and permeability prediction from
dipmeter and image logs: AAPG Bulletin, v. 86, no. 10,
p. 1709 – 1732.
Russell, S. D., K. Sadler, W. H. Weihua, P. Richter, R. Y.
Eyvazzadeh, and E. A. Clerke, 2004, Applications of
image log analyses to reservoir characterization,
Ghawar and Shaybah fields, Saudi Arabia: GEO 2004,
Middle East Geoscience Conference and Exhibition,
abstract, p. 125.
Scotese, C. R., 1998, Quick time computer animations,
Paleomap Project: Department of Geology, University
of Texas at Arlington.
Sharland, P. R., R. Archer, D. M. Casey, R. B. Davies, S. H.
Hall, A. P. Heward, A. D. Horbury, and M. D. Simmons,
2001, Arabian plate sequence stratigraphy: GeoArabia
Special Publication 2, 371 p.
Shearman, D. J., 1978, Evaporites of coastal sabkhas, in
W. E. Dean and B. C. Schreiber, eds., Marine evapo-
rites: SEPM Short Course 4, p. 6 – 42.
Sibley, D. F., 1982, The origin of common dolomite fab-
rics: Clues from the Pliocene: Journal of Sedimentary
Petrology, v. 52, p. 1087 – 1100.
 Ghawar Arab-D Reservoir: Widespread Porosity in Shoaling-upward Carbonate Cycles / 137
Sibley, D. F., 1991, Secular changes in the amount and
texture of dolomite: Geology, v. 19, p. 151 – 154.
Sibley, D. F., and J. M. Gregg, 1987, Classification of
dolomite textures: Journal of Sedimentary Petrology,
v. 57, p. 967 – 975.
Stegner, W., 1971, Discovery: The search for Arabian oil:
Beirut, Middle East Export Press, 190 p.
Steineke, M., R. A. Bramkamp, and N. J. Sander, 1958, Strati-
graphic relations of Arabian Jurassic oil, in L. G. Weeks,
ed., Habitat of oil: AAPG Symposium, p. 1294–1329.
Swart, P. K., D. L. Cantrell, H. Westphal, C. R. Handford,
and C. G. St. C. Kendall, 2005, Origin of dolomite
from Ghawar field, Saudi Arabia: Evidence from petro-
graphic and geochemical constraints: Journal of Sedi-
mentary Petrology, v. 75, p. 476 – 491.
U.S. Geological Survey, 1963, Geologic map of the Arabian
Peninsula, scale 1:2,000,000, and geologic quadrangle
maps, scale 1:500,000: 1 map.
Wender, L. E., J. W. Bryant, M. F. Dickens, A. S. Neville,
and A. M. Al-Moqbel, 1998, Paleozoic (pre-Khuff ) hy-
drocarbon geology of the Ghawar area, eastern Saudi
Arabia: GeoArabia, v. 3, p. 273 – 302.
Xu, C., S. D. Russell, J. Gournay, and P. Richter, 2006,
Porosity partitioning and permeability quantifica-
tion in vuggy carbonates using wireline logs, Permian
Basin, West Texas: Petrophysics, v. 47, no. 1, p. 13 –
22.
Ziegler, M. A., 2001, Late Permian to Holocene paleofacies
evolution of the Arabian plate and its hydrocarbon
occurrences: GeoArabia, v. 6, p. 445 – 504.
FIGURE 8. Core description of the SDGM Arab-D core. Six cores were taken, with a total cored interval of 299 ft (91 m), through the complete Arab-D reservoir.
 Strohmenger, C. J., L. J. Weber, A. Ghani, K. Al-Mehsin, O. Al-Jeelani,

4
A. Al-Mansoori, T. Al-Dayyani, L. Vaughan, S. A. Khan, and J. C. Mitchell,
2006, High-resolution sequence stratigraphy and reservoir characterization
of Upper Thamama (Lower Cretaceous) Reservoirs of a giant Abu Dhabi
oil field, United Arab Emirates, in P. M. Harris and L. J. Weber, eds., Giant
hydrocarbon reservoirs of the world: From rocks to reservoir characterization
and modeling: AAPG Memoir 88/SEPM Special Publication, p. 139 – 171.

High-resolution Sequence
Stratigraphy and Reservoir
Characterization of Upper
Thamama (Lower Cretaceous)
Reservoirs of a Giant Abu Dhabi
Oil Field, United Arab Emirates
Christian J. Strohmenger, Lee Vaughan
Ahmed Ghani, Omar Al-Jeelani, ExxonMobil Exploration Company,
Abdulla Al-Mansoori, Houston, Texas, U.S.A.
and Taha Al-Dayyani,
Abu Dhabi Company for Onshore Oil Sameer A. Khan
Operations, Abu Dhabi, United Arab ExxonMobil Upstream Research Company,
Emirates Houston, Texas, U.S.A.

L. Jim Weber John C. Mitchell

ExxonMobil Exploration Company, Houston, ExxonMobil Exploration Company,
Texas, U.S.A. Houston, Texas, U.S.A.

Khalil Al-Mehsin
Abu Dhabi National Oil Company,
Abu Dhabi, United Arab Emirates

ABSTRACT

I
mportant hydrocarbon accumulations occur in platform carbonates of the
Lower Cretaceous Kharaib (Barremian and early Aptian) and Shuaiba (Aptian)
formations (upper Thamama Group) of Abu Dhabi. The Kharaib and Lower
Shuaiba formations contain three reservoir units separated by three low-porosity
and low-permeability dense zones. From base to top, the thickness of the res-
ervoir intervals range from approximately 80, 170, to 55 ft (24, 51, to 16 m),

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215876M883271

139
140 / Strohmenger et al.

respectively, for the Lower Kharaib, Upper Kharaib, and Lower Shuaiba Reservoir
Units. Core and well-log data of a giant oil field of Abu Dhabi, as well as outcrop
data from Wadi Rahabah in the Emirate of Ras Al-Khaimah were used to establish
a sequence-stratigraphic framework and a lithofacies scheme, applicable to all
three reservoir units and the three dense zones.
The Lower and Upper Kharaib Reservoir Units, as well as the lower, middle,
and upper dense zones are part of the late transgressive sequence set of a second-
order supersequence, made up of two third-order composite sequences. The over-
lying Lower Shuaiba Reservoir Unit belongs to the late transgressive sequence set
and the early highstand sequence set of this second-order supersequence and is
made up of one third-order composite sequence. The three third-order com-
posite sequences are composed of 19 fourth-order parasequence sets that show
predominantly aggradational and progradational stacking patterns, typical of
greenhouse cycles. Conventionally, composite sequence boundaries are placed
at or near the base of the three dense zones. As an alternative scenario, the pos-
sibility that the major composite sequence boundaries actually occur on top of
these dense zones is discussed.
On the basis of faunal content, texture, sedimentary structures, and litho-
logic composition, 13 reservoir lithofacies and 8 nonreservoir (dense) lithofacies
are identified from core. Similar lithofacies are identified in time-equivalent rock
exposures studied in Wadi Rahabah. Depositional environments of reservoir units
range from lower ramp to shoal crest to near-back shoal open-platform deposits.
Dense zones were deposited in an inner-ramp, restricted shallow-lagoonal setting.
Intensively bioturbated wackestone and packstone, and interbedded organic- and
siliciclastic-rich limestone, characterize the dense zones. Locally, mud cracks,
blackened grains, and rootlets are observed.
Outcrop analogs of subsurface reservoirs allow for a detailed investigation
of facies architecture and structure of carbonate bodies. Integration of subsur-
face and outcrop data (e.g., low-angle clinoforms that cannot be seen in core
data) leads to more insightful and realistic geological models of subsurface stra-
tigraphy. Geological model realizations based on core, outcrop, well-log, and
seismic data constrain fluid flow-simulation models. Results mimic known be-
havior in analogous producing fields, and the process of going from rock data to
simulation provides a useful training tool for reservoir characterization methods
and techniques.

INTRODUCTION the Lower Shuaiba Formation. The recoverable oil

Large hydrocarbon accumulations have been dis-
covered and produced from platform carbonates of
the Lower Cretaceous Kharaib and Shuaiba forma-
tions (Thamama Group) in Abu Dhabi (Alsharhan,
1989; Alsharhan and Nairn, 1993, 1997). The focus of
this study is on the sequence stratigraphy and sedi-
mentology of a giant Abu Dhabi oil field referred to
herein as Field B (Figure 1).
Field B was discovered in 1965 and went on pro-
duction in 1973. It is a faulted anticline (Figure 2)
with production of about 408 API oil from the Kha-
raib Formation and planned future production from
reserves are estimated to be in billions of barrels.
More than 350 production and water-injection wells
have been drilled to date.
The Kharaib Formation is of Barremian and early
Aptian age (Vahrenkamp, 1996; Granier, 2000; Pittet
et al., 2002; Granier et al., 2003; C. Liu, T.-C. Huang,
Y.-Y. Chen, 2005, personal communication) (Figure 3)
and contains two reservoir units (Lower Kharaib Res-
ervoir Unit and Upper Kharaib Reservoir Unit) sepa-
rated and encased by three zones of very low po-
rosity and permeability, subsequently referred to as
dense zones (lower, middle, and upper dense zones;
Figure 4). The thickness of the Lower Kharaib Reservoir
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 141

FIGURE 1. Location map showing the major oil fields of Abu Dhabi (green) and Field B (red) where core and well-log
data of the Kharaib and Lower Shuaiba formations were studied. Also shown is the location of the studied outcrops
at Wadi Rahabah, Ras Al-Khaimah.

Unit is about 80 ft (24 m), and the thickness of the
Upper Kharaib Reservoir Unit ranges between 150 ft
(45 m; downflank) and 190 ft (58 m; crestal area).
The thickness of the lower dense zone, underlying
the Lower Kharaib Reservoir Unit, is approximately
50 ft (15 m); the thickness of the middle dense zone,
separating the Upper and Lower Kharaib Reservoir
Units, is approximately 45 ft (14 m), and the thick-
ness of the upper dense zone, overlying the Up-
per Kharaib Reservoir Unit, is approximately 40 ft
(12 m).
The thickness of the Lower Shuaiba Reservoir Unit
(overlying the upper dense zone; Figures 3, 4) is ap-
proximately 55 ft (17 m).
Highly permeable beds (>1 d) that mainly consist
of coated-grain grainstone and rudist floatstone-
rudstone facies are recognized in the upper parts
of the Upper and Lower Kharaib Reservoir Units
(Figure 4). Generally, reservoir quality degrades from
the crest to the flank throughout the field. This is
because of pressure solution (compaction) and ce-
mentation caused by the interaction of the carbon-
ates with formation water and increasing burial
depth (Grötsch et al., 1998a). In contrast to the water
leg, the oil leg (mostly the crestal area) is not as af-
fected by pressure-dissolution processes (Burgess and
Peter, 1985). The presence of hydrocarbons seems to
inhibit the development of stylolites. The distribu-
tion of stylolites therefore shows an apparent dichot-
omy between crest (less stylolites) and flank (more
stylolites), interpreted not to be related to differential
burial effects between crest and flank, but instead, to
the hydrocarbon infilling of the structure. Crestal
wells (oil leg) are slightly thicker and show excellent
porosity and permeability compared to downflank
wells (water leg), which exhibit reduced thickness
and reduced porosity and permeability (Grötsch et al.,
1998a).
Traditionally, the subdivision of the Kharaib and
Lower Shuaiba formations into reservoir units and
142 / Strohmenger et al.

FIGURE 2. Seismic line (west – east) running through the center of Field B shown in Figure 1. The seismic cross section
shows the structure of Field B and the interpreted main stratigraphic horizons from Upper Jurassic to Upper Cretaceous.
LKRU = Lower Kharaib Reservoir Unit; UKRU = Upper Kharaib Reservoir Unit; LSRU = Lower Shuaiba Reservoir Unit.

subunits is based on lithostratigraphic correlation, focused on reservoir characterization (Grötsch et al.,

using the vertical distribution of stylolites. Investi-
gation of stylolite distribution in the Kharaib For-
mation shows that such an approach is not tenable
and, therefore, not recommended for building geo-
logical (static) and, ultimately, reservoir (dynamic)
models. The origin and the degree of stylolitization
depend on the heterogeneity of the sedimentary col-
umn (e.g., preexisting surfaces, lithological contrasts,
facies changes, and type of bedding; Park and Schot,
1968; Nelson, 1984). Moreover, some facies will be
less affected than others by stylolites. For example,
algal lithofacies will be more resistant to compaction
during burial because of calcite cement trapping in
the algal structure. In addition, early diagenetic rim
cement in interparticle pore space of grain-dominated
facies minimizes compaction.
Previous Abu Dhabi Company for Onshore Oil
Operations (ADCO) internal and published studies
conducted on the Kharaib and Lower Shuaiba Res-
ervoir Units of Field B and outcrop analogs of Oman
1998a; Melville et al., 2004), facies analyses, and se-
quence stratigraphy (Grötsch et al., 1998a; Borgomano
et al., 2002; Pittet et al., 2002; van Buchem et al.,
2002; Hillgärtner et al., 2003; Immenhauser et al.,
2004) and agree closely with the results presented in
this chapter. The results of published and unpub-
lished studies conducted at Field B were used to build
three-dimensional (3-D) geological and reservoir mod-
els that adequately describe the reservoir (Grötsch
et al., 1998a; Melville et al., 2004) in its midterm pro-
duction state, but are based on only limited core data.
The results presented in this chapter integrate core
and well-log data from Field B, as well as outcrop data
(Strohmenger et al., 2004a, b, c; Suwaina et al., 2004)
from Ras Al-Khaimah (United Arab Emirates; Figure 1).
For the first time, detailed, sequence stratigraphy-
keyed descriptions of approximately 15,000 ft (4580 m)
of core material from 50 wells of Field B and 4 vertical
outcrop sections of Wadi Rahabah (Ras Al-Khaimah)
have been conducted, considerably improving the
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 143
preexisting facies and sequence stratigraphy models.
The newly developed high-resolution sequence-
stratigraphic framework and facies scheme allow a
better prediction of the vertical and lateral distri-
bution of reservoir quality and reservoir continuity
throughout Field B and have become standard for all
other ADCO fields.
Comparing the location of stylolites or denser,
stylolite-bearing zones with the established high-
resolution sequence-stratigraphic framework of the
Kharaib Formation reveals that some stylolites are re-
lated to third-, fourth-, and fifth-order sequence bound-
FIGURE 3. Upper Thamama Group
(Lower Cretaceous) sequence-
stratigraphic framework.
aries, parasequence set boundaries,
and parasequence boundaries (Boi-
chard et al., 1994; Rebelle et al.,
2004; Strohmenger et al., 2004b).
This can be explained by early dia-
genetic cementation occurring at
the top of the shallowing-upward
cycles, creating a contrast in rock
properties with the overlying rock
type. During burial, these hetero-
geneous intervals undergo prefer-
ential stylolitization. Calcite that
is dissolved during pressure solu-
tion reprecipitates in pores adja-
cent to the stylolites (Park and Schot,
1968; Koepnick, 1984), forming a
nonporous zone adjacent to the
chronostratigraphic boundary. If
the boundary separates grain-
dominated lithofacies (below)
from mud-dominated lithofacies
(above), cementation will be locat-
ed dominantly below the bound-
ary. In the case of grain-dominated
overlying material, cementation
may occur above and below the
chronostratigraphic boundary (Re-
belle et al., 2004; Strohmenger
et al., 2004b). Early diagenetic pro-
cesses thus follow the sequence-
stratigraphic framework and, there-
fore, can be predicted away from
well control.
THAMAMA GROUP
SEQUENCE-STRATIGRAPHIC FRAMEWORK
The Thamama Group can be described by two
second-order supersequences. The older superse-
quence corresponds to the Habshan Formation (Ber-
riasian and early Valanginian). The younger super-
sequence encompasses the interval between the top
Habshan and top Shuaiba formations (Sharland et al.,
2001, 2004; Davies et al., 2002; Droste and van Steen-
winkel, 2004; Strohmenger et al., 2004a, b, c; Haq
and Al-Qahtani, 2005) (Figure 3). The transgressive
144 / Strohmenger et al.
sequence set of this supersequence includes, from
oldest to youngest, the Lekhwair (Valanginian to
Barremian), and the Kharaib (Barremian and early
Aptian) formations; including the Hawar (upper
dense zone). The Shuaiba Formation is of Aptian
age (Vahrenkamp, 1996; Grötsch et al., 1998b; Granier,
2000; Pittet et al., 2002; Granier et al., 2003; C. Liu,
T.-C. Huang, Y.-Y. Chen, 2005, personal communi-
cation) and is part of the late transgressive sequence
set (early Aptian) and the highstand sequence set
(Late Aptian) of this longer term depositional cycle.
FIGURE 4. Type well B-1,
showing the established high-
resolution sequence-stratigraphic
framework for the Kharaib and
Lower Shuaiba formations. ULRU 1 =
Upper Lekhwair Reservoir Unit 1.
See Figures 9 – 29 for color code
of lithofacies types.
Based on facies stacking pat-
terns (shallowing-upward trends),
sedimentary structures (e.g.,
Glossifungites burrows), and sur-
face morphologies (e.g., erosive
surfaces), 4 third-order compos-
ite sequence boundaries (SB),
3 composite maximum flooding
surfaces (MFS), and 13 marine
flooding surfaces (FS), which may
correspond to third- or higher
order flooding surfaces and se-
quence boundaries (flooding sur-
face/sequence boundary [FS/SB])
were identified from core mate-
rial of the Kharaib and Lower
Shuaiba formations and tied to
wire-line logs (Figure 4).
Where possible, third-order
SBs, MFSs, and parasequence
set boundaries (FS) were tied to
the stratigraphic framework
established by Sharland et al.
(2001, 2004) and Davies et al.
(2002) for the Arabian plate
(Figure 4). The nomenclature
proposed by Sharland et al.
(2001, 2004) is based on genetic
stratigraphic sequences (Gallo-
way, 1989), bounded by MFSs.
Each of their identified MFSs
is assigned a name, a number,
and an age, like K60 (123 Ma) and K70 (120 Ma). In
contrast, our sequence-stratigraphic subdivision fol-
lows the sequence-stratigraphic approach estab-
lished by ExxonMobil (Mitchum, 1977; Vail et al.,
1977, 1991; Vail, 1987; Van Wagoner et al., 1987, 1988;
Haq et al., 1988; Sarg, 1988; Sarg et al., 1999), where
depositional sequences are bounded by sequence
boundaries. To tie our identified chronostratigraph-
ic boundaries back to the already existing strati-
graphic framework of the Arabian plate, we had to
update the existing nomenclature proposed by
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 145
Sharland et al. (2001, 2004) and Davies et al. (2002)
in a way that our boundaries are assigned to the
already existing names (e.g., K60 and K70) but also
append information whether these boundaries cor-
respond to sequence boundaries (e.g., K60_SB), maxi-
mum flooding surfaces (e.g., K60_MFS: correspond-
ing to the K60 MFS of Sharland et al., 2001, 2004) or
flooding surfaces (e.g., K60_FS100).
Kharaib and Lower Shuaiba
Sequence-stratigraphic Framework
The Kharaib Formation, including the Hawar
(upper dense zone), corresponds to the late trans-
gressive sequence set (TSS) of a second-order super-
sequence. It is described by a second-order sequence
set (base lower dense zone to base upper dense zone
or Hawar), made up of two third-order composite
sequences (Figure 3). The lower third-order compos-
ite sequence starts at the base of the lower dense zone
and is capped by a pronounced sequence boundary
(exposure surface) about 20 ft (6 m) below the middle
dense zone (Figure 4). It is overlain by a third-order
composite sequence that is bounded on top by a re-
gionally correlative sequence boundary below the
upper dense zone (Figure 4). Fourteen fourth-order
parasequence sets and several fifth-order parase-
quences that make up the two third-order composite
sequences show predominantly aggradational and
progradational stacking patterns, typical for green-
house cycles (Sarg et al., 1999) (Figure 4).
The Lower Shuaiba Formation corresponds to the
late transgressive (TSS) and early highstand sequence
set (HSS) of the same second-order supersequence
and is made up of a third-order composite sequence
(Yose et al., 2004a, b, 2006) (Figure 3), subdivided by
five fourth-order parasequence sets (Figure 4).
Lower Kharaib Third-order Composite Sequence
Intensively burrowed and bioturbated wackestone
and packstone and interbedded organic- and silici-
clastic-rich limestone characterize the lower part
of the Lower Kharaib sequence. Thalassinoides bur-
rows are abundant, and burrow fills are commonly
dolomitized. Well-developed firmground surfaces
(Glossifungites burrows) cap several parasequences
(Sattler et al., 2005). Locally, mud cracks, blackened
grains, and paleosols have been observed. Diversity
of fauna and flora is low. Other characteristics of this
interval are centimeter- and decimeter-scale discon-
tinuous beds. However, the overall interval, referred
to as the lower dense zone (Figure 4), can be
correlated for tens to hundreds of kilometers across
the platform. This interval is interpreted as the early
transgressive systems tract. Low- to moderate-energy,
skeletal wackestone and packstone that overlie the
lower dense zone are interpreted to correspond to the
late transgressive systems tract (Figure 4).
Overlying the transgressive systems tract (TST),
the highstand systems tract (HST) exhibits a normal
and diverse marine fauna and flora characterized by
high-energy coated-grain, algal, skeletal grainstone
and rudstone with sparse rudist fragments (Figure 4).
The depositional environment is interpreted as a
tidal-influenced, high-energy bioclastic shoal envi-
ronment. Water depth probably did not exceed 10 m
(33 ft) even during the highest rates of relative sea
level rise. Normal-marine conditions with good wa-
ter circulation prevailed, as indicated by the generally
diverse faunal content. Nutrients were sufficiently
abundant to sustain the development of the algal
buildups and rudists. Water turbidity was low as in-
dicated by the abundance of Lithocodium/Bacinella
(Dupraz and Strasser, 1999; Immenhauser et al., 2005).
Four fourth-order parasequence sets are capped
by two marine flooding surfaces (K50_FS100 and
K50_FS600), a third-order maximum flooding surface
(K50_MFS; Figure 5A), and a third-order sequence
boundary (K60_SB; Figure 5B). Several small-scale
parasequences are identified in the Lower Kharaib
sequence (K50_SB to K60_SB; Figure 4). These para-
sequences are superimposed on a larger third-order
trend. Successive parasequences show slightly differ-
ent characteristics. Oldest parasequences are mud-
dominated lagoonal sediments and show intense bio-
turbation and firmgrounds at bounding discontinuity
surfaces (Sattler et al., 2005). Firmgrounds formed
during periods of subaerial exposure and subsequent
rapid increase in accommodation and reduced sedi-
mentation rates. Subsequent parasequences are thicker
and are made up of lagoonal wackestone to packstone
and algal, skeletal, peloid floatstone-rudstone. Sedi-
mentation occurred in an open lagoonal environ-
ment with increased water circulation during loss of
accommodation. Upper parasequences are grain domi-
nated (algal, skeletal, peloid floatstone-rudstone, coated-
grain grainstone, and coated-grain, algal, skeletal rud-
stone) and thinner, indicating an overall increase in
current energy and decrease in accommodation. The
uppermost parasequences are thin and dominated by
moderate- to high-energy miliolid shoals (peloid,
skeletal packstone and peloid, skeletal grainstone).
At this time, accommodation was nearly filled.
146 / Strohmenger et al.

FIGURE 5. Core photographs of chronostratigraphic boundaries. (A) Maximum flooding surface K50_MFS showing
burrowed and bored hardground (arrows), as well as patchy dolomitization (Do). (B) Sequence boundary K60_SB showing
stylolitically overprinted erosive surface, iron mineralization, and cementation below sequence boundary. (C) Maximum
flooding surface K60_MFS showing erosive contact between skeletal wackestone (SW, below) and skeletal, peloid packstone
(SPP, above), as well as patchy dolomitization (Do). (D) Sequence boundary K70_SB showing erosive and burrowed
surface (Glossifungites burrows, Gl), as well as cementation below sequence boundary. (E) Sequence boundary K80_SB
showing stylolitically overprinted erosive surface and cementation below sequence boundary. (F) Flooding surface
K70_FS100 on top of the upper dense zone (Hawar), which might also correspond to third-order sequence boundary
Ap3sb (see Figure 4). The pronounced pedogenic overprint of the carbonates below the erosive surface favors the
interpretation of a major subaerial exposure surface.

Upper Kharaib Third-Order Composite Sequence
The early transgressive systems tract of the Upper
Kharaib sequence is composed of Orbitolina-rich, bio-
turbated and burrowed skeletal, peloid wackestone
and packstone of the uppermost Lower Kharaib Res-
ervoir Unit (Figure 4) and is overlain by intensively
burrowed and bioturbated skeletal, peloid wacke-
stone and sparse packstone. This interval is laterally
continuous over considerable distance (tens to hun-
dreds of kilometers) and is referred to as the middle
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 147
dense zone (Figure 4). Cycle tops are capped by firm-
grounds (Glossifungites burrows), with sediment- and
calcite spar-filled burrows (Sattler et al., 2005). The
middle dense zone consists of abundant peloids, small
echinoderm fragments, and some dasycladacean and
red algae. In addition, other green algae, small mol-
lusks, and sponge spicules are common. Locally, mud
cracks, blackened grains, and paleosols have been ob-
served. In outcrops, the organic- and siliciclastic-rich
nature of the Kharaib middle dense zone results in a
friable, nodular weathered outcrop.
During the late transgressive systems tract, an open
lagoonal environment persisted on the platform. Dolo-
mitized Thalassinoides firmgrounds indicate temporary
cessation in sedimentation. Burrowed horizons are con-
tinuous over many kilometers (Sattler et al., 2005). The
fauna is more diverse, with notable orbitolinids, echi-
noderm remains, and bivalves. Carbonate rock textures
are dominated by skeletal wackestone. At the ocean
margin, rudist shoals, microbial mounds, and beaches
developed at this time (van Buchem et al., 2002).
Skeletal, peloid wackestone-packstone, coated-grain
grainstone, coated-grain, algal, skeletal grainstone
and rudstone, rudist and chondrodont floatstone and
rudstone, and peloid, skeletal packstone and grain-
stone, rich in miliolids, gradually covered the entire
platform during the highstand systems tract. The ru-
dists include caprotinids (Glossomyophorus) and mono-
pleurids (Agriopleura) but no caprinids. Rudists, as well
as chondrodonts, reflect deposition in shallow water.
These grain-dominated lithofacies were deposited un-
der moderate- to high-energy, normal-marine condi-
tions and exhibit considerable lateral variability. In
the uppermost part of the sequence, miliolid-rich,
skeletal, peloid packstone and grainstone with planar
bedding and low-angle cross-bedding imply moderate-
to high-energy conditions.
In the Upper Kharaib sequence, shoaling did not
occur until the uppermost part of the highstand sys-
tems tract. An increase in overall accommodation
from the Lower Kharaib to the Upper Kharaib sequence
may reflect a long-term second-order transgressive
trend in relative sea level (Figure 4).
Ten fourth-order parasequence sets bounded on
top by eight marine flooding surfaces (K60_FS100,
K60_FS200, K60_FS300, K60_FS600, K60_FS700,
K60_FS800, K60_FS900, and K60_FS1000), a third-order
maximum flooding surface (K60_MFS; Figure 5C), and
a third-order sequence boundary (K70_SB; Figure 5D),
as well as several small-scale parasequences, are
identified in the Upper Kharaib sequence (K60_SB to
K70_SB; Figure 4).
Lower Shuaiba Third-order Composite Sequence
The early transgressive systems tract of the third-
order Lower Shuaiba sequence corresponds to the up-
per dense zone (Hawar), overlying the Upper Kharaib
sequence (Figure 4). This interval is rich in discoidal
orbitolinids. Peloids, small echinoderm fragments, and
dasycladacean algae also occur. The upper dense zone is
organic and siliciclastic rich, and can be correlated for
tens to hundreds of kilometers across the platform
with varying thicknesses. Decimeter-scale beds are
marked by highly burrowed firmgrounds (Glossifungites
burrows). Mud cracks, blackened grains, and paleosols
are quite common, indicating frequent exposure.
The Lower Shuaiba Reservoir Unit, overlying the
upper dense zone, is a shallow-to deeper water limestone
consisting of algal (Lithocodium/Bacinella) boundstone
(Immenhauser et al., 2005) and skeletal, peloid wacke-
stone to packstone and foraminifera, skeletal wacke-
stone (Figure 4). This succession shows a deepening-
upward trend interpreted to correspond to the late
transgressive systems tract. The overlying highstand
systems tract is dominated by deeper marine forami-
nifera, skeletal mudstone to wackestone that show a
subtle shallowing-upward trend as well as an upward
increase in grain richness (graining upward) to skel-
etal (shell fragments), peloid wackestone to packstone.
The Lower Shuaiba composite sequence is bounded
on top by sequence boundary K80_SB (Figure 5E).
Five fourth-order parasequence sets are capped by
three marine flooding surfaces (K70_FS100, K70_FS200,
and K70_FS300), a third-order maximum flooding
surface (K70_MFS), and a third-order sequence bound-
ary (K80_SB; Figure 5E). Several small-scale parase-
quences are identified in the lower part of the Lower
Shuaiba sequence (K70_SB to K80_SB; Figure 4).
The overlying Upper Shuaiba Reservoir Unit is dom-
inated by deeper marine, planktonic foraminifera
wackestone-mudstone and intercalated low-porosity
and low-permeability, burrowed, and bioturbated
wackestone to mudstone (dense intervals).
Alternative Sequence-stratigraphic Interpretation
of the Kharaib Formation
The identified composite sequence boundaries
K60_SB and K70_SB of the Kharaib Formation clearly
show indications of significant emersion of the Kha-
raib platform carbonates (Figure 5B, D) and can re-
gionally be correlated throughout the entire Arabian
platform. However, the overall facies successions
from shallow-water upper ramp carbonates (reservoir
lithofacies types) below the sequence boundaries to
inner ramp (dense zones lithofacies types) above the
148 / Strohmenger et al.

FIGURE 6. Schematic models showing two possible sequence-stratigraphic interpretations of the dense zones, using
the upper dense zone as an example. (A) Conventional model: upper dense zone deposits correspond to the early
transgressive systems tract of the Lower Shuaiba sequence, overlying major third-order composite sequence boundary
(K70_SB) on top of the Upper Kharaib Reservoir. Minor higher order flooding surface (K70_FS100) occurs on top of the
upper dense zone. (B) Alternative model: upper dense zone deposits correspond to the late highstand systems tract of
the Upper Kharaib sequence, displaying major third-order composite sequence boundary (Ap3sb) at the top of the upper
dense zone. Minor third- or higher order sequence boundaries occur at the base (Ap2sb) and within the upper dense zone.

sequence boundaries follow a normal shallowing-
upward and progradational trend. In addition, the so-
called dense zones, especially the upper dense zone
(Hawar), show frequent omission surfaces (firm-
grounds), most likely corresponding to subaerial expo-
sure, and significant thickness variation throughout
the Arabian platform. For example, the upper dense
zone decreases in thickness from approximately 40 ft
(12 m) in onshore Abu Dhabi to approximately 20 ft
(6 m) in offshore Abu Dhabi and is not present at the
platform margin in Oman (van Buchem et al., 2002).
Therefore, as an alternative sequence-stratigraphic in-
terpretation, composite sequence boundaries could
be placed at the top of the lower (Barr4sb), middle
(Barr6sb), and upper (Ap3sb; Figure 5F) dense zones
(Figure 4). The dense zones would thus correspond
to late highstand systems tracts, representing inner-
ramp, restricted platform deposits, time-equivalent to
outer ramp, prograding carbonates (Figure 6). The omis-
sion surfaces in the dense zones are interpreted to rep-
resent higher order sequence boundaries, indicating the
frequent subaerial exposure of the inner ramp during
the late highstand systems tracts (reduced accommo-
dation). Most of the time represented by the dense
zones would actually correspond to nondeposition.
KHARAIB AND LOWER SHUAIBA FACIES
AND DEPOSITIONAL ENVIRONMENT
On the basis of texture, grain types, sedimentary
structures, faunal content, and lithologic compo-
sition, 21 lithofacies types were defined for the
subsurface Kharaib and Lower Shuaiba formations
(Strohmenger et al., 2004a, b, c). Thirteen lithofacies
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 149

FIGURE 7. Paleobathymetrical profile showing the interpreted depositional environment of each of the thirteen
lithofacies types identified within the Lower and Upper Kharaib Reservoir Units and the Lower Shuaiba Reservoir Unit.

types (LF1–LF13) correspond to the three reservoir (Figure 7), and eight lithofacies types (LF20–LF27)
units: Lower Kharaib Reservoir Unit, Upper Kharaib occur in three dense zones: lower dense zone, middle
Reservoir Unit, and Lower Shuaiba Reservoir Unit dense zone, and upper dense zone (Figure 8). The

FIGURE 8. Paleobathymetrical profile showing the interpreted depositional environment of each of the eight lithofacies
types identified within the lower, middle, and upper dense zone.
150 / Strohmenger et al.
FIGURE 9. Reservoir lithofacies type 1: Rudist, peloid rudstone (RPR). Shown here are summary table, core photograph,
and thin-section photomicrograph (plane polarized light, porosity in blue). Thin-section photomicrograph shows large
chondrodont shells.
lithofacies types and their interpreted depositional
environments are listed below. Interpretations de-
scribed here agree closely with the work of Pittet et al.
(2002) and van Buchem et al. (2002).
Lithofacies Types of the Lower Kharaib,
Upper Kharaib, and Lower Shuaiba Reservoir Units
The analyzed 13 lithofacies types range from open
platform, deeper subtidal, lower ramp, to shallow sub-
tidal to intertidal, upper-ramp environments (Figure 7).
A detailed description of the individual lithofacies
types is given below. For a summary, see Figures 9–21.
Lithofacies LF1: Rudist, Peloid Rudstone (RPR)
 rudists, chondrodonts, and other mollusks, fora-
minifera, and echinoderms
 peloids and coated grains (ooids, superficial ooids)
 rudstone texture
 matrix: grainstone and packstone texture
 moderate to well sorted
 bimodal grain-size distribution
 blocky calcite cement as common pore-filling ce-
ment associated with subaerial exposure surfaces
 shallow subtidal, high-energy open platform
above fair-weather-wave base
 shoal to upper-ramp rudist buildups and re-
worked buildups
 high porosity and moderate to very high (>1 d)
matrix permeability
 moldic, vuggy, and interparticle porosity
 molds and vugs that dominate under exposure
surfaces
 Lower (uncommon or not present at Field B) and
Upper Kharaib Reservoir Units (see Figure 9)
Lithofacies LF2: Rudist, Peloid Floatstone (RPF)
 rudists, chondrodonts and other mollusks, fora-
minifera, echinoderms, and sponge spicules
 peloids
 floatstone texture
 matrix: packstone and wackestone texture
 poor to moderate sorted
 bimodal grain-size distribution
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 151

FIGURE 10. Reservoir lithofacies type 2: Rudist, peloid floatstone (RPF). Shown here are summary table, core photograph,
and thin section photomicrograph (plane polarized light, porosity in blue).

FIGURE 11. Reservoir lithofacies type 3: Skeletal, peloid packstone (SPP) overlain by algal, skeletal, peloid floatstone-
rudstone (ASPF). Shown here are summary table, core photographs, and thin-section photomicrograph (plane polarized
light, stained with alizarin red-S, porosity in blue). Skeletal, peloid packstone may show fenestral structures (arrows;
left core photograph) and commonly occurs on top of shallowing-upward parasequence sets and parasequences
(Glossifungites burrows, Gl; right core photograph).
152 / Strohmenger et al.

FIGURE 12. Reservoir lithofacies type 4: Skeletal, peloid grainstone (SPG). Shown here are summary table, core
photographs, and thin-section photomicrograph (plane polarized light, stained with alizarin red-S, porosity in blue).
Core photograph (left) shows low-angle swash cross-bedding. Thin-section photomicrograph shows abundant miliolids
and small echinoderm plates.

FIGURE 13. Reservoir lithofacies type 5: Coated-grain, skeletal grainstone (CgSG). Shown here are summary table, core
photograph, and thin-section photomicrograph (plane polarized light, porosity in blue).
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 153

FIGURE 14. Reservoir lithofacies type 6: Coated-grain, algal, skeletal rudstone – floatstone (CgASR). Shown here are
summary table, core photograph, and thin-section photomicrograph (plane polarized light, porosity in blue).

FIGURE 15. Reservoir lithofacies type 7: Algal, skeletal, peloid floatstone – rudstone. Shown here are summary table, core
photographs, and thin section photomicrograph (plane polarized light, porosity in blue). Thin-section photomicrograph
shows preserved intraparticle porosity in Lithocodium/Bacinella grain and pore-filling cement (white areas).
154 / Strohmenger et al.

FIGURE 16. Reservoir lithofacies type 8: Algal, skeletal floatstone – boundstone (ASFB). Shown here are summary table,
core photograph, and thin-section photomicrograph (plane polarized light, stained with alizarin red-S, porosity in blue).
Thin-section photomicrograph shows preserved primary intraparticle (framework) porosity in Lithocodium/Bacinella
boundstone.

FIGURE 17. Reservoir lithofacies type 9: Orbitolinid, skeletal packstone (OSP). Shown here are summary table, core
photograph, and thin-section photomicrograph (plane polarized light, stained with alizarin red-S, porosity in blue).
Thin-section photomicrograph displays echinoderm plates and densely packed discoidal orbitolinids with preserved
primary intraparticle porosity.
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 155

FIGURE 18. Reservoir lithofacies type 10: Skeletal, peloid wackestone – packstone (SPWP). Shown here are summary
table, core photograph, and thin-section photomicrograph (stained with alizarin red-S, porosity in blue). Thin-section
photomicrograph shows benthonic foraminifera and small echinoderm plates.

FIGURE 19. Reservoir lithofacies type 11: Orbitolinid, skeletal wackestone (OSW). Shown here are summary table, core
photograph, and thin-section photomicrograph (plane polarized light, porosity in blue). Thin-section photomicrograph
displays discoidal orbitolinids.
156 / Strohmenger et al.

FIGURE 20. Reservoir lithofacies type 12: Skeletal wackestone (SW). Shown here are summary table, core photograph,
and thin-section photomicrograph (plane polarized light, porosity in blue).

FIGURE 21. Reservoir lithofacies type 13: Foraminifera, skeletal wackestone (FSW). Shown here are summary table, core
photograph, and thin-section photomicrograph (plane polarized light, stained with alizarin red-S, porosity in blue). Core
photograph shows bioturbation. Thin-section photomicrograph displays planktonic foraminifera and echinoderm plates.
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 157
 blocky calcite as common pore-filling cement
associated with subaerial exposure surfaces
 shallow subtidal, moderate-energy open platform
above fair-weather-wave base
 upper ramp reworked rudist buildups
 low to high porosity and highly variable matrix
permeability
 moldic, vuggy, intraparticle, and microporosity
 molds and vugs that dominate under exposure
surfaces
 Lower (not present at Field B) and Upper Kharaib
Reservoir Units (see Figure 10)
Lithofacies LF3: Skeletal, Peloid Packstone (SPP)
 miliolids and other foraminifera, echinoderms,
algae, and mollusks
 miliolids that can be very abundant in the upper
part of the Upper and Lower Kharaib sequences
 peloids, aggregate grains, oncoids, and ooids
 packstone texture with minor grainstone interlayer
 moderate to well sorted
 unimodal grain-size distribution
 uncommonly planar laminated and low-angle
cross-bedded
 gradational basal contact and erosive and bur-
rowed upper contact (Glossifungites burrows)
 caps fining- and shallowing-upward cycles
 desiccation cracks and circumgranular cracks
around grains, indicating subaerial exposure
 shallow subtidal to intertidal, moderate-energy
restricted and open platform above fair-weather-
wave base
 inner shoal and upper ramp deposits
 low to high porosity and low to moderate matrix
permeability
 intraparticle, micromoldic, and microporosity
 Lower and Upper Kharaib Reservoir Units (see
Figure 11)
Lithofacies LF4: Skeletal, Peloid Grainstone (SPG)
 miliolids and other foraminifera and echinoderms
 miliolids that can be very abundant in the upper
part of the Upper and Lower Kharaib sequences
 peloids, coated grains (ooids, superficial ooids
and composite grains)
 grainstone texture with minor packstone interlayer
 well sorted
 unimodal grain-size distribution
 planar laminated and low-angle cross-bedded indi-
cating tidal influence
 gradational basal contact and sharp and erosive
(dominantly) or gradational upper contact
 shallow subtidal to intertidal, high-energy open
platform above fair-weather-wave base
 upper ramp beach, near shoal crest, and near
inner shoal deposits
 low to high porosity and low to high matrix
permeability
 interparticle, intraparticle, and micromoldic
porosity
 Lower and Upper Kharaib Reservoir Units (see
Figure 12)
Lithofacies LF5: Coated Grain,
Skeletal Grainstone (CgSG)
 miliolids and other foraminifera, echinoderms, green
algae, Lithocodium/Bacinella fragments, and mollusks
 coated grains (ooids, superficial ooids, compos-
ite grains, and oncoids), aggregate grains, intra-
clasts, and peloids
 grainstone texture with minor packstone interlayer
 moderate to well sorted
 unimodal to bimodal grain-size distribution
 uncommonly cross-bedded
 sharp basal contact and gradational (dominant)
or sharp upper contact
 shallow subtidal, high-energy open platform
above fair-weather-wave base
 upper-ramp, near-shoal crest deposits
 moderate to high porosity and low to very high
matrix permeability (>1 d)
 interparticle, intraparticle, moldic, and vuggy
porosity
 molds and vugs that dominate under exposure
surfaces
 Lower and Upper Kharaib Reservoir Units (see
Figure 13)
Lithofacies LF6: Coated-grain, Algal,
Skeletal Rudstone-Floatstone (CgASR)
 Lithocodium/Bacinella, miliolids and other fora-
minifera, echinoderms, green algae, rudists, and
other mollusks
 coated grains (composite grains, oncoids, super-
ficial ooids, and uncommon ooids), aggregate
grains, intraclasts, and peloids
 rudstone (predominant) to floatstone texture
 matrix: grainstone and mud-lean packstone texture
 poor to moderate sorted
 bimodal grain-size distribution
158 / Strohmenger et al.
 sharp basal contact and gradational upper contact
 shallow subtidal, high-energy open platform above
fair-weather-wave base
 upper-ramp, near-shoal crest deposits
 moderate to high porosity and low to very high
(> 1 d) matrix permeability
 interparticle, intraparticle, moldic, and vuggy
porosity
 molds and vugs that dominate under exposure
surfaces
 Lower and Upper Kharaib Reservoir Units (see
Figure 14)
Lithofacies LF7: Algal, Skeletal,
Peloid Floatstone-Rudstone (ASPF)
 Lithocodium/Bacinella, foraminifera, echinoderms,
rudists and other mollusks, and sponge spicules
 peloids, oncoids, aggregate grains, and uncom-
mon coated grains (superficial ooids)
 floatstone (predominantly) to rudstone texture
 matrix: packstone to mud-lean packstone texture
 poor to moderate sorted
 bimodal grain-size distribution
 sharp basal contact and gradational fining-
upward upper contact
 shallow subtidal, moderate- to high-energy open
platform above fair-weather-wave base
 upper-ramp algal buildups and reworked algal
buildups
 moderate to high porosity and low to high
matrix permeability
 interparticle, intraparticle, and microporosity
 Lower and Upper Kharaib Reservoir Units (see
Figure 15)
Lithofacies LF8: Algal,
Skeletal Floatstone-Boundstone (ASFB)
 Lithocodium/Bacinella, orbitolinids and other
foraminifera, mollusks, and echinoderms
 algal-binding activity forms low-relief digitate
growth framework and encrusting masses
 peloids
 floatstone to boundstone (bindstone) texture
 matrix: packstone and wackestone texture
 poor to moderate sorted
 bimodal grain-size distribution
 subtidal, low- to moderate energy open platform
near fair-weather-wave base
 lower, upper-ramp algal buildups and reworked
algal buildups
 moderate to high porosity and low to high ma-
trix permeability
 intraparticle (framework), shelter, and microporosity
 Lower Shuaiba Reservoir Unit (see Figure 16)
Lithofacies LF9: Orbitolinid,
Skeletal Packstone (OSP)
 orbitolinids and other foraminifera, mollusks,
echinoderms, and sponge spicules
 peloids
 packstone texture with minor wackestone inter-
layer
 moderate to well sorted
 bimodal grain-size distribution
 bioturbation
 shallow subtidal, low- to moderate-energy re-
stricted platform, and subtidal, open platform
near fair-weather-wave base deposits
 inner ramp, restricted lagoonal deposits, and
upper- to middle ramp deposits
 low to high porosity and low matrix permeability
 intraparticle, micromoldic, and microporosity
 Lower and Upper Kharaib Reservoir Units (see
Figure 17)
Lithofacies LF10: Skeletal,
Peloid Wackestone-Packstone (SPWP)
 foraminifera, sponge spicules, algae, and echinoderms
 peloids and uncommon coated grains
 wackestone to packstone texture
 moderately sorted
 unimodal grain-size distribution
 bioturbation
 thin dolomitized intervals common
 subtidal, low-energy open platform near fair-
weather-wave base
 upper to middle ramp deposits
 low to high porosity and low to moderate matrix
permeability
 intraparticle, micromoldic, and microporosity
 Lower and Upper Kharaib Reservoir Units, and
Lower Shuaiba Reservoir Unit (see Figure 18)
Lithofacies LF11: Orbitolinid,
Skeletal Wackestone (OSW)
 orbitolinids and other foraminifera, mollusks,
echinoderms, and sponge spicules
 peloids
 wackestone texture with minor packstone interlayer
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 159

 moderate sorted  fine-grained, microporous limestone with low

 bimodal grain-size distribution matrix permeability
 bioturbation  microporosity, intraparticle, and micromoldic
 thin dolomitized layers porosity
 shallow subtidal, low-energy restricted platform,  Lower Shuaiba Reservoir Unit (see Figure 21)
and deeper subtidal, open platform below fair-
weather-wave base
 inner-ramp, restricted lagoonal deposits, and Lithofacies Types of the Lower, Middle,
middle-ramp deposits and Upper Dense Zones
 low to high porosity and low matrix permeability The analyzed eight lithofacies types represent
 microporosity, intraparticle, and micromoldic organic- and siliciclastic-rich, inner-platform, restricted
porosity lagoonal deposits (Figure 8). The frequent occurrence
 Lower and Upper Kharaib Reservoir Units (see of blackened pebbles and blackened grains suggests
Figure 19) a coastal, peritidal, oxygen-depleted environment
(Strasser and Davaud, 1983; Strasser, 1984). This is sup-
Lithofacies LF12: Skeletal ported by the elevated organic content of the carbon-
Wackestone (SW) ates, interpreted from high uranium readings of the
spectral gamma-ray log. Sedimentary structures like
 foraminifera, sponge spicules, echinoderms, and Glossifungites burrows, desiccation cracks, root marks,
algae and erosive surfaces, as well as paleosols indicate re-
 peloids peated subaerial exposures. A detailed description of
 wackestone texture the individual lithofacies types is given below. For a
 moderately sorted summary see Figures 22–29.
 well sorted
 unimodal grain-size distribution Lithofacies LF20: Wispy-laminated, Burrowed,
 bioturbation Skeletal Packstone (WBSP)
 thin dolomitized layers
 subtidal, low-energy open platform below fair-  low diversity, abundant, and small-size fauna
weather-wave base  foraminifera, mollusks, echinoderms, sponge
 middle-ramp deposits spicules, and ostracods
 fine-grained, microporous limestone with low to  peloids
moderate matrix permeability  blackened pebbles and blackened grains
 microporosity, intraparticle, and micromoldic  pyrite
porosity  siliciclastics present
 Upper Kharaib Reservoir Unit (see Figure 20)  packstone texture
 wispy- and nodular-bedded
Lithofacies LF13: Foraminifera,  burrows and bioturbation
Skeletal Wackestone (FSW)  shallow subtidal, low- to moderate-energy re-
stricted platform
 planktonic and benthonic foraminifera, sponge  inner shoal to inner ramp, restricted lagoonal
spicules, and echinoderms deposits
 peloids  low porosity and low matrix permeability
 pyrite  lower and middle (uncommon) dense zones (see
 wackestone texture with minor mudstone inter- Figure 22)
layer
 moderately sorted Lithofacies LF21: Wispy-laminated, Burrowed,
 well sorted Orbitolinid, Skeletal Packstone (WBOSP)
 unimodal grain-size distribution
 bioturbation  low diversity and high abundance of fauna
 subtidal, low-energy open platform below fair-  discoidal orbitolinids and other foraminifera,
weather-wave base green algae (dasycladacean), mollusks, echino-
 middle to lower ramp deposits derms, sponge spicules, and ostracods
160 / Strohmenger et al.

FIGURE 22. Nonreservoir lithofacies type 20: Wispy-laminated, burrowed, skeletal packstone (WBSP). Shown here
are summary table, core photograph, and thin-section photomicrograph (plane polarized light). Thin-section
photomicrograph displays abundant skeletal debris.

FIGURE 23. Nonreservoir lithofacies type 21: Wispy-laminated, burrowed, orbitolinid, skeletal packstone (WBOSP).
Shown here are summary table, core photograph, and thin-section photomicrographs (plane polarized light). Thin-
section photomicrographs show densely packed discoidal orbitolinids and wispy stylolites.
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 161

FIGURE 24. Nonreservoir lithofacies type 22: Wispy-laminated, burrowed, orbitolinid, skeletal wackestone (WBOSW).
Shown here are summary table, core photograph, and thin-section photomicrograph (plane polarized light).

FIGURE 25. Nonreservoir lithofacies type 23: Wispy-laminated, burrowed, skeletal wackestone – mudstone (WBSW).
Shown here are summary table, core photograph, and thin-section photomicrograph (plane polarized light). Thin-
section photomicrograph shows dolomite rhombs and pyrite.
162 / Strohmenger et al.

FIGURE 26. Nonreservoir lithofacies type 24: Burrowed, bioturbated, skeletal packstone (BBSP). Shown here are
summary table, core photograph, and thin-section photomicrograph (plane polarized light). Core photograph shows
intensive burrowing and bioturbation.

FIGURE 27. Nonreservoir lithofacies type 25: Burrowed, bioturbated, orbitolinid, skeletal packstone (BBOSP). Shown here
are summary table, core photograph, and thin-section photomicrographs (plane polarized light, stained with alizarin
red-S). Thin-section photomicrograph shows densely packed discoidal orbitolinids and glauconite (green grains).
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 163

FIGURE 28. Nonreservoir lithofacies type 26: Burrowed, bioturbated, orbitolinid, skeletal wackestone (BBOSW). Shown
here are summary table, core photographs, and thin-section photomicrograph (plane polarized light). Thin-section
photomicrographs show glauconite (green grains, left thin section photomicrograph), discoidal orbitolinids, and green
algae (right thin-section photomicrograph).

FIGURE 29. Nonreservoir lithofacies type 27: Burrowed, bioturbated, skeletal wackestone (BBSW). Shown here are
summary table, core photograph, and thin-section photomicrograph (plane polarized light, stained with alizarin red-S).
164 / Strohmenger et al.
 peloids
 blackened pebbles and blackened grains
 dolomite rich
 pyrite rich
 siliciclastics present
 packstone texture
 wispy and nodular bedded
 burrows and bioturbation
 maroon lithoclasts, red paleosols, and caliche
crusts
 rhizoliths (root tubes) and desiccation cracks
 frequent subaerial exposure
 shallow subtidal, low- to moderate-energy re-
stricted platform
 inner-ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 upper dense zone (Hawar) (see Figure 23)
Lithofacies LF22: Wispy-Laminated, Burrowed,
Orbitolinid, Skeletal Wackestone (WBOSW)
 low diversity and high abundance of fauna
 discoidal orbitolinids and other foraminifera,
green algae (dasycladacean), mollusks, echino-
derms, sponge spicules, and ostracods
 peloids
 blackened pebbles and blackened grains
 dolomite rich
 pyrite rich
 siliciclastics present
 wackestone texture with minor packstone inter-
layer
 wispy and nodular bedded
 burrows and bioturbation
 maroon lithoclasts, red paleosols, and caliche crusts
 rhizoliths (root tubes), and desiccation cracks
 frequent subaerial exposure
 shallow subtidal, low-energy restricted platform
 inner-ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 upper dense zone (Hawar) (see Figure 24)
Lithofacies LF23: Wispy-laminated, Burrowed,
Skeletal Wackestone-Mudstone (WBSW)
 low diversity, abundant, and small-size fauna
 foraminifera, mollusks, echinoderms, sponge
spicules, and ostracods
 peloids
 blackened pebbles and blackened grains
 dolomite rich
 pyrite rich
 siliciclastics present
 wackestone to mudstone texture with minor
packstone interlayer
 wispy and nodular bedded
 burrows and bioturbation
 paleosols and caliche crusts
 rhizoliths (root tubes) and desiccation cracks
 frequent subaerial exposure
 shallow subtidal, low-energy restricted platform
 inner-ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 lower and middle dense zones (see Figure 25)
Lithofacies LF24: Burrowed, Bioturbated, Skeletal
Packstone (BBSP)
 low diversity, abundant, and small-size fauna
 foraminifera, echinoderms, sponge spicules, and
ostracods
 peloids
 blackened pebbles and blackened grains
 pyrite
 packstone texture with minor wackestone inter-
layer
 burrows and bioturbation
 shallow subtidal, low- to moderate-energy re-
stricted platform
 inner-ramp, restricted lagoonal deposits
 low to moderate porosity and low matrix
permeability
 lower and middle (uncommon) dense zones (see
Figure 26)
Lithofacies LF25: Burrowed, Bioturbated,
Orbitolinid, Skeletal Packstone (BBOSP)
 low diversity and high abundance of fauna
 discoidal orbitolinids and other foraminifera, green
algae (dasycladacean), echinoderms, sponge spic-
ules, and ostracods
 peloids
 blackened pebbles and blackened grains
 pyrite
 glauconite
 packstone texture with minor wackestone inter-
layer
 burrows and bioturbation
 shallow subtidal, low- to moderate-energy re-
stricted platform
 inner ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 upper dense zone (Hawar) (see Figure 27)
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 165
Lithofacies LF26: Burrowed, Bioturbated,
Orbitolinid, Skeletal Wackestone (BBOSW)
 low diversity and high abundance of fauna
 discoidal orbitolinids and other foraminifera, green
algae (dasycladacean), echinoderms, sponge spic-
ules, and ostracods
 blackened pebbles and blackened grains
 pyrite
 glauconite
 wackestone texture with minor packstone inter-
layer
 burrows and bioturbation
 shallow subtidal, low-energy restricted platform
 inner-ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 upper dense zone (Hawar) (see Figure 28)
Lithofacies LF27: Burrowed, Bioturbated, Skeletal
Wackestone–Mudstone (BBSW)
 low diversity, abundant, and small-size fauna
 foraminifera, mollusks, echinoderms, sponge
spicules, and ostracods
 peloids
 blackened pebbles and blackened grains
 pyrite
 wackestone to mudstone with minor packstone
interlayer
 burrows and bioturbation
 shallow subtidal, low-energy restricted platform
 inner ramp, restricted lagoonal deposits
 low porosity and low matrix permeability
 lower and middle dense zones (see Figure 29)
Sequence Stratigraphy-keyed
Lithofacies Distribution
Lithofacies distribution in the Lower and Upper
Kharaib Reservoir Units is closely linked to the over-
all sequence-stratigraphic framework. The transgres-
sive systems tracts of both the Lower and Upper Kha-
raib sequences are clearly more mud dominated (high
porosity but low to moderate permeability), whereas
the grain-dominated lithofacies types that show
excellent porosity and permeability are nearly ex-
clusively restricted to the highstand systems tracts
(Figure 4). The identified 16 parasequence sets that
build the Lower (4 parasequence sets) and Upper
(8 parasequence sets) Kharaib Reservoir Units and the
Lower Shuaiba Reservoir Unit (4 parasequence sets)
define 16 reservoir subunits; the basic building blocks
for the geological (static) and reservoir (dynamic)
model. A generalized southwest–northeast cross sec-
tion shows the vertical and horizontal lithofacies
and facies association distribution of the Lower and
Upper Kharaib Reservoir Units as well as the Lower
Shuaiba Reservoir Unit in the established sequence-
stratigraphic framework (Figure 30). The lateral and
horizontal variations of the grain-dominated facies
associations and lithofacies types in the highstand
systems tracts of the Lower and Upper Kharaib se-
quences are evident (Figure 30).
KHARAIB FORMATION OUTCROP ANALOGS AT
WADI RAHABAH, MUSANDAM PENINSULA
(RAS AL-KHAIMAH)
The Musandam Peninsula consists of a series of
upthrusted blocks (Hagab thrust) of Permian to Cre-
taceous strata. The dominant structures in the pen-
insula are north – south arcuate thrust faults, asso-
ciated tear faults, and north –south folds. A slight
regional northward dip of the peninsula toward Iran
exists. The Ruus Al Jibal forms the backbone of the
peninsula, consisting of a series of upthrusted blocks
(Hudson, 1960). One of these upthrusted blocks is
the 4760-ft (1450-m) high Jabal Hagab to the east of
the peninsula opposite to the city of Ras Al-Khaimah.
Rock exposures at Wadi Rahabah (north of the city
of Ras Al-Khaimah) are time equivalent to the Lower
Cretaceous upper Thamama Group, the most prolific
onshore oil- and gas-producing zones in Abu Dhabi.
The Lower Cretaceous Kahmah Group (Oman) and
Musandam Group M4 (Musandam Peninsula) are
time equivalent to the Thamama Group (United
Arab Emirates, Saudi Arabia, and Bahrain), which con-
tains (from oldest to youngest) the reservoirs of the
Habshan, Lekhwair, Kharaib, and Shuaiba forma-
tions of the subsurface of the United Arab Emirates.
Outcrop analogs of subsurface reservoirs allow for
a detailed investigation of the facies architecture
and structure of these carbonate bodies, keyed to the
overall sequence-stratigraphic framework (Alsharhan
and Nairn, 1993; Borgomano et al., 2002; van Buchem
et al., 2002; Hillgärtner et al., 2003; Immenhauser
et al., 2004; Strohmenger et al., 2004b; Suwaina et al.,
2004).
The 21 lithofacies types that were defined for the
subsurface Lower and Upper Kharaib and Lower
Shuaiba Reservoir Units and the lower, middle, and
upper dense zones (Strohmenger et al., 2004a, b, c) are
broadly applicable to the Kharaib and Lower Shuaiba
166 / Strohmenger et al.

FIGURE 30. Southwest – north-

east cross section through
Field B showing the established
sequence-stratigraphic frame-
work and the vertical and hori-
zontal distribution of facies
associations (lithofacies 1 [RPR] +
lithofacies 2 [RPF]; lithofacies 4
[SPG] + lithofacies 5 [CgSG] +
lithofacies 6 [CgASR]; and litho-
facies 9 [OSP] + lithofacies 11
[OSW]) and lithofacies types
(lithofacies 3: SPP; lithofacies 7:
ASPF; lithofacies 8: ASFB; litho-
facies 10: SPWP; lithofacies 12:
SW; and lithofacies 13: FSW).
Transgressive systems tract
(TST): green color; highstand
systems tract (HST): orange
color (see Figure 4).

thymetry and energy. The top

of a parasequence is deter-
mined by a flooding surface.
A sharp change in lithofacies
from shallow high-energy shoals
to slightly deeper water mud-
dominated lagoonal deposits
is recognized across parase-
quence boundaries. Mud-
dominated rocks grade upward
into packstone and floatstone
and subsequently into grain-
stone and rudstone with abun-
dant rudist debris. These rocks
indicate a progressive increase
in energy and shallowing of
the depositional environment.
The top of the parasequence
is marked either by well-sorted
grainstone and packstone
with abundant miliolids or
by deposits of the intershoal-
formations rock exposures. The depositional envi- protected areas that may become restricted and ex-
ronments do not occur randomly in time and space. hibit mud-dominated rocks with abundant forami-
They occur systematically in specific systems tracts nifera, green dasycladacean algae, small mollusks,
of the third-order composite sequences (Strohmenger and sparse rudists. The overlying parasequence com-
et al., 2004a, b, c; Suwaina et al., 2004; Vaughan et al., monly indicates a change to deeper and quieter de-
2004). positional environments as described, or it displays
The southeast wall of the wadi shows a carbonate a more gradual deepening with initial high-energy
succession that corresponds to the Lower and Upper deposits above a reactivation surface.
Kharaib Reservoir Units (Strohmenger et al., 2004a, Composite sequence boundary K70_SB at the
b, c; Suwaina et al., 2004) (Figure 31A). Meter-scale top of the Upper Kharaib sequence (top Upper Kha-
lithofacies stacking patterns reflect variations in ba- raib Reservoir Unit) displays pedogenic overprint of
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 167

FIGURE 31. Southeast wall of Wadi Rahabah outcrop. (A) Shown here is the Kharaib Formation measured section (base
lower dense zone to top upper dense zone; 250 ft [76 m]). (B) The top of the Upper Kharaib Reservoir (Upper Kharaib
sequence) displays pedogenic overprint of limestone beds (peds) below sequence boundary K70_SB. (C) Base of lower
dense zone showing erosive surface and mud cracks, corresponding to sequence boundary K50_SB.

limestone beds below the sequence boundary (paleo-
sol: peds; Figure 31B). Composite sequence boundary
K50_SB at the base of the Lower Kharaib sequence
(base lower dense zone) shows an erosive surface dis-
playing desiccation cracks (Figure 31C).
Outcrop analogs are a valuable source of data for
reservoir characterization. They are particularly use-
ful for visualizing interwell variability because sub-
surface wellbores represent discrete, widely spaced
data. In other words, outcrops provide the continu-
ous, large-scale coverage of a seismic line and provide
the fine-scale resolution of core measurements. Mod-
eling input parameters such as lateral and vertical
continuity (variogram range), object dimensions (as-
pect ratios), detailed reservoir architecture (layering),
facies relationships, and the nature and extent of
diagenetic features can be observed and measured.
To better understand reservoir properties of the
Upper Thamama producing zones in Abu Dhabi, a
geologic model was developed from descriptions of
time-equivalent, analogous outcrops in the Emirate
of Ras Al-Khaimah. Parasequences, maximum flood-
ing surfaces, sequence boundaries, and lithofacies were
described in four measured sections and tied to the
subsurface data of Field B. Porosity measurements de-
rived from logs, core porosity, and core permeability
data from an Abu Dhabi oil field (Field B) were ap-
plied to similar lithofacies observed in outcrop. These
subsurface data were used in the outcrop 3-D model to
more precisely reflect subsurface reservoir behavior
because outcrops in the United Arab Emirates expe-
rienced a significantly different burial history relative
to the subsurface reservoirs and, thus, have different
porosity and permeability (Strohmenger et al., 2004b;
Suwaina et al., 2004). The objective of the reservoir
modeling was to investigate the impact of reservoir
architecture, porosity-permeability relationships, facies,
and scale-up on the development of quantitative 3-D
geologic and fluid flow-simulation models. The use of
continuously exposed outcrops that are time equivalent
and have similar geology to the subsurface Upper
Thamama reservoirs of Abu Dhabi provide an opportu-
nity for improving geologic models of these reservoirs.
For the stratigraphic interval studied in outcrop
(base lower dense zone to top upper dense zone), four
composite reservoir facies (three permeable reservoir
facies and one dense facies) adequately described the
flow behavior. These reservoir facies were modeled
168 / Strohmenger et al.

FIGURE 32. Integration of outcrop data into geological, porosity, and permeability (fluid-flow) models. (A) North wall
of Wadi Rahabah showing clinoforms in the Upper Kharaib sequence. (B) Information gathered from outcrop studies
(occurrence of clinoforms) should be incorporated into geological (static) and reservoir (dynamic) models. (C) Porosity
model showing porosity distribution and the influence of clinoforms. High porosity is shown in red. No porosity is shown
in pink. (D) Permeability model showing resulting permeability distribution and the influence of clinoforms. Fluids will
encounter more baffles and travel more slowly through these dipping layers than through the flat-lying proportional
layers. High permeability is shown in red. No permeability is shown in pink.

and were scaled up into a simulation model. In addi-
tion to the four measured sections treated as wells, six
pseudowells were generated to help constrain the
mapping and modeling algorithms. These pseudo-
wells represented copies of the measured sections and
contained realistic unit thickness changes based on
outcrop observations. Multiple scenarios, including
various injection patterns, scale-up methods, and lo-
cations of thin, high-permeability streaks, were inves-
tigated. Results mimic known behavior in analogous
producing fields, and the process of going from rock
data to simulation provide a useful training tool for
reservoir characterization methods and techniques
(Vaughan et al., 2004).
The north and south wall of the wadi show clino-
forms in the Upper Kharaib sequence. Local clusters
of fractures and offset of clinoforms by minor faults
can also be observed (Figure 32A). Even if prograda-
tional patterns (clinoforms) cannot be identified in
core material, the information gathered from out-
crop data (Figure 32A) should be used to build re-
alistic geological (static; Figure 32B) and reservoir (dy-
namic; Figure 32C, D) models (Strohmenger et al.,
2004a, b, c; Suwaina et al., 2004; Vaughan et al., 2004).
Porosity (Figure 32C) and permeability (Figure 32D)
models show that low-angle clinoforms will act as
baffles, and fluid flow through these dipping layers
will therefore be slower compared to flat-lying pro-
portional layers. Sector model simulation results ad-
dress issues of scale-up, stratigraphic architecture, and
diagenesis, as well as alternative recovery mecha-
nisms and field development scenarios.
CONCLUSIONS
The Kharaib Formation (Barremian and early Ap-
tian) contains two reservoir units (Lower and Upper
Kharaib Reservoir Unit) separated and encased by
three so-called dense zones (lower, middle, and up-
per dense zones). It is part of the late transgressive
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 169
sequence set of a second-order supersequence, made
up of two third-order composite sequences (base lower
dense zone to base upper dense zone or Hawar). The
overlying Lower Shuaiba Reservoir Unit belongs to
the late transgressive sequence set and the early high-
stand sequence set and is made up of one third-order
composite sequence. Nineteen fourth-order parase-
quence sets were identified in core and on well-log data.
Four parasequence sets build into the Lower Kharaib,
ten parasequence sets build into the Upper Kharaib,
and five parasequence sets build into the Lower
Shuaiba third-order composite sequences (Figure 4).
These parasequence sets show predominantly aggra-
dational and progradational stacking patterns, typical
of greenhouse cycles (Sarg et al., 1999). The identified,
chronostratigraphically bounded parasequence sets
were used to subdivide the three reservoir units into
a total of 16 subunits.
As an alternative sequence-stratigraphic interpre-
tation, the major third-order composite sequence
boundaries (Figures 4, 6) may actually occur at the
tops of the dense zones (lower, middle, and upper
dense zones) and not, as conventionally interpreted
(van Buchem et al., 2002) and discussed above, be-
low the dense zones (Figure 6).
Thirteen lithofacies types correspond to the three
reservoir units (Lower and Upper Kharaib Reservoir
Unit and Lower Shuaiba Reservoir Unit) and eight
lithofacies types occur in the three dense zones (lower,
middle, and upper dense zones; Figures 7–29). Mud-
dominated lithofacies types dominate the transgres-
sive systems tracts, whereas grain-dominated, highly
porous and permeable lithofacies types make up the
highstand systems tracts of the Lower and Upper Kha-
raib third-order composite sequences (Figure 30).
These lithofacies types are broadly applicable to the
Kharaib and Shuaiba formations rock exposures (time-
equivalent to the Upper Thamama reservoir units)
studied in Wadi Rahabah, Ras Al-Khaimah (United
Arab Emirates).
Many of the so-called stylolitic intervals correspond
to major facies changes related to third-, fourth-, and
fifth-order sequence boundaries, parasequence set
boundaries, and parasequence boundaries. Early dia-
genetic processes follow the sequence-stratigraphic
framework and, therefore, can be predicted away from
well control.
Outcrops of the Kharaib Formation at Wadi
Rahabah are of seismic scale and, thus, provide field-
scale cross sections that help to refine sequence-
stratigraphic and facies models and aid to the un-
derstanding of reservoir geometry distribution and
reservoir continuity in the subsurface (Vaughan et al.,
2004) (Figures 31, 32). Outcrops are particularly use-
ful for filling the holes common in subsurface well
data sets used as input to geologic (static) and reser-
voir (dynamic) models.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the manage-
ments of Abu Dhabi Company for Onshore Oil Op-
erations (ADCO), Abu Dhabi National Oil Company
(ADNOC), and ExxonMobil Exploration Company
for permission to publish this study. We thank Prasad
Patannakara (ADCO) and his team for drafting the fig-
ures. Appreciation is extended to Ismail Al-Mansouri,
Majeed Haq, Mohammed Hilal, Hassan Ahmed, and
Abdulsalam Khoury from ADCO core facility (Mus-
safah) for their help in preparing and laying out
the core material and for preparing thin sections.
Sean A. Guidry, Christine I. Gonzalez, Martin A. Herr-
mann, Lisa A. Roehl, Chengije Liu, Susan M. Agar,
Jerry J. Kendall, and Mike G. Kozar (ExxonMobil), as
well as Andrew Clark, Srikant Guruswamy, Hafez H.
Hafez, Jorge S. Gomes, Michel J.-M. Rebelle, William L.
Soroka, and Mohammed Ayoub (ADCO) are thanked
for their valuable discussions. We extend special
thanks to Omar Suwaina (ADNOC) for his valuable
contributions to this study. We appreciate the help-
ful comments and suggestions of reviewers José E.
de Matos, Jürgen Grötsch, and Jonathan Kaufman,
which greatly improved the manuscript.
REFERENCES CITED
Alsharhan, A. S., 1989, Petroleum geology of the United
Arab Emirates: Journal of Petroleum Geology, v. 12,
no. 3, p. 253 – 288.
Alsharhan, A. S., and A. E. M. Nairn, 1993, Carbonate plat-
form models of Arabian Cretaceous reservoirs, in J. A. T.
Simo, R. W. Scott, and J.-P. Masse, eds., Cretaceous
carbonate platforms: AAPG Memoir 56, p. 173 – 184.
Alsharhan, A. S., and A. E. M. Nairn, 1997, Sedimentary
basins and petroleum geology of the Middle East:
Amsterdam, Elsevier, 843 p.
Boichard, R., A. S. Al Suwaidi, and H. Karakhanian, 1994,
Sequence boundary types and related porosity evolu-
tions: Example of the Upper Thamama Group in field
‘‘A’’ (offshore Abu Dhabi, UAE): 6th Abu Dhabi Interna-
tional Petroleum Exhibitions & Conference, Abu Dhabi
Society of Petroleum Engineers Paper 76, Abu Dhabi,
p. 417 – 428.
Borgomano, J., J.-P. Masse, and S. Al Maskiry, 2002, The
lower Aptian Shuaiba carbonate outcrops in Jebel
Akhdar, northern Oman: Impact on static modeling
170 / Strohmenger et al.
for Shuaiba petroleum reservoirs: AAPG Bulletin, v. 86,
no. 9, p. 1513 – 1529.
Burgess, C. J., and C. K. Peter, 1985, Formation, distribution
and prediction of stylolites as permeability barriers
in the Thamama Group, Abu Dhabi: 4th Middle East
Oil Show, Bahrain, Society of Petroleum Engineers
Paper 13698, p. 165 – 174.
Davies, R. B., D. M. Casey, A. D. Horbury, P. R. Sharland,
and M. D. Simmons, 2002, Early to mid-Cretaceous
mixed carbonate-clastic shelfal systems: Examples,
issues and models from the Arabian plate: GeoArabia,
v. 7, no. 3, p. 541 – 598.
Droste, H., and M. van Steenwinkel, 2004, Stratal geome-
tries and patterns of platform carbonates: The Creta-
ceous of Oman, in G. P. Eberli, J. L. Masaferro, and J. F.
Sarg, eds., Seismic imaging of carbonate reservoirs and
systems: AAPG Memoir 81, p. 185 – 206.
Dupraz, C., and A. Strasser, 1999, Microbialites and micro-
encrusters in shallow coral bioherms (middle – late
Oxfordian, Swiss Jura Mountains): Facies, v. 40, p. 101 –
130.
Galloway, W. E., 1989, Genetic stratigraphic sequences in
basin analysis: I — Architecture and genesis of flood-
ing-surface bounded depositional units: AAPG Bulle-
tin, v. 73, no. 2, p. 125 – 142.
Granier, B., 2000, Lower Cretaceous stratigraphy of Abu
Dhabi and the United Arab Emirates — A reappraisal:
9th Abu Dhabi International Petroleum Exhibition &
Conference Proceedings, Paper 0918, Abu Dhabi,
p. 526 – 535.
Granier, B., A. S. Al Suwaidi, R. Busnardo, S. K. Aziz, and
R. Schroeder, 2003, New insight on the stratigraphy
of the ‘‘Upper Thamama’’ in offshore Abu Dhabi
(U.A.E.): Carnets de Géologie, Article 2003/5, p. 1 – 17.
Grötsch, J., O. Al-Jeelani, and Y. Al-Mehairi, 1998a, Inte-
grated reservoir characterization of a giant Lower
Cretaceous oil field, Abu Dhabi, U.A.E.: 8th Abu Dhabi
International Petroleum Exhibition & Conference
Proceedings, Society of Petroleum Engineers Paper
49454, Abu Dhabi, p. 77 – 86.
Grötsch, J., I. Billing, and V. Vahrenkamp, 1998b, Carbon-
isotope stratigraphy in shallow-water carbonates:
implications for Cretaceous black-shale deposition:
Sedimentology, v. 45, no. 4, p. 623 – 634.
Haq, B. U., and A. M. Al-Qahtani, 2005, Phanerozoic cycles
of sea-level change on the Arabian platform: Geo-
Arabia, v. 10, no. 2, p. 127 – 160.
Haq, B. U., J. Hardenbol, and P. R. Vail, 1988, Mesozoic
and Cenozoic chronostratigraphy and cycles of sea-
level change, in C. K. Wilgus, B. S. Hastings, C. G. St.
C. Kendall, H. W. Posamentier, C. A. Ross, and J. C.
Van Wagoner, eds., Sea-level changes: An integrated
approach: SEPM Special Publication 42, p. 71 – 108.
Hillgärtner, H., F. S. P. van Buchem, F. Gaumet, P. Razin,
B. Pittet, J. Grötsch, and H. Droste, 2003, The Barre-
mian – Aptian evolution of the eastern Arabian carbon-
ate platform margin (northern Oman): Journal of
Sedimentary Research, v. 73, no. 5, p. 756 – 773.
Hudson, R. G. S., 1960, The Permian and Triassic of the
Oman Peninsula, Arabia: Geological Magazine, v. 18,
no. 4, p. 299 – 309.
Immenhauser, A., H. Hillgärtner, and E. Van Bentum, 2005,
Microbial-foraminiferal episodes in the early Aptian of
the southern Tethyan margin: Ecological significance
and possible relation to oceanic anoxic event 1a: Sedi-
mentology, v. 52, no. 1, p. 77 – 99.
Immenhauser, A., et al., 2004, Barremian – lower Aptian
Qishn Formation, Haushi-Huqf area, Oman: A new
outcrop analogue for the Kharaib/Shu’aiba reservoirs:
GeoArabia, v. 9, no. 1, p. 153 – 194.
Koepnick, R. B., 1984, Distribution and vertical perme-
ability of stylolites within a Lower Cretaceous carbon-
ate reservoir, Abu Dhabi, United Arab Emirates —
Stylolites and associated phenomena: Relevance to
hydrocarbon reservoirs, Abu Dhabi, U.A.E.: Abu Dhabi
Reservoir Research Foundation Special Publications,
p. 261 – 278.
Melville, P., O. Al Jeelani, S. Al Menhali, and J. Grötsch,
2004, Three-dimensional seismic analysis in the char-
acterization of a giant carbonate field, onshore Abu
Dhabi, United Arab Emirates, in G. P. Eberli, J. L. Masa-
ferro, and J. F. Sarg, eds., Seismic imaging of carbonate
reservoir systems: AAPG Memoir 81, p. 123– 148.
Mitchum Jr., R. M., 1977, Seismic stratigraphy and global
changes in sea level: Part 11 — Glossary of terms used
in seismic stratigraphy, in C. E. Payton, ed., Seismic
stratigraphy — Applications to hydrocarbon explora-
tion: AAPG Memoir 26, p. 205 – 212.
Nelson, R. A., 1984, Geologic analysis of naturally fractured
reservoirs: Contributions in petroleum geology and
engineering: Oxford, Gulf Publishing Company, 352 p.
Park, W. C., and E. H. Schot, 1968, Stylolites: Their nature
and origin: Journal of Sedimentary Petrology, v. 38,
p. 175 – 191.
Pittet, B., F. van Buchem, H. Hillgärtner, J. Grötsch, and
H. Droste, 2002, Ecological succession, paleoenvi-
ronmental change, and depositional sequences of
Barremian – Aptian shallow-water carbonates in north-
ern Oman: Sedimentology, v. 49, no. 3, p. 555– 581.
Rebelle, M., C. J. Strohmenger, A. Ghani, K. Al-Mehsin, and
A. Al-Mansouri, 2004, Lower Cretaceous Upper Tha-
mama reservoir high-resolution sequence stratigraphy,
United Arab Emirates (abs.): GeoArabia, v. 9, no. 1, p. 136.
Sarg, J. F., 1988, Carbonate sequence stratigraphy, in C. K.
Wilgus, B. S. Hastings, C. G. St. C. Kendall, H. W.
Posamentier, C. A. Ross, and J. C. Van Wagoner, eds.,
Sea-level changes: An integrated approach: SEPM
Special Publication 42, p. 155 – 181.
Sarg, J. F., J. R. Markello, and L. J. Weber, 1999, The second-
order cycle, carbonate-platform growth, and reservoir,
source, and trap prediction, in P. M. Harris, A. H. Saller,
and J. A. T. Simo, eds., Advances in carbonate sequence
stratigraphy: Application to reservoirs, outcrops, and
models: SEPM Special Publication 63, p. 11 – 34.
Sattler, U., A. Immenhauser, H. Hillgärtner, and M. Esteban,
2005, Characterization, lateral variability and lateral
 Sequence Stratigraphy and Reservoir Characterization of Upper Thamama Reservoirs / 171
extend of discontinuity surfaces on a carbonate plat-
form (Barremian to lower Aptian, Oman): Sedimen-
tology, v. 52, no. 2, p. 339 – 361.
Sharland, P. R., R. Archer, D. M. Casey, R. B. Davies, S. H.
Hall, A. P. Heward, A. D. Horbury, and M. D. Simmons,
2001, Arabic plate sequence stratigraphy: GeoArabia
Special Publication 2, 371 p.
Sharland, P. R., R. Archer, D. M. Casey, R. B. Davies, M. D.
Simmons, and O. E. Sutcliffe, 2004, Arabic plate se-
quence stratigraphy — Revisions to SP2: GeoArabia,
v. 9, no. 1, p. 199 – 214.
Strasser, A., 1984, Black-pebble occurrence and genesis in
Holocene carbonate sediments (Florida Keys, Baha-
mas, and Tunisia): Journal of Sedimentary Petrology,
v. 54, no. 4, p. 1097 – 1109.
Strasser, A., and E. Davaud, 1983, Black pebbles of the
Purbeckian (Swiss and French Jura): Lithology, geo-
chemistry and origin: Eclogae Geologicae Helveticae,
v. 76, p. 551 – 580.
Strohmenger, C. J., L. J. Weber, A. Ghani, K. Al-Mehsin,
and O. Al-Jeelani, 2004a, Sequence stratigraphy and
reservoir characterization of the Lower Cretaceous
Kharaib Formation, Abu Dhabi (abs.): GeoArabia, v. 9,
no. 1, p. 136.
Strohmenger, C. J., L. J. Weber, A. Ghani, M. Rebelle, K. Al-
Mehsin, O. Al-Jeelani, A. Al-Mansoori, and O. Suwaina,
2004b, High-resolution sequence stratigraphy of the
Kharaib Formation (Lower Cretaceous, U.A.E.): 11th
Abu Dhabi International Petroleum Exhibition &
Conference Proceedings, Society of Petroleum Engi-
neers Paper 88729, Abu Dhabi, 10 p.
Strohmenger, C. J., L. J. Weber, A. Ghani, A. Al-Mansoori, K.
Al-Mehsin, O. Suwaina, O. Al-Jeelani, and L. Vaughan,
2004c, Sequence stratigraphy and reservoir character-
ization of the Kharaib Formation comparing outcrop
and subsurface data (Lower Cretaceous, U.A.E.) (abs.):
AAPG International Conference and Exhibition
(Abstracts), Cancun, p. A75.
Suwaina, O. A., L. J. Weber, C. J. Strohmenger, L. Vaughan,
A. Al-Mansoori, S. Khan, and A. Ghani, 2004, Sequence
stratigraphy and reservoir characterization of the
Thamama reservoirs and outcrop equivalents: A core
workshop and field seminar (abs.): GeoArabia, v. 9,
no. 1, p. 137 – 138.
Vahrenkamp, V. C., 1996, Carbon isotope stratigraphy of
the Upper Kharaib and Shuaiba formations: Impli-
cations for the Early Cretaceous evolution of the Ara-
bian Gulf region: AAPG Bulletin, v. 80, no. 5, p. 647 –
662.
Vail, P. R., 1987, Seismic stratigraphy interpretation using
sequence stratigraphy: Part 1. Seismic stratigraphy in-
terpretation procedure, in A. W. Bally, ed., Atlas of
seismic stratigraphy, v. 1: AAPG Studies in Geology 27,
p. 1 – 10.
Vail, P. R., R. M. Mitchum Jr., R. G. Todd, J. M. Widmier, S.
Thompson III, J. B. Sangree, J. N. Bubb, and W. G.
Hatlelid, 1977, Seismic stratigraphy and global changes
of sea level: Part 1. Overview, in C. E. Payton, ed.,
Seismic stratigraphy — Applications to hydrocarbon
exploration: AAPG Memoir 26, p. 49 – 212.
Vail, P. R., F. Audemard, S. A. Bowman, P. N. Eisner, and
C. Perez-Cruz, 1991, The stratigraphic signatures of
tectonics, eustasy and sedimentology — An overview,
in G. Einsele, W. Ricken, and A. Seilacher, eds., Cycles
and events in stratigraphy: Berlin, Springer, p. 617 –
659.
van Buchem, F. S. P., B. Pittet, H. Hillgärtner, J. Grötsch,
A. I. Al Mansouri, I. M. Billing, H. H. J. Droste, W. H.
Oterdoom, and M. van Steenwinkel, 2002, High-
resolution sequence stratigraphic architecture of Bar-
remian/Aptian carbonate systems in northern Oman
and the United Arab Emirates (Kharaib and Shuaiba
formations): GeoArabia, v. 7, no. 3, p. 461 – 500.
Van Wagoner, J. C., R. M. Mitchum Jr., H. W. Posamentier,
and P. R. Vail, 1987, Seismic stratigraphy interpreta-
tion using sequence stratigraphy: Part 2. Key defini-
tions of sequence stratigraphy, in A. W. Bally, ed., Atlas
of seismic stratigraphy: AAPG Studies in Geology 27,
v. 1, p. 11 – 14.
Van Wagoner, J. C., H. W. Posamentier, R. M. Mitchum
Jr., P. R. Vail, J. F. Sarg, T. S. Loutit, and J. Hardenbol,
1988, An overview of the fundamentals of sequence
stratigraphy and key definitions, in C. K. Wilgus, B. S.
Hastings, C. G. St. C. Kendall, H. W. Posamentier,
C. A. Ross, and J. C. Van Wagoner, eds., Sea-level
changes: An integrated approach: SEPM Special
Publication 42, p. 39 – 46.
Vaughan, R. L., S. A. Khan, L. J. Weber, O. Suwaina, A. Al-
Mansoori, A. Ghani, C. J. Strohmenger, M. A. Herrmann,
and D. Hulstrand, 2004, Integrated characterization of
UAE outcrops: From rocks to fluid flow simulation: 11th
Abu Dhabi International Petroleum Exhibition & Con-
ference Proceedings, Society of Petroleum Engineers
Paper 88730, Abu Dhabi, 17 p.
Yose, L. A., S. Bachtel, L. J. Weber, C. J. Strohmenger, A.
Al-Mansoori, and O. Suwaina, 2004a, Integrated ap-
proaches to carbonate reservoir characterization and
prediction: Examples from the United Arab Emirates
fields and outcrops (abs.): GeoArabia, v. 9, no. 1, p. 146.
Yose, L. A., et al., 2004b, New frontiers in 3D seismic char-
acterization of carbonate reservoirs: Examples from a
supergiant field in Abu Dhabi: 11th Abu Dhabi Inter-
national Petroleum Exhibition & Conference Proceed-
ings, Society of Petroleum Engineers Paper 88689, Abu
Dhabi, 16 p.
Yose, L. A., A. S. Ruf, C. J. Strohmenger, J. S. Schuelke, A.
Gombos, I. Al-Hosani, S. Al-Maskary, G. Bloch, Y. Al-
Mehairi, and I. G. Johnson, 2006, Three-dimensional
characterization of a heterogeneous carbonate reser-
voir, Lower Cretaceous, Abu Dhabi (United Arab
Emirates), in P. M. Harris and L. J. Weber, eds., Giant
hydrocarbon reservoirs of the world: From rocks to
reservoir characterization and modeling: AAPG Mem-
oir 88/SEPM Special Publication, p. 173 – 212.
 Yose, L. A., A. S. Ruf, C. J. Strohmenger, J. S. Schuelke, A. Gombos, I. Al-Hosani,

5
S. Al-Maskary, G. Bloch, Y. Al-Mehairi, and I. G. Johnson, 2006, Three-
dimensional characterization of a heterogeneous carbonate reservoir,
Lower Cretaceous, Abu Dhabi (United Arab Emirates), in P. M. Harris
and L. J. Weber, eds., Giant hydrocarbon reservoirs of the world: From
rocks to reservoir characterization and modeling: AAPG Memoir 88/SEPM
Special Publication, p. 173 – 212.

Three-dimensional Characterization
of a Heterogeneous Carbonate
Reservoir, Lower Cretaceous,
Abu Dhabi (United Arab Emirates)
Lyndon A. Yose Jim S. Schuelke
ExxonMobil Qatar, Inc., Doha, Qatar ExxonMobil Upstream Research Company,
Houston, Texas, U.S.A.
Amy S. Ruf
ExxonMobil Upstream Research Company, Andy Gombos
Houston, Texas, U.S.A. ExxonMobil Upstream Research Company,
Houston, Texas, U.S.A.
Christian J. Strohmenger,
Ismail Al-Hosani, Shamsa Al-Maskary, Imelda G. Johnson
Gerald Bloch, and Yousuf Al-Mehairi ExxonMobil Exploration Company, Houston,
Abu Dhabi Company for Onshore Oil Texas, U.S.A.
Operations, Abu Dhabi, United Arab
Emirates

ABSTRACT

H
igh-effort three-dimensional (3-D) seismic data collected by the Abu
Dhabi Company for Onshore Oil Operations (ADCO) are some of the
highest quality data ever collected for a carbonate field. The 3-D seismic
data were integrated with core and log data to develop a new, volume-based
framework for enhanced reservoir characterization. The Lower Cretaceous
(Aptian) reservoir is positioned over a platform-to-basin transition and records
a diverse range of depositional facies and stratal geometries. Reservoir properties
vary predictably based on position along the platform-to-basin profile and po-
sition in the sequence-stratigraphic framework.
The Aptian reservoir interval (Shuaiba Formation) records a second-order
sequence set that is divided into five depositional sequences. Sequences 1 and
2 were deposited during the transgressive phase of the sequence set. These se-
quences are retrogradational, record the initial formation of a low-relief ramp,

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215877M882562

173
174 / Yose et al.

and are dominated by algal-prone facies. Ramp interior and margin facies of
the transgressive phase are characterized by high porosity and low perme-
ability because of mud-dominated textures and development of microporos-
ity. Sequence 3 was deposited during the highstand phase of the sequence set,
is mainly aggradational, and records the proliferation or rudists across the plat-
form top. Grain-dominated platform interior and margin facies of the high-
stand phase are the highest quality reservoir facies in the Shuaiba reservoir.
Sequences 4 and 5 were deposited during the late highstand phase of the se-
quence set. These sequences are progradational and record the progressive
downstepping of the platform margin onto a low-angle (1 – 28) slope. Clino-
forms of the late highstand phase are characterized by alternations of high and
low reservoir quality developed in response to relative sea level changes.
Sequence 6 was deposited during the second-order lowstand and forms the base
of the next overlying sequence set. Sequence 6 is composed primarily of fine-
grained siliciclastics and is a nonreservoir.
Results from the study have led to an improved understanding of platform
evolution and a volume-based framework for reservoir characterization. The in-
tegrated data set provides new insights on platform paleogeography, carbonate
facies architecture, and the geometry and mechanisms of carbonate platform
progradation. In the platform interior area, 3-D seismic data reveal a complex
mosaic of tidal channels, high-energy rudist shoals, and intershoal ponds that
impact reservoir sweep and conformance. At the basin margin, the seismic data
provide high-definition images of platform-margin clinoforms that impact
reservoir architecture and well-pair connectivity.
Business applications of the volume-based reservoir framework include
(1) use of 3-D seismic visualization technology for optimizing well placement,
identifying bypassed reservoirs, and evaluating reservoir connectivity; (2) inte-
gration of quantitative, volume-based seismic information into reservoir models;
(3) maximizing recovery through full integration of all subsurface data; and
(4) enhanced communication among geoscientists and engineers, leading to
improved reservoir management practices.

INTRODUCTION
The stratigraphic and diagenetic complexities in-
herent in carbonate reservoirs require accurate reser-
voir descriptions and models to optimize recovery.
Three-dimensional (3-D) seismic data provide the
only continuous source of information on reservoir
properties in the subsurface. With increasingly more
3-D seismic data available over carbonate fields, the
challenge is to maximize the value of the seismic data
for characterization of carbonate reservoir architec-
ture and rock properties (Eberli et al., 2003; Sarg and
Schuelke, 2003; Masafarro et al., 2004; Yose et al., 2004).
This study used high-quality seismic data collected
over a supergiant onshore field in Abu Dhabi to il-
lustrate the value of 3-D seismic data for carbonate
reservoir characterization. The field produces mainly
from the Lower Cretaceous (Aptian) Shuaiba reser-
voir and has been under production since the 1960s.
As the field matures, a more detailed understand-
ing of the reservoir is required to optimize recovery.
To address this need, the Abu Dhabi Company for
Onshore Oil Operations (ADCO) acquired a high-
effort, 3-D seismic survey over the field. A joint study
between ADCO and ExxonMobil was undertaken to
fully integrate the 3-D seismic data with core, well,
and production data, leading to a new volume-based
reservoir description. The integrated data set provides
new perspectives on carbonate platform evolution,
the response of carbonate systems to sea level change,
and the influence of carbonate facies and stratal ar-
chitecture on reservoir properties and architecture.
The volume-based reservoir framework provides a
powerful tool for evaluating full-field production and
performance issues, leading to improved reservoir
management.
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 175

FIGURE 1. Regional map showing major hydrocarbon accumulations in the United Arab Emirates and the locations of
Field A1 and Field A2. The fields are on separate structures, but both produce from the Shuaiba reservoir. The box
around the studied fields in the inset map highlights the area covered in Figure 2.

STUDY AREA
Field Location
The location of the study area in the United Arab
Emirates is shown in Figure 1. Onshore Abu Dhabi in-
cludes five supergiant fields, each of which is located
over large north-northeast–trending anticlinal struc-
tures. The present study is focused on the Lower Cre-
taceous Aptian reservoir interval (Shuaiba Formation)
of one of the onshore fields (Field A1 in Figure 1). As
shown in Figure 2, Field A1 occurs over a large, dou-
bly plunging anticline that is about 35  20 km (22 
12.5 mi) in dimension. A small field is also present
just south of Field A1 (Field A2 in Figure 2) that occurs
over a separate and much smaller closure. Fields A1
and A2 were evaluated jointly in this study because
the data sets overlap, and both fields are producing
mainly from the Shuaiba reservoir.
Field A1, discovered in 1962, is estimated to con-
tain more than 20 billion bbl of original oil in place
and is one of the ten largest carbonate fields in the
world. The field has been under peripheral water in-
jection since the mid-1970s. Pattern gas and water-
injection schemes are also being implemented in
parts of the field. A more detailed understanding of
reservoir architecture and connectivity is required to
optimize the new injection strategy and future field
development to maximize recovery.
Database
The study area provides a world-class data set and
presents an opportunity to develop and test high-end
reservoir characterization technologies. Greater than
500 wells exist in the field, more than 100 of which
are cored. Most wells contain a standard suite of open-
hole logs. More than 40 yr of production and per-
formance data are available to validate geologic in-
terpretations and calibrate reservoir models. Despite
the high number of wells in the field, the average well
176 / Yose et al.

FIGURE 2. Time structure map on the top of the reservoir interval at Field A1. The field is located over a large anticlinal
structure that plunges to the north and south. Field A2 is a small field on a separate closure developed just south of
Field A1. As shown, separate seismic surveys exist over each field.

spacing in the field is 1 km (0.62 mi). A 3-D seismic
survey was recently acquired over Field A1, covering
an area of nearly 2000 km2 (772 mi2). The 3-D seismic
data are 300-fold with 25-m (82-ft) trace spacing and
have a peak frequency of 30 –35 Hz at reservoir depth
(2440–2745 m; 8000– 9000 ft). A separate, but over-
lapping, seismic survey was also recently acquired
over Field A2 (Figure 1). Seismic data in the southern
survey are 150-fold and also provide high-quality
imaging of the Shuaiba reservoir. Subsequent figures
show extracts from the merged seismic surveys to
highlight stratigraphic and paleogeographic varia-
tions across the entire study area.
Data evaluated as part of the present study include
the 3-D seismic surveys, a strike and dip cross section
grid that intersects more than 100 wells, 25 cored
wells, and available production and performance data.
Regional Stratigraphy and Paleogeography
The stratal hierarchy and nomenclature used in the
present study is shown in Figure 3. The Shuaiba res-
ervoir interval is part of a second-order supersequence
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 177

FIGURE 3. Chronostratigraphic chart for the Lower Cretaceous of the United Arab Emirates. The Lower Cretaceous
comprises a hierarchy of second- and third-order depositional cycles. The reservoir interval at Field A1 spans most of
the Aptian and forms the second-order Aptian sequence set. The Aptian sequence set, in turn, is divided into five third-
order depositional sequences (1 – 5). A sixth depositional sequence (6) forms a composite lowstand sequence and is
part of the next overlying supersequence. OAE1a is a globally recognized oceanic anoxic event and correlates to the
maximum flooding interval of the Lower Cretaceous supersequence (coincident with the maximum flooding interval of
the Aptian sequence set). The Hardenbol et al. (1998) time sequence terminology is keyed to a nannofossil zonation.
Nannofossil dating in the present study confirms the position of the Ap 3, Ap 4, and Ap 5 time sequences (x in the time-
sequence nomenclature means unconfirmed). MFS = maximum flooding surface.

that spans the Valanginian to Aptian stages and can
be correlated across the Arabian platform (Sharland
et al., 2001). Based on interpretation of facies and
their stacking patterns in the present study, and by
Strohmenger et al. (2006), the Lower Cretaceous
supersequence is subdivided into three second-order
sequence sets (Figure 3). The Shuaiba reservoir is with-
in the Aptian sequence set, culminating in the second-
order supersequence boundary at the top. A compar-
ison of the sequence-stratigraphic – based framework
with the lithostratigraphic nomenclature is shown
in Figure 3. Lithostratigraphic nomenclature in the
Shuaiba varies regionally, underscoring the need for
a regional, sequence-stratigraphic framework based
on chronostratigraphy.
The Aptian paleogeography had a pronounced in-
fluence on depositional patterns in the study area. As
shown in Figure 4, Field A1 straddles the margin of
the Bab basin. Accordingly, the Shuaiba reservoir re-
cords a range of platform, margin, slope, and basin
environments and is ideally positioned to record the
full evolution of the platform-basin system. The Bab
basin differentiated in the early Aptian as a shallow
intrashelf seaway in the central Arabian platform
(Figure 4) (Alsharhan, 1985; Alsharhan and Nairn,
1993). The basin is bounded by an extensive carbon-
ate platform, and restricted, organic-rich facies were
deposited within the Bab basin proper. Carbonate
platform and margin facies in the Shuaiba form a
regional reservoir system, and organic-rich basinal
facies provide a potential source rock. The Bab basin is
filled by prograding carbonates and siliciclastics that
were deposited during the late Aptian sea level low-
stand. Shales and argillaceous carbonates of the Nahr
178 / Yose et al.

FIGURE 4. Aptian paleogeographic framework showing the extent of the Bab basin. Field A1 straddles the platform-basin
transition. A south-facing margin is also present in Field A2, just south of Field A1. The northern extent of the Bab
basin is uncertain. The basin may have been open to the north and connected to the Neo-Tethys open ocean, or isolated
in the Arabian platform and surrounded by a shallow, carbonate shelf (Sharland et al., 2001).

Umr Formation were deposited during the subse-
quent transgression and form a regional top seal over
the Shuaiba reservoir.
WORKFLOW
Improved recovery from carbonate reservoirs re-
quires a 3-D understanding of reservoir architecture
and properties. The 3-D seismic visualization environ-
ment provides a powerful platform for data integra-
tion and evaluation and a catalyst for multidisciplin-
ary interactions across geoscience and engineering
boundaries. A volume-based reservoir characterization
workflow was applied to maximize information from
the 3-D seismic data (Figure 5). Several key elements
of the workflow will be highlighted, including (1) post-
stack optimization of seismic data; (2) calibration of
seismic with core and log data; (3) application of vol-
ume interpretation techniques for rapid and accurate
definition of reservoir frameworks; (4) direct detection
and prediction of reservoir properties from seismic
data; and (5) integration of seismic data into reservoir
models.
As illustrated in Figure 6, a combination of volume
types and corendering techniques are used to maxi-
mize the information extracted from the 3-D seismic
data. The approach used in this study included a com-
bination of poststack filtering of the data to improve
the signal-to-noise ratio (compare Figure 6A, B), se-
lection and generation of optimum seismic attribute
volumes to image geologic features (Figure 6B), and
corendering of multiple seismic attributes to bring
out more detail (Figure 6C). The seismic discontinu-
ity volume shown in Figure 6B highlights the edges
of structural and stratigraphic features and is one of
the most useful volumes for characterization of the
Shuaiba reservoir. The volume mathematically quan-
tifies trace-to-trace variations in the seismic data,
with areas of high continuity shown in white and
areas of low continuity shown in black. In the current
example, facies geometries are resolved with clarity
comparable to aerial photographs of modern carbon-
ate systems (Figure 6D).
Calibration of the enhanced seismic data with core
and well data is required to understand the geologic
origin of seismic features and their impact on field
performance. In this study, images, such as those
shown in Figure 6, were used for optimum selection
of wells and cores to sample the full range of seismic
variability. Seismic information on reservoir archi-
tecture and rock properties is compared with reser-
voir performance data to validate predictions and
assess new opportunities. Seismic interpretations and
predictions are also incorporated into the reservoir
model to guide reservoir framework development, dis-
tribute reservoir features extracted from the seismic
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 179
FIGURE 5. Integrated seismic characterization workflow. Visualization technology allows for improved data integration
and evaluation during all stages of the workflow.
data, and to estimate rock properties outside of well
control. Many elements of the seismic workflow are
applicable to other carbonate reservoirs, although the
details will vary on a reservoir-specific basis, depend-
ing on such factors as reservoir complexity, data
quality and availability, and the business purpose.
STRUCTURAL FRAMEWORK
The impact of faults and fractures on reservoir
performance is a key uncertainty for many large, low-
relief structures in the Middle East, including on-
shore fields in Abu Dhabi. Prior to collection of 3-D
seismic data over the study area, the distribution and
geometry of faults were not well understood (Alshar-
han, 1993; Marzouk and Sattar, 1993). As illustrated
in Figure 7, the 3-D seismic data provide a new level
of detail on the structural framework. Several dif-
ferent seismic attribute volumes were evaluated as
part of the structural analysis, including disconti-
nuity, azimuthal amplitude variation with offset, and
fault enhancement volumes. The seismic discontinu-
ity volume (Figure 7) reveals a network of dominantly
northwest-trending faults that intersect the Lower
Cretaceous reservoirs. In the study area, faults are
organized into a system of grabens that are best de-
veloped in the central and northern regions of the
field. The faults are low offset, most commonly with
less than 18 ms (40 m; 130 ft) of vertical displace-
ment, and are commonly arranged en echelon along
strike (lateral offsets). These faults are consistent
with a regional fault system that has been mapped
across onshore and offshore Abu Dhabi (Marzouk
and Sattar, 1993; Johnson et al., 2005). The regional
faults are interpreted as low-offset wrench faults
that developed in response to west-northwest com-
pression during the Late Cretaceous ( Johnson et al.,
2005).
The impact of the faults and associated fractures
on fluid flow and reservoir connectivity in the study
area is uncertain at present. The Shuaiba is a matrix-
dominated flow system, but fracturing could enhance
fluid flow in some parts of the field and may become
more pronounced with time. As will be discussed fur-
ther, premature water breakthrough along the north-
ern platform-margin trend may be in part related to
enhanced fluid flow along faults. Additional work
and data are required to further evaluate the impact
of the faults and associated fractures on fluid flow.
180 / Yose et al.

FIGURE 6. Impact of data optimization on imaging of reservoir features in the platform interior area. The images of
(A – C) are identical time slices from three different attribute volumes. (A) Time slice of the original, unfiltered amplitude
data with reservoir features poorly defined. (B) Enhanced seismic volume that incorporates ExxonMobil proprietary
filtering and seismic attribute technologies (see text for explanation). (C) Image of seismic discontinuity corendered
with the isochron of the reservoir interval. Isochron thins highlight lower porosity areas (faster velocities), and isochron
thicks highlight higher porosity areas (slower velocities). The corendered image brings out even more detail on the
platform architecture and qualitative indications of reservoir quality. (D) Aerial photograph of modern carbonates of the
Great Barrier Reef illustrating ponds and channels similar to those imaged in seismic data (photo courtesy of National
Geographic Society).

The new 3-D fault framework can be used to monitor
future production trends and to guide future data
collection.
SEQUENCE-STRATIGRAPHIC OVERVIEW
Stratigraphic variability is the primary control on
Shuaiba reservoir architecture in the study area. An
overview of the new sequence model for the Shuaiba
is summarized in Figures 3 and 8. Regional perspec-
tives on stratigraphic and paleogeographic variations
in the Shuaiba, including summaries of previous work,
are provided in Sharland et al. (2001), van Buchem
et al. (2002), and Greselle and Pittet (2005).
Sequence Age Dating
Age control in the study area is based on unpub-
lished work by I. Billing and E. J. Oswald (1995, per-
sonal communication) and new age dating conducted
by ExxonMobil. Age dates from I. Billing and E. J.
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 181

FIGURE 7. Seismic characterization of structural framework. Image shown includes a 3-D discontinuity probe cut by
a vertical seismic amplitude profile. Most of the linear, black discontinuities on the surface of the 3-D cube are faults
(fault interpretations shown in inset). Faults are low offset, are best developed in the platform-margin area, and trend
parallel to the depositional strike of the margin.

Oswald (1995, personal communication) are based
on a combination of macrofossil dating (mainly ru-
dists) and carbon isotopes and provide an important
constraint on the position of the lower–upper Aptian
boundary across the study area. The ExxonMobil
age dating is based on nannofossils that established
ties to the global Aptian time sequences defined in
Hardenbol et al. (1998) (Ap terminology in Figure 3)
and provided a higher resolution time-stratigraphic
zonation for the lower and upper Aptian.
Results from the age dating are summarized in
Figure 3. Sequence 1 is of Ap 2 age, sequences 2 and 3
are of Ap 3 age, sequence 4 is of Ap 4 age, and se-
quence 5 is of probable Ap 5 age (T. C. Huang, 2004,
personal communication). The age of sequence 6 is
not confirmed, but is interpreted as Ap 6 time. The
overlying Nahr Umr Shale is Albian in age based on
regional age dates (Sharland et al., 2001). The position
of the Aptian stage boundaries in the study area is
uncertain (Figure 3). The Barremian–Aptian bound-
ary is interpreted to be within the Kharaib Forma-
tion (Strohmenger et al., 2006), and the Aptian–Albian
boundary is interpreted to be within the Bab Member
of the Shuaiba. These results are generally consistent
with age dates reported in Azer and Toland (1993),
Hughes (2000), Sharland et al. (2001), and van Buchem
et al. (2002). Based on these results, the time span of
the Aptian sequence set is estimated to be on the or-
der of 7 –9 m.y.
Stratal Hierarchy and Reservoir Impact
A hierarchy of stratigraphic cyclicity is developed
within the Shuaiba that has a pronounced influence
on reservoir properties. Stratal geometries developed
within the Aptian sequence set record a complete
transgressive-regressive cycle that developed in re-
sponse to a second-order rise and fall in sea level
(Figures 3, 8). A similar depositional pattern is rec-
ognized regionally within the Aptian (Sharland et al.,
2001; van Buchem et al., 2002; Droste, 2004). Second-
order stacking patterns control the overall architec-
ture of the Shuaiba platform.
The Aptian sequence set is subdivided into three
major phases of deposition, corresponding to the
182 / Yose et al.

FIGURE 8. Sequence- and seismic-stratigraphic overview of the Aptian sequence set. See Figure 13 for explanation facies
codes. (A) Sequence-stratigraphic framework and reservoir implications. The reservoir is subdivided into four major phases
of deposition, corresponding to phases of a long-term (second-order) sea level cycle, and six depositional sequences (1– 6),
corresponding to third-order sea level changes. Wells A – D are type logs for distinct areas of the field and are shown
in Figures 17 – 20, respectively. (B) Seismic geometries displayed on 2-D vertical slice (amplitude data, flattened on base
reservoir). Reservoir top is highlighted in blue, and reservoir base is highlighted in green. Amplitude anomalies in the
platform interior correspond to channel and pond features, as shown in time slice views in Figures 6 and 10.

transgressive, early highstand, and late highstand
phases of the second-order sea level cycle (Figure 8).
A subsequent lowstand phase forms the base of the
overlying second-order sequence set. Stratal geometry,
stacking patterns, and facies development vary sys-
tematically between each major phase of deposition
(Figure 8).
The Aptian sequence set is further subdivided into a
series of five depositional sequences that are each inter-
preted to record a third-order cycle of relative sea level
rise and fall. Third-order sequences control the distri-
bution of major reservoir barriers and flow compart-
ments in the Shuaiba reservoir (Figure 8A). Third-order
sequences are partitioned within the second-order
sequence set as follows: sequences 1 and 2 comprise
the transgressive phase; sequence 3 comprises the
highstand phase; and sequences 4 and 5 comprise the
late highstand phase. Sequence 6 forms the lowstand
phase of the next overlying sequence set.
As shown in Figure 9, third-order depositional
sequences are divided into systems tracts that are tied
to different phases of the third-order relative sea level
cycles. Most sequences display well-developed trans-
gressive and highstand systems tracts (TST, HST), al-
though the character of systems tracts varies signifi-
cantly based on position in the second-order cycle.
Higher frequency (fourth- or fifth-order) cycles are
best developed within the HSTs of third-order se-
quences and control the distribution and character of
flow units. Lowstand systems tracts (LST) (e.g., basin-
ward shifts of the carbonate system) are generally
absent in the transgressive and early highstand phases
of the Aptian sequence set because of a combination
of long-term accommodation increase, and the lack of
high-amplitude, shorter term sea level changes during
the Cretaceous greenhouse period. Lowstand systems
tracts are developed during the late highstand and
lowstand phases of the Aptian sequence set, as long-
term sea level was falling, and are represented by
basinward shifts in carbonate sedimentation. At the
second-order scale, sequences 4–6 are all basinally re-
stricted sequences (Figure 8). At the third-order scale,
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 183

FIGURE 9. Time-stratigraphic evolution of the Aptian sequence set, showing the sequential evolution of the carbonate
buildup. See Figure 13 for explanation of facies codes. HST = highstand systems tract; LST = lowstand systems tract;
TST = transgressive systems tract.

lowstand and TSTs that developed during the late platform interior (also see Figure 6). The width of the
highstand are difficult to distinguish because both platform interior (along depositional profile) varies
systems tracts are basinally restricted caused by long- throughout time, but is approximately 20 km (12.5 mi)
term sea level fall (see sequences 4 and 5, Figure 8). wide in the seismic time slice shown in Figure 10.

PALEOGEOGRAPHIC ELEMENTS Northern Platform Margin

Paleogeographic variations across the area had a
large impact on sequence and reservoir develop-
ment (Figure 4). Paleogeographic features are well
imaged by the seismic data (see cross sectional view
in Figure 8B and time slice view in Figure 10). An
introduction to the paleogeographic elements and
their seismic character is provided below.
Platform Interior
The central part of the field area is dominated by
stacked, shallow-water platform facies. Seismic ho-
rizons are concordant but highly discontinuous be-
cause of rapid lateral changes in facies and reservoir
properties. The seismic discontinuity time slice shown
in Figure 10 highlights a pond-channel complex that
gives rise to the complex seismic character of the
The northern platform margin is a broad zone of
relatively transparent seismic amplitude character. The
margin records the aggradation of high-energy facies,
with generally high reservoir quality. The discontinu-
ity time slice shows that the margin is characterized by
relatively high continuity, consistent with the high
and continuous reservoir quality trends observed in
log and core data. The width of the high-energy north-
ern margin averages 8 km (5 mi). Maximum relief on
the northern margin is estimated at 75 m (250 ft).
Clinoform Belt
The northern part of the field area comprises clino-
forms (sequences 4 and 5) that record more than 10 km
(6.25 mi) of progradation. The clinoforms are charac-
terized by alternations of low-angle (1–28), sigmoidal
seismic reflections and higher angle (3–48), oblique
184 / Yose et al.

FIGURE 10. Seismic discontinuity time slice from flattened volume (flattened on base reservoir). Seismic surveys over
Fields A1 and A2 are spliced together to create this volume. Seismic variations highlight the main paleogeographic
elements across the field area, each of which has distinct reservoir characteristics. Platform margins are recognized
north and south of the central buildup. Green line shows location of cross sections in Figure 8. Red dashes show location
of calibration wells that were evaluated in the study.

seismic reflections (Figures 8, 9). Reservoir properties the south. Based on platform-margin geometries in
and stratal geometries are highly variable in the clino- seismic data, maximum depositional relief on the
forms, resulting in significant reservoir heterogeneity. southern margin is estimated to be 30 m (100 ft).

Bab Basin Intrashelf Embayment

The Bab basin is characterized by restricted, low-
energy facies, reflected in a relatively thin, condensed
interval. This interval is separated from the overlying
clinoform sequences by a seismic downlap surface
(Figure 8). Basinal facies are characterized by highly
continuous and concordant seismic reflections.
Southern Platform Margin
The 3-D seismic data reveal a previously unrecog-
nized platform margin in the southern part of the
study area. The southern margin differentiates later
in time than the northern platform margin (Figure 8).
The southern platform margin is aggradational, with
very limited progradation. Clinoforms that devel-
oped seaward of the northern margin are absent to
The southern platform margin passes rapidly into
low-energy, restricted facies interpreted to have been
deposited in a shallow intrashelf embayment that
developed during the time of sequence 3. The inferred
geometry of the embayment is shown in Figure 4, but
is poorly constrained outside the 3-D seismic data
area. At a regional scale, it is possible that the Shuaiba
platform margin may have evolved into a series of
low-relief, isolated banks.
SEQUENCE RECOGNITION
Criteria for recognition of depositional sequences
and surfaces in the Shuaiba vary depending on po-
sition in the Aptian sequence set and position along
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 185
FIGURE 11. Paired seismic section and line drawing illustrating the major phases of platform evolution during the
deposition of the Aptian sequence set. The maximum flooding surface (MFS) shown occurs in sequence 2 and is the
second-order MFS for the Aptian sequence set.
the depositional profile (platform, margin, slope, and
basin). Sequence-stratigraphic correlations are guid-
ed by a combination of seismic, log, and core data.
The seismic-stratigraphic architecture of the Aptian
sequence set is summarized in Figure 11. Preliminary
sequence interpretations are weighted largely on seis-
mic geometries and termination patterns (e.g., down-
lap, toplap, and onlap). Because seismic surfaces cor-
respond closely to time lines, seismic interpretations
are used as a template to guide higher resolution
correlations derived from core and log observations.
In general, seismic surfaces are well expressed in the
clinoforms of the northern field area where there are
significant amplitude variations and seismic geome-
tries (Figure 11). Seismic surfaces in platform interior
and margin areas are difficult to constrain because of
less geometric variation and minimal impedance con-
trast (Figures 8, 11). Further complexity is introduced
by the high-amplitude bursts in the platform interior
associated with the pond and tidal-channel complex
(Figure 8B). Seismic interpretation in platform inte-
rior and margin areas is guided by well ties and the use
of multiple seismic volumes (e.g., amplitude, band
limited impedance, and discontinuity).
Some of the sequence-stratigraphic surfaces are
well expressed in core (e.g., exposure surfaces, hard-
grounds, stylolitic boundaries, shale seams), as illus-
trated in Figure 12. Most of the surfaces are accom-
panied by rapid facies changes above and below and
by changes in cycle stacking patterns. Transects of
cored wells were selected to sample the full range
of stratigraphic and seismic variability observed
across the platform (Figures 8, 10). Surface correla-
tions through the northern platform margin are dif-
ficult because of the thick stack of relatively contin-
uous, high-porosity rock in sequences 2 and 3.
Sequence Boundaries
Because of the prevailing Cretaceous greenhouse
climate pattern during the Aptian, sequence bound-
aries are generally not pronounced. The best devel-
oped sequence boundaries are those developed at the
base and top of the Aptian sequence set, especially
the upper sequence set boundary (Figures 8, 9). Third-
and fourth-order sequence boundaries are best devel-
oped in the highstand and late highstand phases of
the sequence set because of a decrease in long-term
accommodation (Figures 8, 9).
In updip areas of the Shuaiba (platform position),
sequence boundaries are commonly expressed as thin,
iron-oxide – encrusted surfaces, with pyrite mineral-
ization (Figure 12A, B). Evidence for dissolution is
limited to within a few feet or inches of the exposure
surfaces. Coarse, blocky calcite cements are developed
186 / Yose et al.

FIGURE 12. Core expression of sequence-stratigraphic surfaces. SB = sequence boundary; MFS = maximum flooding
surface; HFS = high-frequency (fourth-order) surface; HST = highstand systems tract. Surfaces in (A, B) are long-term
exposure surfaces, forming the sequence boundary of the Lower Cretaceous supersequence. Surfaces shown in (D, E)
are candidates for composite exposure and flooding surfaces, as discussed in Immenhauser et al. (2000). Characteristics
of the surfaces are discussed in the text.

locally below sequence boundaries and can penetrate
as deep as 3 m (10 ft) below the exposure surface.
Sequence boundaries are also marked by rapid facies
changes above and below and are sometimes overlain
by coarse skeletal and/or lithoclastic lag deposits.
In downdip areas (slope and basin positions), se-
quence boundaries become conformable and lack evi-
dence for exposure. However, some sequence bound-
aries record major facies shifts. For example, in the
clinoform area, sequence boundaries correspond to
a shift from grainy peloidal-skeletal facies (porous
zones) in the HST to burrowed, miliolid-rich wack-
estone (dense zones) in the LST to TST above (Figure 9).
In some cases, the downdip sequence boundaries are
manifested by clay-rich seams that occur in the same
position as the facies shifts (Figure 12C) and generate
a clear gamma-ray log response to the surface.
Maximum Flooding Surfaces
The maximum flooding surface of the Aptian se-
quence set is interpreted at the base of the last platform-
margin backstep in the platform area, coincident with
the maximum flooding surface of sequence 2 (Figures 8,
11). The equivalent surface in the slope and basin
environment is interpreted within the condensed in-
terval that separates shallow-water platform facies be-
low from deeper water clinoform facies above (coinci-
dent with the seismic downlap surface; Figures 8, 11).
Flooding surfaces in the platform area generally oc-
cur at or near a transition in facies and stacking pat-
terns that record a change from deepening-upward to
shoaling-upward characteristics. In some cases, it is
difficult to pick an exact surface and is more appro-
priate to identify a flooding interval. Flooding surfaces
(or intervals) are commonly well expressed in slope
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 187

FIGURE 13. Facies explanation and key to core log symbols used in this paper. Facies are subdivided based on the
position in the Aptian sequence set and the position along the platform to basin profile. Core photos and reservoir-quality
data for reservoir facies are illustrated in Figures 14 – 16. Core-to-log facies calibrations are shown in Figures 17 – 20. MLP =
mud-lean packstone; skel = skeletal; pel = peloidal; bndst. = boundstone; fltst. = floatstone; HST = highstand systems tract;
TST = transgressive systems tract.

and proximal basin environments and are generally Transgressive Phase

expressed as condensed intervals that include concen-
trations of clay or muddy carbonate sediments, bur-
rowed hardgrounds, and/or stylolites (Figure 12D–F).
SEQUENCE DESCRIPTION
A general description of depositional sequences in
relationship to evolution of the Shuaiba platform is
provided below. The discussion is keyed to the four
main phases of the Aptian accommodation cycle in-
troduced above. A key to the facies nomenclature and
symbols used in this report is provided in Figure 13.
Core photos of representative facies types for each
major phase of the Aptian sequence set are displayed
in Figures 14 – 16 (transgressive, early highstand, and
late highstand phases, respectively). Type logs for each
of the main areas of the buildup are provided in
Figures 17 – 20 (platform interior, platform margin,
proximal clinoform sequence, and distal clinoform
sequence, respectively).
The transgressive phase records the deposition dur-
ing the rising limb of the long-term sea level cycle
and includes sequences 1A–1B and 2 (Figure 9). Dur-
ing this period, the rate of relative sea level rise was
greater than the carbonate accumulation rate. As
a consequence, depositional relief was established
over time, marking the initiation of the Bab basin
in the area. The carbonate buildup differentiated
as a low-relief ramp that attained as much as 45 m
(150 ft) of depositional relief during the transgressive
phase. Ramp interior facies are dominated by an algal-
stromotoporoid assemblage, with some associated
corals and rudists (Figure 14). The faunal assemblage
is interpreted to indicate relatively restricted circula-
tion during the initial development of the Bab in-
trashelf basin. Ramp interior and margin facies of
the transgressive phase are characterized by high po-
rosity (10–25%) and low permeability (0.1–5 md) be-
cause of mud-dominated textures and development
of microporosity.
188 / Yose et al.

FIGURE 14. Examples of depositional facies and reservoir properties in the composite transgressive interval (sequences 1
and 2). Colors in the upper right of photos are keyed to Figure 13. Scale bar is in 1-cm (0.4-in.) increments. IP = intraparticle;
BP = between particle; md = millidarcys; Perm = permeability; Por = porosity; HST = highstand systems tract; TST =
transgressive systems tract.

The transgressive phase comprises as many as
three depositional sequences (sequences 1A, 1B, and
2) as summarized below.
Sequence 1A
Sequence 1A includes the Hawar shale (Figure 9)
and was not extensively evaluated in this study. The
Hawar comprises argillaceous, Orbitolina-rich pack-
stone and wackestone and contains cyclic alterna-
tions of more and less argillaceous deposits. As dis-
cussed further in Strohmenger et al. (2006), there are
alternative sequence-stratigraphic interpretations of
the Hawar shale. Strohmenger et al. (2006) inter-
preted that the Hawar was deposited in relatively
shallow, restricted environments with episodic expo-
sure. Strohmenger et al. (2006) also interpreted that
the Hawar is bounded above and below by sequence
boundaries (interpretation shown in Figure 3). In
contrast, van Buchem et al. (2002) interpreted the
Hawar to record a regional transgression associated
with the initial formation of the Bab basin. In this
interpretation, the Hawar may be the TST of se-
quence 1, as opposed to a separate depositional se-
quence. More regional study is required to test these
alternative interpretations.
Reservoir Quality
The Hawar is of very low porosity and permeability
and forms an effective barrier (dense) between the
Kharaib reservoir below and Shuaiba reservoir above.
Sequence 1B
Sequence 1B comprises the basal part of the
Shuaiba Formation and records a major change in
deposition relative to sequence 1A. The argillaceous
content in sequence 1B drops off significantly, and
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 189

FIGURE 15. Examples of depositional facies and reservoir properties in the composite highstand interval (sequence 3).
Colors in the upper right of photos are keyed to Figure 13. Scale bar is in 1-cm (0.4-in.) increments. IP = intraparticle;
BP = between particle; md = millidarcys; Perm = permeability; Por = porosity; TST = transgressive systems tract.

there is a shift from Orbitolina-dominated facies to
algal-dominated facies. Similar to sequence 1A, se-
quence 1B records shallow-water deposition across the
study area, with no indication of depositional relief.
The sequence is divided into three shoaling-upward
cycles that can be correlated across the study area
based on core and log character (Figures 17, 18). Fa-
cies stacking patterns in the high-frequency cycles
are variable, but a general progression, from base to
top, would include (1) skeletal-peloidal wackestone,
(2) algal boundstone-floatstone, and (3) chalky peloidal-
skeletal packstone (with microporosity).
Reservoir Quality
The algal boundstone and floatstone represent a
range of facies that have variable reservoir quality
(Figures 17, 18). Porosity ranges from 10 to 20% and
permeability from 0.1 to 5 md. The primary controls on
reservoir quality variability include (1) sparse versus
dense algal boundstone fabrics (intraparticle porosity in
algal boundstone is better connected if the algal compo-
nents are touching [dense]) and (2) grainy versus muddy
matrix in the algal floatstone (grainier fabrics have higher
porosity and permeability). The chalky, fine-grained
facies (skeletal and peloidal wackestones and packstones)
are characterized by the development of microporos-
ity with relatively high porosity, but low permeability.
Sequence 2
Sequence 2 records the initial differentiation of
the platform margin and inception of the Bab basin
in the study area (Figure 9). The sequence is differ-
entiated into distinct TST and HST, each with dif-
fering reservoir architecture and quality (Figure 9).
190 / Yose et al.

FIGURE 16. Examples of depositional facies and reservoir properties in the composite late highstand interval (sequences 4
and 5). Colors in the upper right of photos are keyed to Figure 13. Scale bar is in 1-cm (0.4-in.) increments. IP = intraparticle;
BP = between particle; md = millidarcys; Perm = permeability; Por = porosity; HST = highstand systems tract; TST =
transgressive systems tract.

The TST is characterized by a backstepping ge-
ometry that records the development of the northern
ramp margin. The initial margin develops via two
small-scale backsteps that, together, created on
the order of 20 m (65 ft) of relief (Figure 9). Algal-
stromotoporoid boundstone is the dominant facies
in the TST (Figure 14A, B). Cycle stacking patterns
change from exposure-capped shoaling-upward
cycles below (sequence 1A), to amalgamated algal-
stromotoporoid facies above (sequence 2, TST). The
basin is characterized by deposition of organic-rich
wackestone and packstone (Figure 14F). These basinal
facies immediately overlie shallow-water carbonate
facies of sequence 1B and record the initial platform
drowning (Figure 12F). The basinal deposits are in-
terpreted as the condensed interval equivalent to the
maximum flooding interval of the Aptian sequence
set. Nannofossil age dates indicate that the con-
densed interval corresponds with early Aptian global
oceanic anoxic event (OAE1a; Figure 3) (Bralower et al.,
1994; Erba et al., 1999; Premoli Silva et al., 1999).
The HST records aggradation of the northern ramp
margin and the inception of the southern ramp mar-
gin. A diverse suite of facies is represented including
shallow-water, algal-prone facies in the ramp interior
and fine-grained mud-lean packstone on the ramp
margin (Figure 14B). Faunal diversity increases slight-
ly in the HST, with an increased abundance of corals
and rudists (Figure 14C). The increase in faunal diver-
sity may be related to more open-circulation patterns
in the Bab intrashelf basin in response to the marine
transgression. Evidence for decreasing accommoda-
tion during the HST phase includes a shift from re-
trogradational to aggradational geometry along the
northern ramp margin and increased evidence for
exposure. An exposure surface is present at the base
Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 191

of a prominent biostromal unit in the HST interval

(Figure 8). The biostromal unit corresponds to the
mid-C coral zone of Russell et al. (2002), ranges be-
tween 3 and 8 m (10 and 25 ft) thick, and comprises a
complex assemblage of corals, stromotoporoid, and
algae, most of which are in growth position (Figure 14C).
The biostromal unit is interpreted to have nucleated
on the underlying exposure surface. Algal bound-
stone and floatstone with increasing amounts of rudist
debris overlie the biostrome interval and continue up
to the sequence boundary at the top of sequence 2.
The upper sequence boundary is not a prominent
exposure surface, but records a major shift from algal-
dominated facies below to rudist-dominated facies
above. The faunal shift is rapid, commonly occurring
in less than 1 ft (0.3 m) of vertical section. The surface
recording the faunal shift is of regional extent (Hughes,
2000; van Buchem et al., 2002; Droste, 2004) and ap-
pears to indicate a major change in circulation pat-
terns across the Bab basin.

Reservoir Quality
Vertical variations in facies and reservoir quality in
sequence 2 are shown in Figures 17 and 18. Facies
textures in the HST are generally grainier that those in
the underlying TST. The HST is dominated by algal-
boundstone and floatstone with a grain-prone ma-
trix, whereas the underling TST is dominated by algal-
boundstone and floatstone with a mud-prone matrix.
As a result, reservoir quality increases upward in se-
quence 2 (Figures 17, 18). In the TST, porosity ranges
between 15 and 25%, and permeability ranges from
1 to 20 md. In the HST, porosity ranges between 20
and 30%, and permeability ranges from 10 to 100 md.
The condensed interval deposited within the basinal
area is dense and forms a barrier between the shallow-
water platform facies of sequence 1B below and
overlying slope facies of sequences 4 and 5 above.

Early Highstand Phase

During the early highstand phase, the rate of long-
term relative sea level rise was in balance with the
carbonate sedimentation, resulting in aggradation of

FIGURE 17. Type log for the platform interior (well A in

Figure 8). Vertical scale is in 10-ft (3-m) depth increments.
Porosity values below 5% are shaded in gray. The 1-md
permeability line is highlighted in red. See Figure 13 for
key to facies codes and symbols. HST = highstand systems
tract; TST = transgressive systems tract; SB = sequence
boundary; MFS = maximum flooding surface.
192 / Yose et al.

the carbonate platform. Maximum depositional re-

lief was attained during the aggradational phase,
with more than 75 m (250 ft) of relief established
between the platform and basin (Figure 9). Facies and
faunal diversity reached a maximum during the early
highstand phase, with the proliferation of rudists
across the bank top (Figure 15A, B, D) and the de-
velopment of margins to the north and south of the
platform. The development of an open, high-energy
platform may have been triggered by changes in circu-
lation patterns in the Bab basin, in combination with
relative sea level rise. Platform interior and margin
facies deposited during the early highstand phase
are the highest reservoir quality facies in the Shuaiba.
Sequence 3 comprises the early highstand phase of
deposition and is subdivided into distinct TST and
HST, as discussed below.

Sequence 3 (TST)
The TST of sequence 3 records significant changes
in platform and margin deposition. The southern
platform margin attained maximum development at
this time, with as much as 30 m (100 ft) of relief be-
tween the platform and shallow embayment to the
south. Depositional patterns in the platform interior
changed significantly in response to changes in plat-
form paleogeography. The platform interior differen-
tiated into a complex mosaic of tidal channels, rudist
shoals, and intershoal depressions or ponds, marking
the establishment of significant cross-bank currents
and open-marine conditions (Figures 9, 10).
Platform-margin facies differ significantly between
the northern and southern margins. The northern
platform margin is characterized by a broad (5–7-km;
3.1 –4.4-mi) grainstone belt that passes northward
into deeper water slope facies. A narrow band (1 –
2 km; 0.6 – 1.2 mi) of in-situ rudist and stromotopor-
oid boundstone and bafflestone is located platform-
ward of the grainstone belt and is interpreted to form
a discontinuous band of stromotoporoid and rudist
patch reefs along the northern platform. The north-
ern margin sand belt is characterized by skeletal-to-
peloidal packstone and grainstone that are inter-
preted to have been deposited in a high- to moderate

FIGURE 18. Type log for the northern platform margin

(well B in Figures 8, 11). Vertical scale is in 10-ft (3-m)
depth increments. Porosity values below 5% are shaded
in gray. The 1-md permeability line is highlighted in red.
See Figure 13 for key to facies codes and symbols. HST =
highstand systems tract; TST = transgressive systems tract;
SB = sequence boundary; MFS = maximum flooding surface.
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 193
FIGURE 19. Type log for the clinoform area (well C in
Figures 8, 11). Well is in updip position of sequence 4,
highlighting the development of the dense TST. Vertical
scale is in 10-ft (3-m) depth increments. Porosity values
below 5% are shaded in gray. The 1-md permeability
line is highlighted in red. See Figure 13 for key to facies
codes and symbols. HST = highstand systems tract;
TST = transgressive systems tract; SB = sequence boundary;
MFS = maximum flooding surface.
energy carbonate sand flat. Skeletal components are
diverse and include local concentrations of Orbitolina-
rich and miliolid-rich grains. Textures are relatively
fine grained and become coarser grained and cleaner
(less muddy) upward within the section. Textures be-
come finer and muddier toward the basin (downslope).
In contrast, the southern platform margin is charac-
terized by high-energy, rudist-dominated facies that
cover a broad region of the outer platform. Deposi-
tional facies along the southern margin include a
stack of rudist-dominated grainstone and boundstone
(Figure 15A). Little vertical variation in this facies exists,
and no internal surfaces or bedding features were iden-
tified. High-energy facies along the southern margin
pass gradationally into the mosaic of rudist shoals,
intershoal depressions, and channels characteristic
of the platform interior to the north and grade rapidly
into Orbitolina-rich packstone and grainstone of the
upper- to middle-slope environment to the south.
Sequence 3 (HST)
The HST records a significant change in platform
and margin deposition relative to the underlying TST
(Figure 9). Depositional energy levels in the platform
interior become much lower. The complex associa-
tion of channels, ponds, and shoals developed in the
TST are absent in the HST. In contrast, the HST re-
cords a major decrease in depositional energy across
the bank top and an attendant decrease in reservoir
quality. Carbonate facies are generally mud domi-
nated, with a notable increase in argillaceous con-
tent. In addition, in contrast to the amalgamated
rudist shoal facies characteristic of the TST, the HST is
characterized by the development of shoaling-up-
ward parasequences with well-developed surfaces.
Platform-margin facies and geometries also change
significantly. The width of the southern margin
narrows to 2 km (1.25 mi) and is characterized by
much lower energy facies than in the underlying
TST (Figure 9). The northern margin likewise narrows
relative to the underlying TST and evolves from a
fine-grained sand-flat environment into coarser rudist
rudstone and floatstone deposited as shoals. Facies
relationships and seismic geometries indicate that
the northern margin prograded approximately 3 km
(1.9 mi) during the HST as the rate of sea level rise
began to decrease (Figure 8).
The controlling factor on restriction of the plat-
form interior during the HST may be related to ge-
ometry changes at the platform margins. In partic-
ular, the northern margin appears to become a more
pronounced topographic feature, possibly in response
194 / Yose et al.
to shoaling and progradation of the margin. Progra-
dation and increased topographic expression of the
northern platform margin may have restricted cir-
culation in the platform interior and in the intrashelf
embayment to the south.
FIGURE 20. Type log for the clinoform area (well D in
Figures 8, 11). Well is in downdip position of sequence 4,
highlighting the development of the porous HST. Vertical
scale is in 10-ft (3-m) depth increments. Porosity values
below 5% are shaded in gray. The 1-md permeability line
is highlighted in red. See Figure 13 for key to facies codes
and symbols. HST = highstand systems tract; TST =
transgressive systems tract; SB = sequence boundary;
MFS = maximum flooding surface.
Reservoir Quality
Reservoir quality varies both vertically and laterally
within sequence 3 (Figures 17–20). Rudist shoal facies
deposited in the platform interior of the TST contain
the highest reservoir quality of any interval in the
Shuaiba. Porosity ranges from 20 to 30%, and perme-
ability ranges from 10 to 500 md. The TST also marks
the period of maximum facies diversity caused by the
development of the pond-channel network in the plat-
form interior. Ponds and channels are filled with mud-
prone facies that have variable, but generally low, reser-
voir quality. Many of these features are nonreservoir.
Reservoir quality varies significantly between the
northern and southern margins (Figure 15). Rudist-
dominated facies of the southern margin have poros-
ity of 20 –30% and permeability in the 100–1000-md
range. In contrast, fine-grained grainstone and mud-
lean packstone deposited along the northern margin
has similar porosity to the southern margin, but per-
meability is much lower, in the 10– 100-md range.
The HST of sequence 3 is of generally low reservoir
quality, with the best reservoir facies corresponding
to rudist shoals developed along the northern mar-
gin. These facies range from 20 to 30% porosity, and
permeability ranges from 10 to 100 md.
Late Highstand Phase
The late highstand phase marks a fundamental
turning point in the history of the Aptian platform.
Relative sea level rise slowed to a near standstill and
then started to gradually fall. The carbonate platform
had limited space for vertical accumulation and be-
gan to move laterally, and downslope, into the Bab
basin. In response, the northern platform margin pro-
graded more than 10 km (6.2 mi) into the Bab basin.
Seismic geometries in the clinoform belt define two
major phases of progradation that correspond to two
depositional sequences (sequences 4 and 5; Figure 11).
Because of a gradual, long-term fall in sea level, each
slope sequence steps lower into the basin. Sequences 4
and 5 are thus detached from the main platform and
are slope restricted. The older Aptian platform was
exposed during most of the late hightstand phase,
and a long-term subaerial exposure surface was de-
veloped. The exposure surface becomes progressively
younger toward the basin in response to the down-
stepping geometry (Figure 9).
Sequences 4 and 5 have very similar architecture,
facies, and reservoir quality. Each sequence includes
a muddy, lowstand-to-transgressive phase (low reser-
voir quality) with low-angle clinoforms, followed by
a grainy, highstand phase (moderate to high reservoir
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 195
quality) with higher angle clinoforms (Figures 10, 11).
Facies relationships in sequences 4 and 5 are sum-
marized in Figure 16.
Sequences 4 and 5 (TST)
The TST is dominated by muddy, nonporous facies
that create extensive barriers to fluid flow. The facies
model for the TST includes a downdip transition
from (1) caprotinid floatstone, bafflestone, and rud-
stone at the margin (Figure 16D) to (2) mixed-peloidal
and skeletal packstone and grainstone in the upper-
most slope (Figure 16E) to (3) burrowed wackestone
(denses) in the upper- and lower-slope environments
(Figure 16F). The TST is dominated by thick, mud-
prone facies deposited in the upper- and middle-
slope environments. As will be discussed below, these
muds are interpreted to have been generated in situ
as mud mounds.
Sequence 4 and 5 (HST)
The HST is dominated by grainy, porous clino-
forms alternating with higher frequency (fourth-
order) dense intervals. Clinoforms of the HST exhibit
higher slope angles (2 – 38) than the muddy TST cli-
noforms (1 – 28). A typical facies progression in the
HST includes rudist rudstone at the shelf margin
(Figure 16A), skeletal grainstone in the upper slope,
peloidal packstone and grainstone in the middle
slope (Figure 16B), and foraminifera-packstone in the
lower-slope environment (Figure 16C).
Reservoir Quality
Reservoir quality varies dramatically between the
TST and HST components of the slope sequences and
along the depositional profile (Figures 16, 19, 20).
The mud-dominated TST of each sequence is dense
and forms an effective and laterally continuous bar-
rier. Grainier facies in the updipmost part of the TST
exhibit porosity that ranges from 10 to 20% and per-
meability ranging from 10 to 50 md, thus providing
for possible updip connectivity between the slope se-
quences. The grain-dominated HST of each sequence
has much higher reservoir quality, although the in-
ternal architecture of the HST clinoforms is complex,
resulting in a high degree of heterogeneity. The fourth-
order cyclicity generates a high-frequency alternation
of flow units and flow barriers. Reservoir quality var-
ies systematically down the profile of each clinoform
in response to pore-type changes. Porosity values
are fairly consistent (15–25%) down the profile, but
permeability varies significantly, ranging from the
tens to hundreds of millidarcys in grain-prone updip
facies, to 0.1–1 md in the downdip mud-prone facies
of the lower slope. The dominant pore type in lower-
slope facies is microporosity. Microporosity develop-
ment may have been enhanced through meteoric dia-
genesis, also in association with subaerial exposure.
Lowstand Phase
The maximum lowstand of sea level is marked by a
significant influx of fine-grained siliciclastics and the
demise of the Aptian carbonate platform. Lowstand
facies form a basin-restricted wedge (sequence 6) that
onlaps the Aptian platform margin (Figures 10, 11)
and forms a lateral seal to the reservoir. The lowstand
wedge comprises a shoaling- and cleaning-upward se-
quence that includes an upward succession from fine-
grained siliciclastics and carbonate mud to skeletal and
oolitic carbonate grainstone. Sequence 6 is bounded at
the top by an exposure surface that precedes deposition
of the overlying Nahr Umr Shale. In the study, only the
updip part of sequence 6 is present. Regional relation-
ships indicate that sequence 6, corresponding in large
part to the Bab Member of the Shuaiba, continues to
prograde into the Bab basin. The Nahr Umr Shale re-
cords a rise in sea level that floods back across the pre-
viously exposed Aptian platform. However, in contrast
to carbonate platform facies below, the marine envi-
ronment was dominated by fine-grained siliciclastics,
with only thin intervals of argillaceous carbonates. The
Nahr Umr forms a top seal for the Shuaiba reservoir.
VOLUME-BASED RESERVOIR EVALUATION
The sequence-stratigraphic framework and 3-D
seismic data are used to guide an integrated evalua-
tion of the Aptian reservoir. Volume interpretation
and visualization workflows are applied to further
evaluate and quantify geologic variations and their
impact on reservoir performance. A corendered im-
age of seismic discontinuity and seismic porosity is
shown in Figure 21 and provides a seismic-scale per-
spective of reservoir architecture and properties.
Field A1 is extremely heterogeneous and can be
subdivided into three main areas, each with diag-
nostic geologic, seismic, and production expressions.
These areas include the platform interior to the south,
the northern platform margin in the central field
area, and the clinoform belt to the north. Geologic
controls that drive these variations are directly related
to the sequence-stratigraphic framework as discussed
above and summarized in Figure 8.
A peripheral waterflood was implemented during
earlier stages of field development (prior to acquisition
196 / Yose et al.

FIGURE 21. Volume-based reservoir evaluation. The volume shown is total porosity corendered with seismic discontinuity.
The volume highlights the reservoir heterogeneity resulting from sequence-stratigraphic variations and the impact on
field development. Time slice is from a flattened volume datumed on the top of sequence 1 (base of reservoir) and
represents a slice through the TST of sequence 3 in the platform interior and margin, cutting through sequences 4, 5,
and 6 in the clinoform belt and the Bab basin. The total porosity volume includes the seismic-derived porosity (band-
limited porosity) plus the low-frequency porosity trend. High-frequency porosity data from core and well data are added
during the 3-D geologic modeling stage.

of 3-D seismic data) and does not adequately accom-
modate the reservoir heterogeneity that is now ap-
parent. Several field management challenges have
developed, including differential sweep in the platform
interior, early water encroachment along the platform-
margin grainstone belt, and pressure support in the
clinoforms (Figure 21). The ADCO has implemented
several field management strategies to further opti-
mize field development. As demonstrated below, the
3-D seismic data will greatly assist in guiding future
well placement and field optimization strategy.
Seismic Characterization of the
Platform Interior
The platform interior is characterized by a com-
plex spatial distribution of porosity that is related to
the channel, pond, and rudist shoal complex devel-
oped during the early highstand phase (Figure 21).
This interval contains between 4 and 5 billion bbl of
in-place oil reserves. Thus, a small improvement in
recovery will have a large economic impact. Under-
standing the origin and distribution of reservoir
quality anomalies in the platform area has pre-
sented a long-standing field development challenge,
both in terms of avoiding these anomalies with pro-
ducer and injector wells and the impact of these fea-
tures on reservoir sweep and flood-front advance
(Figure 21).
Calibration with Well and Core Data
The 3-D seismic data provide a blueprint of the
channel and pond network in the platform interior
(Figures 6, 21) and a new tool for reservoir evaluation.
In our initial reconnaissance of the seismic data,
seismic features were identified within the platform
interior that were interpreted either as karst features
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 197

FIGURE 22. Integrated characterization of the platform interior region. (A) Reservoir isochron corendered with seismic
discontinuity (red = isochron thick; blue = isochron thin). Blue well sticks are water injector wells, and red well sticks
are selected producers. Ponds and channels correspond to isochron thins, and porous rudist shoals correspond to
isochron thicks. Prominent isochron thick to the north is the northern platform margin. Peripheral water injection is
diverted by ponds and channels resulting in differential sweep. (B) Example of high-porosity, high-permeability rudist
shoal facies developed among the ponds and channels. (C) Example of nonporous submarine pond fill (open-marine
skeletal wackestones and packstones with rudist debris). Por = porosity; Perm = permeability; md = milidarcy.

developed in association with the top Shuaiba super-
sequence boundary or as syndepositional features,
such as tidal channels and submarine ponds. As sum-
marized in Figure 22, calibration of the seismic data
with well data supports a syndepositional origin for
the seismic features. Volume corendering of the seis-
mic discontinuity volume with the reservoir isochron
(time thickness) enhanced the detail of stratigraphic
features in the platform interior (Figures 6, 22) and
guided the selection of calibration wells. Calibration
of the seismic data with log data shows that isochron
thins correspond to dense, nonporous carbonate with
fast seismic velocities, whereas isochron thicks corre-
spond to porous carbonates with slower seismic ve-
locities. These observations are consistent with the
seismic porosity predictions shown in Figure 21, con-
firming that the pond and channel complex is gen-
erally of lower reservoir quality than the surrounding,
more continuous seismic facies.
Cores from within the pond features comprise
largely carbonate mudstone and wackestone that
was deposited in open- to restricted-marine settings
198 / Yose et al.
(Figure 22). A range of carbonate facies was observed
within the ponds, including open-marine carbonate
mudstone; wackestone; peloidal-skeletal packstone;
and restricted-marine, organic-rich mudstone and
algal laminites. In contrast, cores from wells located
outside the ponds comprise rudist-dominated grain-
stone, floatstone, and rudstone that were deposited
in high-energy carbonate shoals and patch reefs. The
open-marine facies in the pond fill contain rudists
and other sediments that are interpreted to have
been transported into the ponds from the adjacent
rudist shoals. No evidence was found to indicate that
karsting or dissolution was involved in the formation
of the pond and channel features. Collapse breccias,
fractures, vadose silts, and other diagenetic features
typical of karst environments are absent.
The ponds and channels appear to be time equiv-
alent to the surrounding rudist shoals and patch reefs.
The initial distribution of the rudist shoals and ponds
may have been controlled by antecedent relief on the
underlying sequence boundary (top of sequence 2),
with shoals nucleating on subtle highs. The initial
relief was then enhanced through aggradation of the
shoals in response to sea level rise during the trans-
gressive phase of sequence 3. This increase in accom-
modation resulted in the differentiation of the south-
ern platform margin, thus opening the platform top
up to strong cross-bank currents and development of
submarine channels. As the rate of accommodation
began to decrease, the rudist shoals coalesced across
the bank top, thus infilling most of the pond and
channel topography (Figure 8). The HST of sequence
3 extends across the tidal channels, ponds, and shoals
of the underlying TST, with little change in facies
(Figures 8, 9). This relationship indicates that the
channels and ponds were developed and infilled
during the TST of sequence 3 and are not related to the
long-term exposure event on the top of the Shuaiba
(Lower Cretaceous supersequence boundary).
Reservoir Implications
Seismic porosity distribution in the channel-pond
network shows that, although porosity is generally
lower than the surrounding rudist shoal facies, there
is considerable variability, and not all of the vol-
ume is nonreservoir (Figure 21). Our calibrations
with core and well data indicate that porosity varies
with respect to several factors, including the size of
the ponds and position of wells in the ponds. Larger
ponds are more restricted from marine currents than
smaller ponds and are filled with muddier, tighter rock.
In addition, reservoir quality tends to increase pro-
gressively from the pond interior to the pond margins.
Less data are available to assess facies and reservoir
quality variations in the large tidal channels, but seis-
mic porosity indicates that, in general, the channels
are filled with higher porosity facies than the ponds.
Visualization sessions provide a powerful mecha-
nism for evaluating reservoir sweep and conformance
and for identification of bypassed reserves in the
platform interior. Visualization of seismic volumes
such as those shown in Figures 21 and 22 enabled
reservoir engineers to rapidly resolve long-standing
production anomalies. For example, many producers
with abnormally high or low water cuts can be ex-
plained by their position relative to injector wells and
the pond-channel complex (Figure 22). Producers
that are sheltered from injectors by ponds and chan-
nels show less water breakthrough (higher oil cuts)
than producers that are not as sheltered. The 3-D
seismic data provide new, detailed information on
the distribution of reservoir features that is being used
to guide reservoir development and placement of
future producer and injector wells.
Seismic Characterization of the Platform Margin
The northern platform-margin area is character-
ized by the highest and most continuous porosity in
the field area (Figures 18, 21). Production data indi-
cate that premature water breakthrough is occurring
along the platform margin in response to peripheral
water injection. The seismic data reveal possible strat-
igraphic and structural controls on fluid flow in the
platform-margin area.
Calibration with Well and Core Data
Seismic and sequence-stratigraphic relationships
of the platform-margin area are summarized in
Figure 23. The seismic character of the margin area
is generally transparent because of the stack of porous
and relatively continuous reservoir facies. Discontinu-
ous basinward-dipping seismic reflectors (clinoforms)
are present in the HST of sequence 3. Calibration with
core data show that platform-margin facies is comprised
of mainly skeletal and nonskeletal grainstone that is
remarkably fine grained and well sorted (Figures 15E,
F; 16B, C; 23C). The carbonate sand shoals aggraded
vertically and then, as the rate of sea level rise began
to decrease, prograded basinward. As a result, facies
transition upward from a grainy platform margin into
rudist-dominated outer platform facies (Figure 23C).
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 199

FIGURE 23. Seismic and depositional relationships in northern platform margin area of the reservoir. Depositional sequence
shown in (C) is sequence 3 in Figure 8. The vertical stacking of facies in the platform margin records progradation, with
open-platform interior facies prograding over platform-margin facies. Reservoir quality in the margin area is generally
high and relatively continuous. (A) Seismic transect through northern platform-margin area (amplitude data). (B) Seismic
transect through northern platform-margin area (impedance data; orange = lower porosity, gray = higher porosity).
(C) Facies and reservoir-quality relationships in the platform-interior, platform-margin to upper-slope environments.
BP = between-particle porosity; IP = intraparticle porosity; micro = microporosity; Por = porosity; Perm = permeability;
HST = highstand systems tract; TST = transgressive systems tract; SB = sequence boundary. See Figure 13 for explanation
of facies codes.
200 / Yose et al.

FIGURE 24. Three-dimensional perspective on clinoform architecture, showing a time slice of seismic discontinuity
with 2-D seismic amplitude cross sections. Systematic variations in clinoform architecture are clearly imaged and are
related to the third-order sequence framework as shown. Colored well sticks show the location of wells that were
evaluated as part of the current study. Intersections of third-order sequence boundaries with the seismic time slice
are denoted by the dotted red lines.

Reservoir Implications High-porosity grainstone in the platform-margin tran-

The data indicate that the northern platform mar-
gin is a broad zone of stacked, moderate- to high-
reservoir-quality facies. Seismic data show that this
trend is continuous across the field area and well con-
nected to the peripheral water injectors (Figure 21),
consistent with the observed water breakthrough pat-
terns. However, as previously discussed, faults seen
in the seismic data trend parallel to the platform mar-
gin and may influence fluid flow and water break-
through patterns (Figure 7). Sensitivity testing can be
conducted via flow simulation to determine if ma-
trix properties can match the performance data or
whether additional permeability associated with faults
and fractures is required. The 3-D fault framework
derived from seismic data (Figure 7) can be leveraged
in the sensitivity testing.
The seismic and sequence-stratigraphic frame-
works also provide an improved understanding of lat-
eral facies relationships and reservoir connectivity in
a dip direction across the platform margin (Figure 23).
sitions into the more variable reservoir quality facies
of the platform interior. The platform margin and in-
terior areas are thus in pressure communication, and
injected water moving along the strike of the platform
margin can also interfinger laterally into higher per-
meability facies in the platform interior. Water break-
through has been observed in the proximal platform
interior area and may increase over time.
Seismic Characterization of the Clinoforms
Slope clinoforms are characterized by complex
stratal geometries and significant variations in res-
ervoir properties (sequences 4 and 5; Figure 8). As a
consequence, peripheral water injection has not pro-
vided adequate pressure support or sweep to this part
of the reservoir (Figure 21). In response, ADCO has
implemented a pattern gas-injection flood to add pres-
sure support and optimize recovery. As illustrated in
Figure 24, seismic data provide new 3-D information
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 201

FIGURE 25. Seismic and depositional anatomy of clinoform sequences. (A) Seismic amplitude section across the entire
clinoform belt, highlighting seismic expression of sequences 4 and 5 (TST = transgressive systems tract; HST = highstand
systems tract). (B) Close-up of sequence 4 with seismic interpretation added. Well sticks in black are featured in (C).
(SB = third-order sequence boundaries; MFS = third-order maximum flooding surface; fourth-order surfaces shown with
blue double arrows). (C) Facies schematic for sequence 4 showing impact of clinoforms on flow layering and well pair
communication.

on stratal architecture and reservoir quality variations by finer scale alternations of porous and dense facies
that can be used to assess reservoir connectivity and that result from higher order depositional cyclicity
optimize gas flood management. (Figure 25).

Calibration with Well and Core Data Reservoir Implications

The detailed anatomy of a slope-restricted deposi-
tional sequence is provided in Figure 25. The seismic
character of each sequence includes a low-angle (1 –
28), high-amplitude reflector, followed by a series of
shorter, steeper (3– 48) reflectors. Cores were evalu-
ated across several slope sequence transects to cali-
brate the seismic response (see Figure 24 for location
of calibration wells). The low-angle, high-amplitude
reflector at the base of the cycle is generated by the
impedance contrast between more porous rock below
and extremely dense rock above. In the sequence mod-
el, this interface is interpreted as a sequence boundary
separating more porous highstand deposits below,
from muddier, transgressive deposits above. Dense,
muddy, lowstand-to-transgressive deposits overlie the
sequence boundary and form the base of each depo-
sitional sequence. Higher angle reflectors in the high-
stand part of each depositional sequence are generated
Reservoir connectivity and quality variations in
the clinoforms are summarized in Figure 25. Reser-
voir implications are discussed below in the context
of the sequence hierarchy. At the sequence scale,
transgressive dense intervals at the base of each de-
positional sequence form effective barriers to fluid
flow and partition the clinoform area into separate
pressure compartments (Figure 8). These dense inter-
vals are well imaged by the seismic data, and 3-D ge-
ometries of the reservoir barriers can be mapped across
the northern field area (Figures 25, 26). Highstand
parts of sequences comprise larger scale flow com-
partments and contain the bulk of the pore volume
in the clinoform region.
The highstand flow compartments are, in turn,
subdivided into a series of alternating flow units and
discontinuous flow barriers, corresponding to the de-
velopment of fourth-order cyclicity. As illustrated
202 / Yose et al.

FIGURE 26. Three-dimensional visualization of clinoforms highlighting reservoir connectivity and continuity. Clinoform
geometries have a large impact on connectivity between injectors (yellow) and producers (red). Volume shown is
instantaneous phase with opacity filter applied to enhance visualization of seismic-scale flow units and flow barriers.
Basal structural surface is draped with seismic porosity and discontinuity from within the reservoir.

in Figures 25 and 26, seismic reflection data are able to
resolve many of these fourth-order clinoform pack-
ages and provide a template to guide interwell cor-
relations. Without seismic data, one may be tempted
to correlate on the basis of facies and log similarities,
resulting in horizontal, instead of inclined, correlation
surfaces (i.e., connecting the colors in Figure 25C).
Such correlations crosscut seismic reflectors and geo-
logic time lines and result in incorrect flow layering
and rock property distributions in the reservoir model.
Integration of the seismic with core and well data
provides information on subseismic variations in
reservoir properties. Depositional facies change sys-
tematically down the profile of each clinoform be-
cause of changes in water depth and wave energy.
Grain-rich, porous facies deposited in updip posi-
tions pass gradually into finer grained, less-porous
facies deposited in downdip positions. These facies
changes, coupled with diagenetic effects, result in
systematic pore-type changes that control porosity
and permeability variations in individual flow units
(Figure 25C).
The geometry and continuity of the flow units,
barriers, and baffles will have a pronounced influ-
ence on injector-producer well connectivity in the
pattern gas flood area. As illustrated in Figure 26, 3-D
seismic visualization is used to evaluate the distribu-
tion of seismic-scale flow units and flow barriers in
the clinoforms in relation to existing producer and
injector wells. Such tools are used to assess 3-D res-
ervoir connectivity, to evaluate producer-injector con-
formance, to resolve specific well-pair performance
issues, and to guide future well placement.
IMPLICATIONS FOR RESERVOIR MODELING
Three-dimensional geologic and flow-simulation
models provide the ultimate reservoir management
tools. The inherent complexity of carbonate reser-
voirs presents significant challenges to distributing
3-D rock properties in reservoir models. The new
volume-based reservoir framework provides a tem-
plate for distributing flow units and flow barriers in
the reservoir and for distributing facies and rock prop-
erties in the 3-D framework. Likewise, the 3-D seismic
data constrain the structural mapping of the reservoir
and have provided a 3-D fault framework that can be
incorporated into the reservoir model framework.
An integrated static and dynamic reservoir mod-
eling effort that incorporates results from the pres-
ent study is underway, but was not concluded at the
time of this publication. Some examples are provid-
ed below to show how a volume-based workflow can
lead to more accurate reservoir models and can also
reduce the reservoir-modeling cycle time.
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 203

FIGURE 27. Seed-detected clinoforms extracted from 3-D seismic volume. Colors correspond to connected clinoform
segments. Volume shown is the cosine of instantaneous phase, which enhances continuity of the seismic reflections.

Rapid Interpretation Techniques Volume-based Data Integration

Development of the reservoir framework is one of
the most time-consuming steps in the construction
of a 3-D geologic model. The volume-interpretation
techniques applied in the present study reduced the
time required to construct the structural and strati-
graphic frameworks. For example, seismic mapping
of individual clinoform surfaces in the northern field
area is critical for the reservoir description, but is time
consuming because of the stratigraphic complexity.
Quantitative seed detection techniques were used to
expedite the correlation process. The ExxonMobil
proprietary seed detection approach applied in this
study uses a combination of amplitude and trace-
shape attributes to autopick the clinoform surfaces.
Results are illustrated in Figure 27. The colored sur-
faces are autopicked clinoform surfaces, and each
color corresponds to a connected clinoform surface.
This technique is rapid (interpretations were gener-
ated in a few hours) and allows for quantitative as-
sessment of clinoform connectivity. The fast-track
interpretations must be further evaluated and opti-
mized by the seismic interpreter. Ultimately, the seis-
mic clinoform surfaces are depth converted and used
to guide the correlation of higher order surfaces iden-
tified from logs and core data.
Most conventional 3-D modeling workflows focus
on the integration of one-dimensional and two-
dimensional (2-D) data sets, and a full 3-D represen-
tation of the reservoir is not developed until the later
stages of the workflow, following the construction of
the 3-D geologic model. A volume-based workflow
provides the basis for 3-D reservoir characterization
throughout the workflow. Figure 28 illustrates how
the volume-based environment is used to visualize
information on structure, stratal architecture, reser-
voir properties, and static and dynamic well data in a
common 3-D environment. Such techniques are used
to evaluate relationships between seismic and other
geologic and production data and to identify key
reservoir issues early in the workflow. Volume-based
data integration leads to more accurate reservoir mod-
els and promotes a 3-D reservoir perspective through-
out the workflow.
Extraction of Reservoir Features from Seismic Data
As illustrated in Figure 29, seismic features in the
platform interior can be extracted from the seismic
volume and incorporated directly into the 3-D geo-
logic model. In this workflow, geologic and seismic
attributes are used to guide the extraction process.
204 / Yose et al.

FIGURE 28. Seismic image highlighting variations in reservoir architecture and quality at the full-field scale. Amplitude
data are shown in 2-D vertical slice. Discontinuity draped on time structure and corendered with band-limited seismic
porosity predictions shown in time slice view. Selected calibration wells are shown as yellow sticks. Major paleographic
elements are annotated.

The extraction was constrained vertically using seis-
mic horizons tied to the sequence-stratigraphic frame-
work. Integration of the seismic with well and core
data indicate that the pond-channel network is con-
fined to a specific sequence-stratigraphic interval
(sequence 3, HST; Figure 9), and this knowledge is
used to guide the extraction of these features and
their placement into the 3-D geologic model. The
lateral extent of individual ponds and channels is
constrained using a combination of seismic attri-
butes, including discontinuity, isochron, and poros-
ity. Once constrained vertically and laterally, the fea-
tures can be extracted from any seismic volume,
along with the seismic properties. Figure 29B and C
show 2-D and 3-D perspectives of the pond and
channel features extracted from the seismic porosity
volume. The extracted features, populated with seis-
mic porosity estimates, were converted from time to
depth and incorporated directly into the 3-D geo-
logic model. Seismic porosity predictions are used in
combination with well log and core information to
produce the final 3-D porosity distribution in the
geologic model.
Rock Property Modeling
Accurate representation of rock properties in res-
ervoir models is challenging because of the different
scales of information on rock properties (seismic, well
tests, logs, and core) and the impact of differing pore
types on the porosity-permeability functions. Rock
property modeling requires a combined approach,
using 3-D seismic data and the sequence-stratigraphic
framework to guide large-scale trends between wells,
and high-resolution information from log and core
data to characterize pore-scale relationships.
Integration of Seismic Porosity Data
The average well spacing in Field A1 is 1 km (0.6 mi),
and seismic data provides the only 3-D information on
rock properties between wells. Several seismic attributes
Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 205

were calibrated to log-derived porosity data using a

neural network approach (Hampson et al., 2001) and
used to generate a full-field prediction. A more de-
tailed explanation of the seismic porosity prediction
workflow conducted in association with this study
is provided by Al-Menhali et al. (2005) and Schuelke
et al. (2005). Most of the seismic porosity informa-
tion is contained in the impedance attribute, but
adding additional attributes via the neural net ap-
proach improved the calibration. Seismic porosity
predictions are validated through blind tests, where
wells are held out of the calibration process and then
compared against the predictions to characterize
uncertainty. Visualization techniques are then used
to evaluate the seismic porosity predictions relative
to the sequence-stratigraphic framework and perfor-
mance data. Statistical error in the seismic prediction
is generally low but increases in areas where rock
properties change rapidly over short vertical or lat-
eral distances (e.g., clinoforms of sequences 4 and 5,
and tidal channels and ponds in sequence 3). As il-
lustrated in Figures 21 and 28, there is a strong tie
between seismic-based porosity predictions and the
geologic framework. The seismic porosity volume
provides a valuable reservoir evaluation tool and will
help to constrain porosity trends in the 3-D geologic
model.

Porosity and Permeability Predictions

Prediction of porosity-permeability relationships
requires understanding pore-type distributions,
which, in turn, define reservoir rock types. Reservoir
rock types reflect a combination of diagenetic and
depositional processes, and depending on the level of
diagenesis, rock types may not equate to depositional
facies. In such cases, the diagenetic component must
be modeled separately. Whereas a considerable dia-
genetic overprint does exist in the Shuaiba reservoir,
most of the diagenesis is facies selective and related
to sequence-stratigraphic surfaces (e.g., exposure

FIGURE 29. Seismic extraction of platform interior pond

and channel features for input into the 3-D geologic
model. (A) Boundaries of detected ponds and channels
shown on discontinuity volume corendered with reservoir
isochron (blue = thin). The combination of these two
seismic attributes provides the best lateral constraint on
these features. (B) Time slice view of pond and channel
features extracted from the seismic porosity volume
(blue = low porosity; red = high porosity). (C) Three-
dimensional perspective of pond and channel complex
extracted from seismic porosity volume (blue = low
porosity; red = high porosity).
206 / Yose et al.

FIGURE 30. Porosity-permeability crossplot for sequences 4 and 5. Depositional facies provide a close proxy for reservoir
rock types in this system. Porosity-permeability relationships vary as a function of sequence, systems tract, and depositional
environment and are thus predictable with the sequence framework. Dominant pore types are indicated: BP = between-
particle porosity; IP = intraparticle porosity; MO = moldic porosity; VUG = vuggy porosity.

surfaces). Thus, depositional facies, in combination
with the sequence-stratigraphic framework, provide
a good understanding of reservoir rock-type distri-
bution. Figure 30 illustrates the value of using res-
ervoir rock types and the sequence-stratigraphic
framework to guide porosity-permeability model-
ing in the clinoform area. Reservoir rock types and
their properties vary predictably based on systems
tract (TST and HST) and position along the deposi-
tional profile (updip and downdip) (Figure 30).
Preliminary 3-D modeling results from the clino-
form area are shown in Figure 31. The flow lines in
the model correspond to sequence-stratigraphic sur-
faces and represent geologic time lines. Facies changes
that occur along and across these time lines have
a significant impact on rock property distribution.
Note the strong conformance between the sequence-
stratigraphic framework of the clinoforms (Figure 8),
the seismic porosity response (Figures 21, 28), and
the 3-D geologic model (Figure 31). A key uncertainty
in the clinoform area is how connected the clino-
forms are in the updip area, below the top Shuaiba
supersequence boundary (Figures 8, 31). A clear com-
partmentalization exists between the TST and HST
clinoform intervals downdip, but updip, relationships
are not clear. Only a few wells penetrate the narrow
updip extension of the TST intervals, so there are
limited data. Reconciliation of this issue will require
feedback between the geologic and flow-simulation
models through integration of production and per-
formance data.
LEARNINGS ON CARBONATE SYSTEMS
The volume-based approach used to characterize
the Shuaiba reservoir provides details on carbonate
system development that are sparsely observed in
outcrop or subsurface studies. Some highlights and
key learnings are discussed below.
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 207

FIGURE 31. Paired porosity and permeability extracts from the 3-D geologic model in the clinoform area along the
northern platform margin. HST = highstand systems tract; TST = transgressive systems tract.

Controls on Platform Architecture 4 and 5. The southern margin differentiated mainly

and Evolution
Integrated characterization of the Shuaiba reservoir
reveals a very ordered system, where changes in stratal
geometries and facies distribution are tied closely to
the long-term accommodation history (Figure 9).
However, despite the order and predictability im-
parted by changes in relative sea level, there is vari-
ability in platform development that may be related
to other competing factors, such as local tectonics,
sediment supply, and cross-bank energy flux.
Platform Asymmetry
The Shuaiba platform records a marked asymme-
try in depositional facies and stratal geometries from
north to south across the study area. Despite a similar
accommodation history, the northern and southern
margins are very different in terms of geometry and
facies (Figure 9). The northern margin is character-
ized by the development of a broad (10-km; 6.25-mi),
grain-dominated sand belt that persisted during
most of the time of sequences 2 and 3, followed by
strong progradation during deposition of sequences
during the time of sequence 3 and is characterized by
high-energy rudist shoals that occupy a broad area of
the margin and outer platform (7 –9 km; 4.4 – 5.6 mi)
(Figure 9). The southern margin is mainly aggrada-
tional, and sequences 4 and 5 appear to be repre-
sented as restricted, low-energy mudstone deposited
in the intrashelf embayment (Figure 9).
Carbonate platform asymmetry is common in many
ancient and modern carbonate systems and may de-
velop in response to a variety of controls, including
differential subsidence, sediment supply, physical
energy flux (e.g., currents, waves), and interference
patterns between isolated platforms (Yose and Col-
lins, 2002). Two primary controls on the asymmetry
observed within the Shuaiba buildup are interpreted.
1) Physical energy flux: At the time of maximum
extent of the Shuaiba platform (sequence 3), we
interpret the southern margin as the current-
facing margin and the northern margin as the
leeward-facing margin. The current-facing mar-
gin was the optimum site for proliferation of the
rudists. Facies comprising rudist rudstone are
208 / Yose et al.
common along the southern margin, indicating
strong wave and current activity. Large tidal chan-
nels imaged by the seismic data all appear to enter
the platform from the south and then meander
their way into lower energy areas of the platform
interior, connecting to many of the submarine
pond features (Figures 6, 9). In contrast, the lee-
ward margin is characterized by the deposition of
fine-grained skeletal-peloidal grainstone and
packstone that grade into mud-dominated slope
facies (Figure 8). The sands and muds along the
northern margin are interpreted to record offbank
transport of fine-grained carbonate sediment along
the leeward margin. The tidal-channel complex
facilitated cross-bank transport of sediments to-
ward the leeward margin. The interplay of pre-
vailing winds, currents, and paleogeography in
controlling the observed patterns in cross-bank
energy flux is not clear at present.
2) Position of sea level: The asymmetry developed
during deposition of sequences 4 and 5 results
from strong progradation along the leeward mar-
gin and virtually no progradation to the south
(Figures 9, 10). These relationships are consistent
with the cross-bank energy flux described above.
Another factor, however, is that the long-term
fall in sea level may have significantly restricted
circulation patterns in the intrashelf embayment
to the south (Figure 4). Connections from the
larger Bab basin into the embayment may have
been restricted as sea level lowered. In contrast, the
northern margin faced into the Bab basin, where
open circulation was maintained, and rudist fring-
ing shoals developed in downslope positions.
Variations in Clinoform Geometry
Variations in the geometry and sedimentology of
the slope clinoform sequences can be observed at a
variety of scales and illustrate a range of possible
controls. At the largest scale, variations in the lateral
width of the clinoform belt show a close correspon-
dence to the present-day structure. The clinoform
belt is widest (more progradation) along the crest of
the structure and narrows offstructure (east and
west) (Figure 24). The offstructure thinning of the
clinoform belt is accommodated by a combination
of increased slope angles to the east and west and a
pronounced narrowing of sequence 5 to the east,
compensated primarily through thinning of the TST
(Figures 9, 24). On the eastern side of the field, it be-
comes difficult to differentiate sequences 4 and 5 on
the basis of seismic time slices because the shingled
HSTs of sequences 4 and 5 are merged (Figure 24).
These relationships demonstrate the variability
that can occur along the strike of carbonate platform
margins. The changes in clinoform geometry could
result from along-strike variations in sediment sup-
ply or subsidence patterns. The offstructure decrease
in the width of the clinoform belt indicates the pos-
sibility of an early structural influence on sedimen-
tation. The primary structural movement of the large
north-northeast structures on the Arabian plate is in-
terpreted to have been in the Late Cretaceous ( Johnson
et al., 2005). However, even a slight increase in sub-
sidence rates offstructure would generate increased
accommodation space and could account for the in-
crease in slope angles and the decrease in the distance
of progradation observed in the offstructure posi-
tions. Conversely, the clinoform geometries could
have been produced by variations in sediment supply,
with areas of higher progradation corresponding
to higher sediment supply. The apparent pinch and
swell of individual clinoforms observed in seismic
data are interpreted to record lobate geometries re-
sulting from variations in erosion and sediment sup-
ply. These variations could also account for geometric
variations at the larger scale.
Platform Response to Sea Level Fall
The long-term fall in sea level during the Aptian
provides an opportunity to evaluate the response of a
carbonate system to sea level fall, including perspec-
tives on stratal architecture, sedimentation patterns,
and diagenesis.
Downslope Shifts of the Carbonate Factory
The Aptian platform demonstrates the potential
for carbonate systems to shift downslope in response
to sea level fall. Sequences 4 and 5 are interpreted to
have accumulated as slope-restricted wedges that are
detached from the main platform (Figure 8). Seismic
(time) and log (depth) relationships confirm that
the clinoform sequences are downstepping into the
Bab basin, tracking the second-order sea level fall. In
the literature, these relationships are referred to as
‘‘forced regressions,’’ in that sedimentation is being
forced basinward (downslope) in response to the sea
level fall (Hunt and Tucker, 1992). Younger sequences
onlap the margin of older sequences, and the com-
posite unconformity becomes progressively younger
toward the basin. Facies relationships from core in-
dicate little to no erosion along the tops of the cli-
noforms; a normal facies progression is observed up to
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 209
the sequence boundary in each clinoform sequence.
Thus, the basinward thinning of the clinoform pack-
age is not caused by erosion. The low-angle (2–38)
slopes flanking the buildup provided a large area for
in-situ carbonate production, and the relative de-
crease in accommodation during the long-term sea
level fall allowed for extensive progradation of the
carbonate system into the shallow basin.
Forced regressions, as observed within the Aptian
sequence set, raise questions on the placement of
sequence boundaries and on systems tract nomen-
clature in the slope sequences. In the Shuaiba ex-
ample, the Aptian sequence set boundary could be
placed below sequences 4 and 5, marking the ini-
tiation of long-term exposure of the central Shuaiba
platform. In this interpretation, sequences 4 and 5
would be considered as part of the composite low-
stand interval. Our preferred interpretation is to place
the sequence set boundary at the base of sequence 6.
In this interpretation, sequences 4 and 5 become part
of the late highstand interval or, alternatively, the
falling-stage systems tract of Plint and Nummedal
(2000). Sequence 6 is viewed as the lowstand phase
of the next sequence set and records a major change
in sedimentation patterns, including a large influx
of siliciclastics. Issues around nomenclature have no
impact on the fundamental sequence interpretations
or the reservoir framework. However, for regional cor-
relations, it is important to have a clear and consis-
tent nomenclature.
Mechanisms of Progradation
The Shuaiba platform provides new insights on
mechanisms of carbonate progradation in response
to sea level changes. Each clinoform slope sequence
(sequences 4 and 5) includes a low-angle, mud-
dominated base (TST), followed by deposition of
higher angle grainy clinoforms (HST) (Figures 9, 10).
Each sequence records on the order of 5 km (3.1 mi) of
progradation, resulting in more than 10 km (6.2 mi)
of total progradation (Figures 24, 25). In the central
part of the clinoform belt, partitioning of prograda-
tion between the TST and HST intervals is roughly
equal, but the mechanisms of progradation are dif-
ferent. A two-phase style of progradation is proposed.
The TST phase includes aggradation of carbonate
mud banks that nucleate downslope in response to
sea level fall and then aggrade vertically during the
subsequent transgression. Aggradation of the mud
banks creates new depositional relief and effectively
steps the margin 2 – 3 km (1.2 – 1.9 mi) outboard rela-
tive to the previous highstand. During the HST
phase, the rate of accommodation begins to slow,
and higher energy shoals initiate along the newly
developed platform margin and then prograde basin-
ward (Figure 25A, B).
The source of the mud in the transgressive interval
is enigmatic because the adjacent bank top is in-
terpreted to be exposed during the deposition of the
slope sequences. The carbonate factory was thus re-
stricted to the slope. The transgressive muds thicken
upslope, forming a basin-thinning wedge. Based on
evaluation of core, we interpret that the upper- to
middle-slope environment was an area of in-situ car-
bonate sediment production. As the carbonate sys-
tem was catching up with rising sea level after the sea
level fall, the main type of sediment produced was
carbonate mud. The source of the muds may have
been sea grass and algal meadows developed in the
upper- to middle-slope environments. The dramatic
upslope thickening of denses supports the concept that
the carbonate muds were produced locally (Figure 9).
The presence and abundance of miliolid foraminif-
era, which lived attached to sea grasses, further sup-
ports the interpretation that sea grass may be an
important factor in both the baffling and production
of carbonate mud. Hydrodynamic ponding of muds
along the upper-slope environment may also be a
factor.
Impact of Long-term Subaerial Exposure
The supersequence boundary that developed on
the top of the Aptian sequence set resulted in long-
term exposure of the carbonate platform, providing
an opportunity to assess the impact of long-term ex-
posure on the carbonate system. The exposure sur-
face is time transgressive and becomes progressively
younger toward the basin because of the progressive
downslope shift of the carbonate system (Figure 9).
The time span of the exposure surface is estimated
to range from 3 – 4 m.y. in the platform interior to
1 – 2 m.y. in sequence 5.
Based on observations from core and seismic data,
there is no evidence for karsting of the platform in
response to the long-term sea level fall. Exposure sur-
faces are marked by thin (2–6-cm; 0.8–2.4-in.) crusts
of iron mineralization, with only minor evidence for
dissolution (Figure 12A, B). Although evidence for
karsting is absent, long-term exposure of the carbon-
ate platform is interpreted to have had a significant
impact on porosity development near sequence bound-
aries and in some sequences. Blocky calcite cements
are common below surfaces and can extend down for
several feet, creating tight intervals immediately below
210 / Yose et al.
sequence boundaries. Subaerial exposure of grain-
prone facies results in the development of abundant
vuggy and moldic porosity through dissolution of
skeletal and nonskeletal grains. Selective dissolution
is observed to penetrate tens of feet below sequence
boundaries into the underlying sequence. Dissolu-
tion enhances porosity in facies that have abundant
primary interparticle porosity, leading to high poros-
ity and permeability (Figure 16A). Abundant micro-
porosity is observed in algal platform facies of se-
quence 1B (Figure 14A, B, E) and in the middle- to
lower-slope facies of sequences 3–5 (Figures 15F, 16C).
Development of microporosity may have been en-
hanced by meteoric fluids. Mircroporosity results in
relatively high porosity but low permeability.
CONCLUSIONS
High-quality 3-D seismic data acquired over an
onshore field in Abu Dhabi demonstrate the value of
seismic data for integrated reservoir characterization
and field optimization. The seismic data quality, abun-
dance of core, well and production data, and the
reservoir heterogeneity all combine to produce an
ideal data set to test the limits of high-end seismic
technologies in carbonates and to demonstrate the
impact on reservoir characterization. Seismic data, in
combination with sequence-stratigraphic concepts,
provide a valuable framework for reservoir evalua-
tion and for incorporating flow layering and rock
property variations into reservoir models.
Key elements of and insights gained from the
workflows applied in the present study are summa-
rized below.
1) Optimizing the seismic data quality: The impact
of high-end seismic technologies increases as
the data quality improves. Efforts on seismic
data optimization, including poststack filtering,
should occur early in the seismic workflow. The
present study demonstrates that poststack filter-
ing can improve the signal-to-noise ratio and
the interpretability of 3-D seismic surveys, even
those with high-effort seismic data acquisition
and prestack processing.
2) Calibration of seismic data to core, well, and pro-
duction data: Seismic data must be calibrated with
other subsurface data to groundtruth the seismic
response to variations in geology and reservoir
properties. A fundamental understanding of the
underlying geology is critical for carbonate res-
ervoir characterization. Volume-based visualiza-
tion is a useful tool to facilitate integration and
evaluation of a wide range of subsurface data and
their relationships.
3) Use of multiple seismic attributes for structural
and stratigraphic interpretations: A range of seis-
mic attribute volumes should be evaluated to
facilitate interpretation of the seismic data. Dif-
ferent attribute volumes provide different infor-
mation and perspectives on the structural and
stratigraphic frameworks and distribution of rock
properties. Attribute volumes that were most use-
ful in this study included amplitude, discontinu-
ity, impedance, isochron, porosity, dip, instanta-
neous phase, and quadrature. Attribute volumes
can also be corendered in different combinations
to provide additional detail and information.
4) Evaluation of seismic data within an integrated
geologic framework: Application of high-end seis-
mic technologies must be conducted within in-
tegrated structural and stratigraphic frameworks.
This is a continuous process because seismic data
is used to develop geologic frameworks, and the
frameworks, in turn, are used to guide more de-
tailed seismic attribute analyses, such as porosity
prediction and fault and fracture analysis.
5) Multidisciplinary approach: Accurate reservoir
characterization and evaluation requires a multi-
disciplinary approach. Three-dimensional visuali-
zation of seismic and other subsurface data fa-
cilitates interactions among geoscientists and
engineers, leading to a collective understanding of
the reservoir and improved reservoir management.
6) Volume-based reservoir optimization: As dem-
onstrated in the present study, calibrated 3-D
seismic data can provide a powerful foundation
for reservoir evaluation and optimization. Visual-
ization of seismic data, along with well and pro-
duction data, can help to resolve longstanding
reservoir performance issues, identify new op-
portunities, and guide future field development
strategies.
ACKNOWLEDGMENTS
The authors thank the Abu Dhabi National Oil
Company, the Abu Dhabi Company for Onshore Oil
Operations, the ExxonMobil Exploration Company,
and the ExxonMobil Upstream Research Company
for supporting this collaborative study and for per-
mission to publish these results. We also thank our
colleagues for their contributions and valuable
 Three-dimensional Characterization of a Heterogeneous Carbonate Reservoir, Abu Dhabi / 211
discussions, including Jim Anderson, Jim Markello,
Peter Holterhoff, Imelda Johnson, Po Tai, Nat Stevens,
Steve Bachtel, Brian Coffey, and Linda Corwin. The
manuscript was improved by reviews from Jim
Weber, Scott Tinker, Tony Simo, Sherry Becker, and
Mike Kozar.
REFERENCES CITED
Al-Menhali, S. S., W. L. S. Abu, and J. S. Schuelke, 2005,
Rock property prediction using multiple seismic and
geologic attributes provides insight to field develop-
ment for a large U.A.E. field: International Petroleum
Technology Conference (Qatar) Paper 10595, CD-ROM.
Alsharhan, A. S., 1985, Depositional environment, res-
ervoir units evolution, and hydrocarbon habitat of
Shuaiba Formation, Lower Cretaceous, Abu Dhabi,
United Arab Emirates: AAPG Bulletin, v. 69, p. 899 –
912.
Alsharhan, A. S., 1993, Bu Hasa field — United Arab
Emirates: Rub’ al Khali basin, Abu Dhabi, in N. H.
Foster and E. A. Beaumont, compilers, Structural traps
VIII: AAPG Treatise of Petroleum Geology, Atlas of Oil
and Gas Fields, v. A-26, p. 99 – 127.
Alsharhan, A. S., and A. E. M., Nairn, 1993, Carbonate
platform models of Arabian Cretaceous reservoirs, in
J. A. Simo, R. W. Scott, and J. Masse, eds., Cretaceous
carbonate platforms: AAPG Memoir 56, p. 173 – 184.
Azer, S. R., and C. Toland, 1993, Sea level changes in the
Aptian and Barremian (upper Thamama) of offshore
Abu Dhabi, U.A.E.: Society of Petroleum Engineers
Middle East Oil Technical Conference and Exhibition,
Bahrain, SPE Paper 25610, p. 141 – 154.
Bralower, T. J., M. A. Arthur, R. M. Leckie, W. V. Sliter, D. J.
Allard, and S. O. Schlanger, 1994, Timing and paleo-
ceanography of oceanic dysoxia/anoxia in the late
Barremian to early Aptian: Palaios, v. 9, p. 335 – 369.
Droste, H. J., 2004, Regional controls on reservoir
properties in the Shuaiba Formation of North Oman
(abs.): 6th Middle East Geoscience Conference and
Exhibition, Bahrain, CD-ROM.
Eberli, G. P., G. T. Baechle, F. S. Anselmetti, M. L. Incze,
2003, Factors controlling elastic properties in carbon-
ate sediments and rocks: The Leading Edge, v. 22,
no. 7, p. 654 – 660.
Erba, E., J. E. T. Channell, M. Claps, C. Jones, R. Larson,
B. Opdyke, I. Premoli-Silva, A. Riva, G. Salvini and
S. Torricelli, 1999, Integrated stratigraphy of the Cismon
APTICORE (southern Alps, Italy): A ‘‘reference section’’
for the Barremian – Aptian interval at low latitudes:
Journal of Foraminiferal Research, v. 29, p. 371 – 391.
Greselle, B., and B. Pittet, 2005, Fringing carbonate plat-
forms at the Arabian plate margin in northern Oman
during the late Aptian – middle Albian: Evidence for
high-amplitude sea-level changes: Sedimentary Geol-
ogy, v. 175, p. 367 – 390.
Hampson, D., J. S. Schuelke, and J. Quirein, 2001, Use of
multi-attribute transforms to predict log properties
from seismic data: Geophysics, v. 66, no. 1, p. 220 –
236.
Hardenbol, J., J. Thierry, M. B. Farley, T. Jacquin, P. De
Graciansky, and P. R. Vail, 1998, Mesozoic and Ceno-
zoic sequence chronostratigraphic framework of Euro-
pean basins, in P. De Graciansky, J. J. Hardenbol, T.
Jacquin, and P. R. Vail, eds., Mesozoic and Cenozoic
sequence stratigraphy of European basins: SEPM Spe-
cial Publication 60, p. 3 – 14.
Hughes, G. W., 2000, Bioecostratigraphy of the Shuaiba
Formation, Shaybah field, Saudi Arabia: GeoArabia,
v. 5, p. 545 – 578.
Hunt, D., and M. E. Tucker, 1992, Stranded parasequences
and the forced regressive wedge systems tract: Depo-
sition during base-level fall: Sedimentary Geology,
v. 81, p. 1 – 9.
Immenhauser, A., A. Cresusen, M. Esteban, and H. B.
Vonhof, 2000, Recognition and interpretation of poly-
genic discontinuity surfaces in the middle Cretaceous
Shuaiba, Nahr Umr, and Natih formations of northern
Oman: GeoArabia, v. 5, no. 2, p. 299 – 322.
Johnson, C. A., T. Hauge, S. Al-Mehhali, S. B Sumaidaa,
B. Sabin, and B. West, 2005, Structural styles and tec-
tonic evolution of onshore and offshore Abu Dhabi,
U.A.E.: International Petroleum Technology Confer-
ence (Qatar) Paper 10646, CD-ROM.
Marzouk, I. M., and M. A. Sattar, 1993, Implications of
wrench tectonics on hydrocarbon reservoirs, Abu
Dhabi, U.A.E.: Proceedings of 8th Middle East Oil Show
and Conference, Bahrain, Society of Petroleum Engi-
neers Paper 25608, p. 119 – 130.
Masafarro, J. L., R. Bourne, and J. C. Jauffred, 2004, Three-
dimensional seismic volume visualization of carbon-
ate reservoirs and structures, in G. P. Eberli, J. L.
Masaferro, and J. F. Sarg, eds., Seismic imaging of
carbonate reservoirs and systems: AAPG Memoir 81,
p. 11 – 41.
Plint, A. G., and D. Nummedal, 2000, The falling stage
systems tract: Recognition and importance in se-
quence stratigraphic analysis, in D. Hunt and R. L.
Gawthorpe, eds., Sedimentary responses to forced
regressions: Geological Society (London) Special Pub-
lication 172, p. 1 – 18.
Premoli Silva, I., E. Erba, G. Salvini, C. Locatelli, and D. Verga,
1999, Biotic changes in Cretaceous oceanic anoxic
events of the Tethys: Journal of Foraminiferal Research,
v. 29, p. 352– 370.
Russell, S. D., M. Akbar, B. Vissapragada, and G. M. Walkden,
2002, Rock types and permeability prediction from
dipmeter and image logs: Shuaiba reservoir (Aptian),
Abu Dhabi: AAPG Bulletin, v. 86, p. 1709 – 1732.
Sarg, J. F., and J. S. Schuelke, 2003, Integrated seismic
analysis of carbonate reservoirs: From the framework
to the volume attributes: The Leading Edge, v. 22,
no. 7, p. 640 – 645.
Schuelke, J. S., L. A. Yose, S. Al-Menhali, and W. Soroka,
212 / Yose et al.
2005, Seismic rock property predictions provide insight
to field development (abs.): AAPG Annual Meeting
Program, v. 14, p. A125.
Sharland, P. R., R. Archer, D. M. Casey, R. B. Davies, S. H.
Hall, A. P. Heward, A. D Horbury, and M. D. Simmons,
2001, Arabian plate sequence stratigraphy: GeoArabia
Special Publication 2, 371 p.
Sharland, P. R., D. M. Casey, R. G. Davies, M. D. Simmons,
and O. E. Sutcliffe, 2004, Arabian plate sequence
stratigraphy — Revisions to SP2: GeoArabia, v. 9, no. 1,
p. 199 – 214.
Strohmenger, C. J., L. J. Weber, A. Ghani, K. Al-Mehsin,
O. Al-Jeelani, A. Al-Mansoori, T. Al-Dayyani, L. Vaughan,
S. A. Khan, and J. C. Mitchell, 2006, High-resolution
sequence stratigraphy and reservoir characterization of
upper Thamama (Lower Cretaceous) reservoirs of a
giant Abu Dhabi oil field, United Arab Emirates, in P. M.
Harris and L. J. Weber, eds., Giant hydrocarbon reser-
voirs of the world: From rocks to reservoir characteri-
zation and modeling: AAPG Memoir 88/SEPM Special
Publication, p. 139– 171.
van Buchem, F. S. P., B. Pittet, H. Hillgartner, J. Grotsch,
A. I. Al Mansouri, I. M. Billig, H. H. J. Droste, W. H.
Oterdoom, and M. van Steenwinkel, 2002, High-
resolution sequence stratigraphic architecture of
Barremian/Aptian carbonate systems in northern
Oman and the United Arab Emirates (Kharaib and
Shuaiba formations): GeoArabia, v. 7, p. 461 – 500.
Yose, L. A., and J. F. Collins, 2002, Windward-leeward
models for carbonate platforms revisited (abs.); AAPG
Annual Meeting Program, v. 11, p. A196.
Yose, L. A., et al., 2004, New frontiers in 3-D seismic char-
acterization of carbonate reservoirs: Example from a
supergiant field in Abu Dhabi: 11th Abu Dhabi In-
ternational Petroleum Exhibition and Conference,
Society of Petroleum Engineers, SPE Paper 88689, 16 p.
 Strohmenger, C. J., P. E. Patterson, G. Al-Sahlan, J. C. Mitchell, H. R. Feldman,

6
T. M. Demko, R. W. Wellner, P. J. Lehmann, G. G. McCrimmon, R. W.
Broomhall, and N. Al-Ajmi, 2006, Sequence stratigraphy and reservoir
architecture of the Burgan and Mauddud formations (Lower Cretaceous),
Kuwait, in P. M. Harris and L. J. Weber, eds., Giant hydrocarbon reservoirs
of the world: From rocks to reservoir characterization and modeling:
AAPG Memoir 88/SEPM Special Publication, p. 213 – 245.

Sequence Stratigraphy and

Reservoir Architecture of the
Burgan and Mauddud Formations
(Lower Cretaceous), Kuwait
Christian J. Strohmenger,1 Timothy M. Demko
John C. Mitchell, Howard R. Feldman, University of Minnesota Duluth, Duluth,
Patrick J. Lehmann, and Minnesota, U.S.A.
Robert W. Broomhall
ExxonMobil Exploration Company, Houston, Robert W. Wellner
Texas, U.S.A. ExxonMobil Upstream Research Company,
Houston, Texas, U.S.A.
Penny E. Patterson
ExxonMobil Upstream Research Company, G. Glen McCrimmon
Houston, Texas, U.S.A. Hibernia Management and Development
Company, St. John’s, Newfoundland
Ghaida Al-Sahlan and Labrador, Canada
Kuwait Oil Company, Ahmadi, Kuwait
Neama Al-Ajmi
Kuwait Oil Company, Ahmadi, Kuwait

ABSTRACT

A
new sequence-stratigraphic framework is proposed for the Burgan and
Mauddud formations (Albian) of Kuwait. This framework is based on the
integration of core, well-log, and biostratigraphic data, as well as seismic
interpretation from giant oil fields of Kuwait.
The Lower Cretaceous Burgan and Mauddud formations form two third-
order composite sequences, the older of which constitutes the lowstand, trans-
gressive, and highstand sequence sets of the Burgan Formation. This composite
sequence is subdivided into 14 high-frequency, depositional sequences that are
characterized by tidal-influenced, marginal-marine deposits in northeast Kuwait
that grade into fluvial-dominated, continental deposits to the southwest.

1
Present address: Abu Dhabi Company for Onshore Oil Operations, Abu Dhabi, United Arab Emirates.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215878M883271

213
214 / Strohmenger et al.

The younger composite sequence consists of the lowstand sequence set of

the uppermost Burgan Formation and transgressive and highstand sequence sets
of the overlying Mauddud Formation. This composite sequence is sand prone
and mud prone in southern and southwestern Kuwait and is carbonate prone
in northern and northeastern Kuwait. The lowstand sequence set deposits of
the Burgan Formation are subdivided into five high-frequency depositional se-
quences, which are composed of tidal-influenced, marginal-marine deposits in
northeastern Kuwait that change facies to fluvial-dominated deposits in south-
western Kuwait. The transgressive and highstand sequence sets of the Mauddud
Formation are subdivided into eight high-frequency, depositional sequences. The
Mauddud transgressive sequence set displays a lateral change in lithology from
limestone in northern Kuwait to siliciclastic deposits in southern and south-
western Kuwait. The traditional lithostratigraphic Burgan –Mauddud contact is
time transgressive. The Mauddud highstand sequence set is carbonate prone and
thins south- and southwestward because of depositional thinning. Significant
postdepositional erosion occurs at the contact with the overlying Cenomanian
Wara Shale.
The proposed sequence-stratigraphic framework and the incorporation of a
depositional facies scheme tied to the sequence-stratigraphic architecture allow
for an improved prediction of reservoir and seal distribution, as well as reservoir
quality away from well control.

INTRODUCTION Al-Eidan et al., 2001) (Figures 1, 2). The recoverable

A regional sequence-stratigraphic analysis of the
Lower Cretaceous Burgan and Mauddud formations
was undertaken through a joint study conducted by
ExxonMobil Exploration Company and Kuwait Oil
Company. The study focused on the stratigraphic ar-
chitecture of selected major oil fields throughout Ku-
wait, including the supergiant Greater Burgan field
in southeastern Kuwait, and Raudhatain and Sabiri-
yah fields in northern Kuwait (Figure 1).
Three culminations constitute the supergiant Great-
er Burgan field: Burgan, Magwa, and Ahmadi. These
three culminations are located near the crest of the
north–south-trending Kuwait arch (Fox, 1961; Ada-
sani, 1965; Brennan, 1990b; Carman, 1996) (Figure 1).
The first well on these structures was drilled in 1938,
followed by wells in 1951 and 1952. Production of
28 – 368 API oil comes from the Burgan (the major
oil-producing reservoir), the Mauddud (a minor oil-
producing reservoir), and the Wara formations (Kauf-
man et al., 1997). The recoverable oil reserves are esti-
mated to be in the tens of billions of barrels (Christian,
1997).
Raudhatain field was discovered in 1955. It is a
faulted anticlinal dome with production of 28 – 408
API oil from the Ratawi, Zubair, Burgan, and Maud-
dud formations (Milton and Davies, 1965; Adasani,
1967; Al-Rawi, 1981; Brennan, 1990a; Carman, 1996;
oil reserves are estimated to be in the billions of bar-
rels (Christian, 1997).
Sabiriyah field was discovered in 1956. It is an
elongated faulted anticline with production of 28 –
328 API oil from the Burgan and Mauddud forma-
tions (Milton and Davies, 1965; Adasani, 1967; Al-
Rawi, 1981; Brennan, 1990a; Carman, 1996; Kaufman
et al., 1997; Al-Eidan et al., 2001) (Figures 1, 2). The
recoverable oil reserves are estimated to be in the bil-
lions of barrels (Christian, 1997).
The Burgan and Mauddud formations are part of
the Wasia Group that overlies the Lower Cretaceous
Thamama Group of the Arabian plate (Alsharhan and
Nairn, 1993, 1997). The lower to middle Albian Bur-
gan Formation is the major oil-bearing sandstone res-
ervoir throughout the Greater Burgan field, as well as
at Raudhatain and Sabiriyah fields in northern Kuwait.
The thickness ranges from approximately 1250 ft
(380 m) at the Greater Burgan field area to approxi-
mately 900 ft (275 m) at Raudhatain and Sabriya fields
area (Bou-Rabee, 1996). The overlying upper Albian
Mauddud Formation is a major oil-bearing carbonate
reservoir in northern Kuwait (Al-Anzi, 1995). Thick-
ness of the Mauddud Formation ranges from only a
few feet at the Greater Burgan field and Minagish field
areas (south and southwest Kuwait) to approximately
450 ft (140 m) at the Abdali, Raudhatain, and Sabi-
riyah fields in northern Kuwait (Bou-Rabee, 1996).
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 215
FIGURE 1. Location map showing the major oil fields of Kuwait (green) and the wells (black dots) studied.
Methodology
The emphasis of this study involved the interpreta-
tion of the sequence-stratigraphic architecture based on
correlation of regional stratal surfaces (sequence bound-
aries [SB], transgressive surfaces [TS], maximum flooding
surfaces [MFS], and flooding surfaces [FS]) through-
out Kuwait, as well as on the sequence-stratigraphy-
keyed facies analyses of the Burgan and Mauddud
formations. Approximately 9600 ft (2930 m) of con-
ventional core from 30 wells penetrating the Burgan
Formation and approximately 5000 ft (1520 m) of con-
ventional core from 35 wells penetrating the Mauddud
Formation were described sedimentologically. The de-
positional environments interpreted from core were
correlated to well-log signatures and used to develop
the regional sequence-keyed sequence-stratigraphic
framework. More than 100 wells were correlated with-
in this sequence-stratigraphic context. In addition,
the results from seismic-stratigraphic and biostrati-
graphic interpretations were integrated into this
study.
The identified chronostratigraphic surfaces (SB, TS,
MFS, and FS) were assigned to high-frequency se-
quences (HFS) and numbered sequentially from the
top down; the underlying sequence boundary giv-
ing name to the overlying high-frequency sequence
(Figure 3). The HFSs were subsequently grouped into the
sequence sets of two third-order composite sequences
(Mitchum, 1977; Mitchum et al., 1977; Vail et al., 1977,
1991; Haq et al., 1987, 1988; Vail, 1987; Van Wagoner
et al., 1987, 1988; Sarg, 1988; Haq, 1991; Mitchum
and Van Wagoner, 1991; Sarg et al., 1999) (Figure 3).
High-frequency Sequences
High-frequency sequences in the siliciclastic Bur-
gan Formation (Figure 3) are defined by the stacking
patterns and facies distributions. These sequences
may contain lowstand, transgressive, and highstand
216 / Strohmenger et al.
FIGURE 2. Seismic cross section oriented northwest – southeast showing Raudhatain and Sabiriyah structures and
the interpreted main stratigraphic horizons. Seismic line runs through the center of Raudhatain and Sabiriyah fields
shown in Figure 1.
systems tracts (LST, TST, and HST). Each systems tract
exhibits distinct facies trends, thickness distribu-
tions, and reservoir quality (Strohmenger et al., 2002;
Demko et al., 2003).
The LSTs consist of incised-valley fills (IVF). These
valleys become thinner, as well as more laterally
discontinuous and tidal influenced downdip to the
northeast. Transgressive systems tracts correspond to
retrogradational successions of coarsening-upward,
marginal-marine mudstones and sandstones that grade
into marine carbonates downdip and mudstone-prone
coastal-alluvial plain deposits updip. Highstand sys-
tems tracts correspond to progadational successions
of coarsening-upward, marginal-marine mudstones
and sandstones that grade updip into mudstone-
prone coastal-alluvial plain deposits (Strohmenger
et al., 2002; Demko et al., 2003).
High-frequency sequences in the carbonate-
dominated Mauddud Formation (Figure 3) are de-
fined by parasequence (PS) stacking patterns, facies
distributions, and microkarst or exposure surfaces.
High-frequency sequences of the Mauddud Forma-
tion may contain TST and HST (Strohmenger et al.,
2002; Demko et al., 2003).
Transgressive systems tracts are generally more mud
dominated with intercalated sandstones and glauco-
nitic sandstones in the northern Kuwait Raudhatain
and Sabiriyah fields area and grade into siliciclastics
toward the south at Greater Burgan field and toward
the southwest at Minagish field areas. Highstand sys-
tems tracts typically show an upward increase in
grain richness (graining upward) as well as porosity
(Strohmenger et al., 2002; Demko et al., 2003).
Composite Sequences
High-frequency sequences are grouped into se-
quence sets of two third-order composite sequences,
based on the stacking patterns and facies distribu-
tions (Figure 3). Each composite sequence consists
of a lowstand sequence set (LSS), transgressive se-
quence set (TSS), and a highstand sequence set (HSS)
(Strohmenger et al., 2002; Demko et al., 2003).
In the siliciclastic Burgan Formation of Kuwait,
LSSs are characterized by an aggradational stacking
of HFSs dominated by braided fluvial deposits. The
TSS exhibits an overall retrogradational stacking
pattern, dominated by nonmarine facies at the base,
whereas the uppermost HFSs contain increasing
marginal-marine components. The HSS forms an overall
progradational succession dominated by marginal-
marine facies, especially in northern Kuwait (Stroh-
menger et al., 2002; Demko et al., 2003).
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 217

FIGURE 3. Mauddud – Burgan sequence-stratigraphic framework showing the lower and upper third-order composite
sequences, as well as the interpreted high-frequency depositional sequences (Raudhatain field type well RA-G). GR =
gamma-ray log; MD = measured depth (feet); RES = resistivity log; NEU = neutron porosity log; DENS = density log.

The Mauddud TSS shows a lateral change in lithol-
ogy from limestone in northern Kuwait to siliciclas-
tics in southern and southwestern Kuwait. An overall
shoaling-upward or progradational signature char-
acterizes the HSS. Most of the upper HSS is removed
by erosion throughout much of southern and south-
western Kuwait (Strohmenger et al., 2002; Demko
et al., 2003).
BURGAN FORMATION
A regional sequence-stratigraphic analysis of the
Burgan and Mauddud formations reveals that the
traditional lithostratigraphic Burgan-Mauddud con-
tact is a time-transgressive facies boundary (Stroh-
menger et al., 2002; Demko et al., 2003). To define
coeval facies successions in both the Burgan and
Mauddud formations, a chronostratigraphically sig-
nificant regional flooding surface (B100_TS) was de-
fined as the Burgan-Mauddud contact (Figure 3).
Within this chronostratigaphic framework, the up-
permost Burgan and overlying Mauddud formations
form a composite sequence (Figure 3) that becomes
more siliciclastic prone to the southwest and carbon-
ate prone to the northeast. A second composite se-
quence encompasses the rest of the underlying Burgan
Formation (Figure 3). This composite sequence is dom-
inated by marginal-marine deposits to the northeast
and nonmarine deposits to the southwest.
The Burgan Formation, as defined in this study,
comprises 19 HFSs (Figure 3). Each of these HFSs
218 / Strohmenger et al.

FIGURE 4. Sequence-keyed depositional facies models for the Burgan Formation. In these models, the lowstand systems
tract consists of incised-valley deposits, whereas the trangressive and highstand systems tracts are composed of wave-
dominated shoreface depositional systems. GR = gamma-ray log.

contains LSTs, TSTs, and HSTs, each with distinct facies
trends, thickness distributions, and reservoir quality.
The LSTs of these sequences consist of IVF. These
valleys become thinner, as well as more laterally dis-
continuous and tidal influenced downdip to the
northeast. The TSTs of the Burgan sequences display a
systematic downdip to updip change from marine
carbonates to marginal-marine mudstones and sand-
stones to mudstone-prone alluvial- and coastal-plain
deposits. The HSTs are dominated by marginal-marine
sandstones and mudstones downdip and alluvial-
and coastal-plain mudstones and sandstones updip.
The shorelines in these highstands trend northwest –
southeast and are best developed in the northern
Raudhatain and Sabiriyah fields area.
The primary reservoirs in the Burgan Formation
are fluvial and tidal deposits that formed within the
incised valleys (Figure 4). Shoreline sandstones in the
TSTs and HSTs are also potential reservoirs (Figure 4).
However, these marginal-marine sandstones have low-
er porosity and permeabilities because of their finer
grained and increased clay matrix due to bioturba-
tion. In summary, the sequence-stratigraphic analy-
sis of the Burgan Formation provides an improved
understanding of the spatial and temporal distribution
of reservoirs that can be used to address exploration-
scale to production-scale issues.
In general, the Burgan Formation is a classic re-
gressive-transgressive-regressive package. It is domi-
nated by sandstone-prone fluvial deposits at its base
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 219
and top and marginal-marine mudstones, sandstones,
and limestones in the middle parts to the northeast.
The Burgan Formation is bounded at its top by a re-
gional flooding surface, referred to as the Burgan trans-
gressive surface B100_TS (Figure 3). In general, this
boundary is marked by a change from blocky, high
net/gross, fluvial-dominated sandstones of the Bur-
gan Formation below to low net/gross marine mud-
stones, sandstones, and limestones of the Mauddud
Formation above. Within this context, fine-grained
siliciclastics, traditionally included at the top of the
Burgan Formation to the south, are assigned to the
clastic member of the Mauddud Formation. This Maud-
dud clastic member occurs above the Burgan transgres-
sive surface B100_TS, a regional flooding surface, de-
fined as the top of the Burgan Formation and beneath
the carbonate strata traditionally included within the
Mauddud Formation (Mauddud carbonate member).
Lithofacies and Depositional Environments
Within the Burgan Formation, 22 lithofacies were
identified and define 6 distinct facies. The physical
criteria used to delineate the individual lithofacies in
this study include grain size, composition, sorting,
grading, physical and biogenic sedimentary struc-
tures, stratal boundaries, presence or absence of clay
drapes, organic-rich drapes, organic debris, and dia-
genetic features. For ease of description, each litho-
facies was classified into six categories based on mud
content and apparent reservoir quality. The six cate-
gories range from well-sorted, very clean sand (litho-
facies 1: excellent reservoir quality) to mudstone (litho-
facies 5: very poor reservoir quality) and, ultimately,
to coal (lithofacies 6). In addition, lithofacies were
assigned an alphabetic qualifier to account for subtle
differences in grain size, mud content, dominant sedi-
mentary structures, and type and extent of bioturba-
tion. Trace fossil assemblages were used to infer en-
vironments of deposition, which are based on models
proposed by Pemberton et al. (1992a, b). In addition,
our regional database includes two lithofacies that
were not observed in the Burgan strata. They are litho-
facies 3A, coarsening-upward, bioturbated mudstone
to sandstone; and lithofacies 4C, bioturbated to pla-
nar laminated mudstone to sandstone.
Lithofacies 1A: Trough Cross-bedded Sandstone
Lithofacies 1A (Figure 5A) consists of poorly to
moderately sorted, fine-grained to coarse-grained sand-
stone. The dominant sedimentary structures are trough
cross-bedding and current-ripple cross-lamination.
Large plant fossils, including compressed branches
and leaves, are common. Mudstone drapes and bio-
turbation are uncommon. Grain size trends are typi-
cally fining upward. Small-scale fining-upward beds
are 0.2–2 ft (0.06–0.6 m) thick and typically asso-
ciated with individual trough cross-bed sets. Larger-
scale, fining-upward trends range from 20 to 50 ft
(6 to 15 m) of thickness.
The cross-bedding and grain-size trends all indicate
deposition in a low-sinuosity fluvial setting. Some
fining-upward trends suggest sedimentation in sandy
point bars. This lithofacies occurs primarily in incised
valleys and in updip braid-plain deposits.
Lithofacies 1B: Trough Cross-bedded Sandstone
with Minor Clay Drapes
Lithofacies 1B (Figure 5B) consists of poorly to mod-
erately sorted, fine-grained to coarse-grained sand-
stones that possess sparse to common clay drapes and
thick-thin couplets. Sedimentary structures are domi-
nated by trough- and current-ripple cross-bedding.
Small mudstone and siderite clasts may be locally
abundant. Small-scale, grain-size trends are typically
fining upward. Large-scale, grain-size trends are ei-
ther coarsening upward or fining upward. Large plant
fossils, such as compressed sticks and leaves, may be
common. Bioturbation is sparse to moderate. Common
burrow types are horizontal sand-filled tubes with a
circular cross section, approximately 1 cm (0.4 in.) in
diameter (Planolites).
The clay drapes and thick-thin couplets indicate
that some tidal influence occurred during deposition.
This lithofacies occurs at the fluvial to tidal transi-
tion in IVF and at the top of coarsening-upward tidal
bars.
Lithofacies 1C: Current-rippled,
Cross-laminated Sandstone
This lithofacies (Figure 5C) consists of well-sorted,
very fine-grained to fine-grained sandstone. Sedimen-
tary structures are characterized by current-ripple
cross-lamination with minor small-scale trough cross-
bedding. Clay drapes are sparse to absent. This facies
occurs in beds typically a few feet thick with no grain-
size trends. Bioturbation is uncommon. Small plant
fossils, such as small leaves and fine plant fragments,
are common.
Lithofacies 1C occurs in a range of environments.
Where it is associated with coastal-plain deposits, it
220 / Strohmenger et al.

FIGURE 5. Siliciclastic lithofacies 1, slabbed core photographs. (A) Lithofacies 1A: trough cross-bedded sandstone.
(B) Lithofacies 1B: trough cross-bedded sandstone with minor clay drapes. (C) Lithofacies 1C: current-rippled, cross-
laminated sandstone, which is interbedded with small-scale, trough cross-beds.

represents splays and channel levees. Within distal
incised valleys, this lithofacies represents sedimen-
tation in tidal channels.
Lithofacies 2A: Clay-draped, Current-rippled to
Laminated Sandstone
Lithofacies 2A (Figure 6A) consists of well-sorted,
very fine- to fine-grained sandstone. Sedimentary
structures include current-ripple cross-laminations
and minor horizontal laminations, with abundant
thin (<1-mm; <0.04-in.) clay and organic drapes, com-
monly on ripple foresets. Some laminations exhibit
centimeter-scale cyclicity in mud-sand couplets. Sider-
ite cement is common. Bedsets typically fine upward.
Bioturbation is uncommon to moderate and consists
of a low diversity of burrows, including Planolites,
Skolithos, and Arenicolites. Fine plant debris and amber
flakes are common.
This lithofacies represents deposition in tidal creeks
or tidal flats.
Lithofacies 2B: Clay-draped, Current-rippled to
Trough Cross-bedded Sandstone
Lithofacies 2B (Figure 6B) consists of very fine-
to fine-grained sandstone. The dominant sedimen-
tary structures are current-ripple cross-laminations
to small-scale, trough– cross-bedding with abundant
thin (1-mm) clay and organic drapes, commonly
on ripple and trough foresets. Bedsets may distinctly
coarsen upward. Bioturbation is sparse to moderate
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 221

FIGURE 6. Siliciclastic lithofacies 2, slabbed core photographs. (A) Lithofacies 2A: clay-draped, current-rippled to
laminated sandstone. (B) Lithofacies 2B: clay-draped, current-rippled to trough cross-bedded sandstone. (C) Lithofacies
2C: clay-draped, horizontally laminated sandstone. (D) Lithofacies 2D: glauconitic sandstone that has been extensively
bioturbated by Teichichnus (Te) and Planolites (Pl) burrows.

and is characterized by the low diversity of burrows
(mostly Planolites). Fine plant debris and amber flakes
are locally abundant.
This lithofacies is inferred to have been deposited
as subtidal bars in estuaries.
Lithofacies 2C: Clay-draped, Horizontally
Laminated Sandstone
Lithofacies 2C (Figure 6C) is dominated by hori-
zontal laminations, minor current ripples with abun-
dant thin (1-mm) clay and organic drapes that de-
fine horizontal laminations. Most laminations show
some centimeter-scale cyclicity of mud-sand couplets.
Lamina and bedsets typically fine upward. Biotur-
bation is sparse to moderate and is characterized by
Planolites, Palaeophycus, Arenicolites, Skolithos, and
Cylindrichnus. Fine plant debris and amber flakes and
pebbles may be locally abundant.
This lithofacies is interpreted to have been depos-
ited on tidal flats and in associated low-energy, tidal
point bars.
Lithofacies 2D: Glauconitic Sandstone
This lithofacies (Figure 6D) consists of fine-grained
to medium-grained sandstone with abundant grains
of glauconite (which are invariably coarser grained
than those of the detrital quartz). Bioturbation is mod-
erate to extensive and is characterized by Thalassi-
noides, Asterosoma, Planolites, Teichichnus, Palaeophycus,
and Scolicia. Siderite concretions are common, par-
ticularly associated with Thalassinoides burrows. This
facies is almost always cemented by calcite and grades
into overlying limestone beds. Mollusk shells and
shell fragments are common.
Lithofacies 2D represents deposition in the distal
parts of low-energy, marine shorelines.
Lithofacies 3B: Bioturbated, Muddy Sandstone
This lithofacies (Figure 7A) consists of poorly to
moderately well-sorted, upper fine-grained to lower
medium-grained muddy sandstone. Bioturbation is
typically extensive and is characterized by multiple
tiers of Thalassinoides, Teichichnus, Asterosoma, Pla-
nolites, Palaeophycus, Skolithos, and Cylindrichnus.
This lithofacies represents deposition along brack-
ish, estuarine, and marine low-energy shorelines.
Lithofacies 3C: Hummocky
Cross-bedded Sandstone
Lithofacies 3C (Figure 7B), which is not often ob-
served in the Burgan Formation, consists of mod-
erately well- to very well-sorted, very fine-grained to
222 / Strohmenger et al.

FIGURE 7. Siliciclastic lithofacies 3, slabbed core photographs. (A) Lithofacies 3B: bioturbated, muddy sandstone
containing shell fragments. (B) Lithofacies 3C: hummocky cross-bedded sandstone. (C) Lithofacies 3D: interlaminated
sandstone, silt, and shale. (D) Lithofacies 3E: current-rippled, interbedded sandstone and mudstone.

lower fine-grained sandstone. The dominant sedi-
mentary structures are hummocky, cross-stratified
beds with wave-rippled and bioturbated upper sur-
faces. Mudstone beds as much as several centimeters
in thickness are only rarely preserved between hum-
mocky bedsets. Bioturbation is typically absent to
sparse, but may increase upward with moderate bur-
rowing at the top of some bed sets. Trace fossils rec-
ognized include Planolites, Palaeophycus, and uncom-
mon Ophiomorpha.
This lithofacies was most likely deposited in ma-
rine, proximal lower shoreface settings.
Lithofacies 3D: Interlaminated Sandstone,
Silt, and Shale
Lithofacies 3D (Figure 7C) consists of interlami-
nated, very fine-grained sandstone, silt, and shale.
Sedimentary structures include horizontal bedding,
starved current-ripple cross-lamination, and uncom-
mon syneresis cracks. Depositional strata range from
mostly clay to mostly sand, but are characterized by
sand and mud couplets, thick and thin couplets, and
centimeter-scale cycles of thickening and thinning
sand laminae. Bioturbation is sparse to moderate and
include trace fossils of Planolites, Palaeophycus, Sko-
lithos, and Cylindrichnus. Plant debris, amber, and
siderite-cemented bands are common.
This lithofacies represents deposition on proximal
(sandy) to distal (muddy) tidal flats, and the centimeter-
scale cycles are interpreted as neap-spring tidal cycles.
Lithofacies 3E: Current-rippled, Interbedded
Sandstone and Mudstone
Lithofacies 3E (Figure 7D) consists of interbedded,
lower medium-grained sandstone and mudstone.
Sedimentary structures are dominated by horizontal
lamination and current-ripple cross-lamination, with
uncommon syneresis cracks. Bioturbation is sparse
to moderate and is characterized mostly by Planolites,
Palaeophycus, Skolithos, and Cylindrichnus. Plant debris
and amber are locally abundant as well as siderite-
cemented bands.
This lithofacies is interpreted to have been depos-
ited in tidal flats and distal tidal bars.
Lithofacies 3F: Calcareous, Bioturbated Sandstone
This lithofacies (Figure 8A) consists of fine-grained
to lower medium-grained sandstone. Bioturbation
is typically extensive (churned). Mollusk shells, shell
fragments, and foraminifera are locally abundant.
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 223

FIGURE 8. Siliciclastic lithofacies 3, slabbed core photographs. (A) Lithofacies 3F: calcareous, bioturbated sandstone
containing shell fragments. (B) Lithofacies 3G: wave-rippled sandstone. (C) Lithofacies 3H: carbonaceous, muddy
sandstone.

This facies is tightly cemented by calcite and grades
into limestone beds above or below.
This lithofacies represents deposition in shallow-
marine shoreline or shoreface settings adjacent to
subtidal carbonate facies.
common to slight and comprises Planolites burrows
and rootworking. Large plant fossils and amber are
very abundant.
This lithofacies represents deposition in flood-
plain, crevasse splay, or levee settings.

Lithofacies 3G: Wave-rippled Sandstone Lithofacies 4A: Heterolithic Siltstone to Mudstone

Lithofacies 3G (Figure 8B) consists of well-sorted,
very fine-grained to fine-grained sandstone. Sedimen-
tary structures consist of wave-ripple cross-laminations.
Clay drapes are common between ripple laminae and
laminae sets. Bioturbation is sparse to moderate and
includes Astersoma, Teichichnus, and Planolites burrows.
This lithofacies is interpreted to have formed in
distal, lower shoreface settings.
Lithofacies 3H: Carbonaceous, Muddy Sandstone
Lithofacies 3H (Figure 8C) consists of very fine- to
fine-grained sandstone with abundant clay and or-
ganic drapes ranging from 5 to 10 mm (0.2 to 0.4 in.)
in thickness. Organic layers range from 1 mm (0.04 in.)
to a few centimeters in thickness. Sedimentary struc-
tures are dominated by horizontal lamination and
current-ripple cross-lamination. Bioturbation is un-
This lithofacies (Figure 9A) consists mostly of mud-
stone with siltstone to very fine-grained sandstone
stringers (>80% mudstone). Sedimentary structures
include parallel-lamination, silt stringers, uncom-
mon syneresis cracks, and starved current-ripple cross-
lamination. Bioturbation is uncommon to slight and
is characterized by Planolites burrows. Plant fossils
and amber are common to abundant.
Lithofacies 4A is interpreted to represent depo-
sition in flood-plain and abandoned channel-fill
settings.
Lithofacies 4B: Bioturbated to Wave-rippled
Sandstone to Mudstone
This lithofacies (Figure 9B) is composed of lower
very fine-grained sandstone to siltstone to clay. Sedi-
mentary structures include isolated to amalgamated
224 / Strohmenger et al.

FIGURE 9. Siliciclastic lithofacies 4, slabbed core photographs. (A) Lithofacies 4A: heterolithic siltstone to mudstone.
(B) Lithofacies 4B: bioturbated to wave-rippled sandstone to mudstone. This core interval has been extensively bio-
turbated by Teichichnus (Te), Asterosoma (As), and Planolites (Pl) burrows. (C) Lithofacies 4D: bioturbated mudstone. The
upper interval of this core has been extensively bioturbated by Teichichnus (Te) burrows. (D) Lithofacies 4E: laminated
siltstone.

wave-ripple cross-laminations. Bioturbation is slight Lithofacies 5A: Laminated Gray Shale

to extensive (churned) and is characterized by Aste-
rosoma, Teichichnus, Thalassinoides, Planolites, Palaeo-
phycus, Scolicia, and Skolithos. Fine plant fragments
and mollusk shells are uncommon.
This lithofacies is interpreted as deposition in ma-
rine, distal lower shoreface settings.
Lithofacies 4D: Bioturbated Mudstone
Lithofacies 4D (Figure 9C) is mostly mudstone,
but may grade vertically into siltstone. Bioturbation
is moderate to extensive (churned) and is charac-
terized by mostly indistinct burrows with some dis-
cernible Planolites, Teichichnus, and Thalassinoides
burrows.
This lithofacies is inferred to have been deposited
in offshore to lower shoreface settings.
Lithofacies 4E: Laminated Siltstone
This lithofacies (Figure 9D) is composed of hori-
zontally laminated to wavy laminated siltstone. Bio-
turbation is uncommon to slight and is character-
ized by small (<5-mm; <0.2-in.) indistinct horizontal
burrows.
Lithofacies 4E represents deposition in lacustrine,
flood-plain, pond, and abandoned-channel settings.
Lithofacies 5A (Figure 10A) is composed of dark-
gray, horizontally laminated shale. Bioturbation is
slight to moderate and is characterized by Planolites,
Teichichnus, sand-filled Chondrites, and sparse Zoo-
phycos burrows. Bivalve and gastropod shells range
from sparse to common, and small plant fragments
are uncommon.
This lithofacies is interpreted to have been deposited
in offshore marine settings, below storm-wave base.
Lithofacies 5B: Carbonaceous Mudstone
Lithofacies 5B (Figure 10B) is composed of car-
bonaceous, horizontally laminated mudstone, which
commonly displays postdepositional, compactional
slickensides. Bioturbation is slight to moderate and
is characterized mostly by rootworking (Figure 11A).
Large leaves and amber (Figure 11B, C) are very
abundant.
This lithofacies represents deposition in clastic
swamp and abandoned-channel settings.
Lithofacies 5C: Laminated Dark Gray Shale
Lithofacies 5C (Figure 10C) is composed of very
dark-gray, horizontally laminated shale. Bioturbation
is uncommon to slight and is characterized by small
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 225

FIGURE 10. Siliciclastic lithofacies 5 and 6, slabbed core photographs. (A) Lithofacies 5A: laminated gray shale.
(B) Lithofacies 5B: carbonaceous mudstone. (C) Lithofacies 5C: laminated dark-gray shale. (D) Lithofacies 6: coal.

Planolites burrows. Large leaves, sticks,

and amber are very abundant.
This lithofacies is interpreted to
represent deposition in lacustrine,
abandoned channel, and flood-
plain settings.

Lithofacies 6: Coal
This lithofacies (Figure 10D) is com-
posed exclusively of coal.
Some coals are rooted and repre-
sent peat swamps. Other coals are
composed of allochthonous plant
detritus and formed in abandoned
channels to estuarine settings with
low clastic input.

Facies, Facies Associations,

Depositional Environments, and
Reservoir Quality
The 22 lithofacies are grouped into
6 distinct facies. These groupings are
either spatially reoccurring groups
FIGURE 11. Slabbed core photographs of (A) rooted horizon, (B) amber, and of different lithofacies or thick oc-
(C) fragment of amber. currences of the same lithofacies. In
226 / Strohmenger et al.
turn, the six facies are grouped into two unique facies
associations. The facies associated with each are mu-
tually exclusive (Figure 4).
Facies Association I
Facies association I (Figure 4) consists of four
unique facies, which are interpreted as the normal,
shoaling-upward facies succession of a low-energy,
wave-dominated, shoreline (sensu Walker and Plint,
1992). This facies association is the basic building
block of highstand and transgressive deposits in the
Burgan Formation. The typical idealized vertical
facies succession in this facies association from top
to base is
 carbonaceous facies (lithofacies 1C, 3H, 5B, and
6): coastal-plain and backshore
 bioturbated to stratified sandstone facies (litho-
facies 3B and 3C): proximal lower shoreface
 bioturbated to interstratified mudstones and sand-
stones (lithofacies 2D, 3F, 3G, 4B, and 4D): distal
lower shoreface
 bioturbated mudstone facies (lithofacies 5A):
offshore
Lithofacies association I is interpreted to represent
the deposits of a low-energy shoreline environment.
Wave-dominated shoreface sandstones are oriented
along a northwest-southeast belt and primarily occur
in northeastern Kuwait. Shoreface deposits change
facies to marginal-marine sandstones and mudstones
and offshore mudstones in a northeast transect. They
change facies updip to coastal-plain deposits, which
dominate the southwestern region of Kuwait. The
shoreface deposits are interpreted to represent a low-
energy environment based on the paucity of high-
energy stratification and the presence of strata that
have been moderately to intensely bioturbated by
trace fossil assemblages indicative of open-marine
conditions.
In general, reservoir quality ranges from moderate
in deposits of the proximal lower shoreface to poor in
the distal lower shoreface and offshore strata. Al-
though sandstone reservoirs may exist within small
(<0.5-km [<0.31-mi]-wide) fluvial channels on the
coastal-plain, for the most part, the coastal-plain part
of this succession is mud prone and nonprospective.
Facies Association II
Facies association II (Figure 4) consists of two
unique facies, which are interpreted as the normal
updip to downdip facies variation in an incised-
valley system (sensu Van Wagoner et al., 1990). The
typical idealized Burgan updip to downdip facies
succession and associated lithofacies are
 cross-stratified sandstone facies (lithofacies 1A
and 1C): fluvial-dominated valley fill
 heterolithic tidal facies (lithofacies 1B, 2A, 2B,
2C, 3D, 3E, and 4A): tidal-dominated valley fill
The incised valleys, identified in this study, typi-
cally trend southwest to northeast across Kuwait and
are perpendicular to somewhat oblique to the shore-
line trend of the underlying highstand and overlying
transgressive deposits. Deposits of the incised valley
are characterized by fluvial-dominated facies in up-
dip regions to the southwest and tidal-dominated
facies in the distal regions to the northeast. The
fluvial facies are interpreted as low-sinuosity, braided-
stream deposits based on the prevalence of fining-
upward, fine-grained to coarse-grained, trough–cross-
beds and bedsets. The fluvial strata change facies
downdip to tidal-influenced deposits that possess more
heterolithic lithofacies, indicative of fluctuation-energy
conditions.
Reservoir quality is excellent in the braided fluvial
strata and in the updip parts of the mixed fluvial-tidal
systems. However, reservoir quality rapidly diminishes
in the seaward direction as the valley systems thin
and become more mudstone prone.
High-frequency Sequences
Nineteen high-frequency depositional sequences
were interpreted within the Burgan Formation, as
defined in this study (Figure 3). From bottom to top,
these are B900, B850, B800, B750, B725, B700, B650,
B600, B550, B500, B450, B400, B350, B300, B250,
B200, B150, B125, and B100 (LST).
In the nonmarine to marine intervals of the
Burgan Formation, high-frequency sequence bound-
aries are interpreted as abrupt vertical changes in
stratal stacking pattern or abrupt basinward shift in
environments of deposition. In the downdip position,
high-frequency sequence boundaries are evident as
abrupt changes from coarsening-upward marginal-
marine parasequences to blocky fluvial-tidal deposits.
Updip, these sequence boundaries can be traced into
an abrupt change from interpreted low net/gross
coastal-alluvial plain (transgressive or highstand) de-
posits (below) into blocky fluvial (lowstand) deposits
(above; Figure 12). Downdip and laterally, the se-
quence boundaries and transgressive surface become
coincident. However, in the most distal sequences in
FIGURE 12. Regional cross section showing facies distribution and sequence-stratigraphic framework of the Burgan Formation. Note that the fluvial-dominated
sandstones of the incised-valley fills, which constitute the lowstand systems tracts, thicken toward the south in the vicinity of the Greater Burgan field area.
The overlying Mauddud clastic member and Mauddud carbonate member also display thickness variations along the depositional transect. The detailed
sequence-stratigraphic framework for the Mauddud Formation is shown in Figures 22 and 23. GR = gamma-ray log; MD = measured depth (feet); RES = resistivity
log; DT = sonic log; NEU = neutron porosity log; DENS = density log.
Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 227
228 / Strohmenger et al.

dominated strata in the updip regions. (B) Incised valleys become thinner and more discontinuous downdip. IVF = incised-valley fills; HST = highstand systems tract.
FIGURE 13. Depositional environment and isopach maps of a lowstand systems tract in the Burgan Formation. These maps are interpreted from the sequence-

upper, third-order composite sequence. (A) The lowstand deposits fill incised valleys and are composed of tidal-influenced strata in the downdip positions and fluvial-
the Burgan Formation, an onlap-

stratigraphic architecture of the B100 high-frequency, depositional sequence, which is the uppermost lowstand systems tract in the lowstand sequence set of the
ping succession of carbonates ap-
pears to mark a depositional shelf
break and downdip limit of clas-
tic progradation in each sequence.
In more marine-dominated
parts of the Burgan Formation,
the TSs are interpreted at the ver-
tical change from prograding
to retrograding marine parase-
quences. In more medial settings,
these surfaces coincide with the
vertical facies change, from blocky,
fluvial-tidal deposits below to
retrograding marine parase-
quences above. In more proxi-
mal (updip) settings, the TSs are
placed at the vertical change from
blocky fluvial-tidal below to low
net/gross coastal- and alluvial-
plain deposits above (Figure 12).
The MFSs of these sequences
are interpreted at major marine
incursions marked by a change
from retrograding to prograd-
ing stacking of marine parase-
quences. In the downdip parts
of the most distal sequences, ma-
rine limestones occur beneath in-
terpreted MFSs (Figure 12).
Lowstand systems tracts pri-
marily consist of incised-valley
systems that are filled by tidal-
influenced sandstones and mud-
stones in the downdip regions
and braided fluvial sandstones
in updip regions (Figure 13A). In
the low-relief, HFSs identified
within the Burgan Formation, on-
lapping lowstand clastic wedges,
as well as basin-floor lowstand
fans are absent. Clastic lowstand
deposition appears limited to
IVF. The valleys become thinner
and more discontinuous down-
dip (Figure 13B) and do not ap-
pear to extend to the maximum
downdip progradational limit of
underlying sandstone-prone
highstands. These relationships
strongly suggest that transgres-
sive erosion has modified the
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 229

FIGURE 14. Paleogeographic map showing the depositional environments of a highstand systems tract in the Burgan
Formation. This paleogeographic map is interpreted from the sequence-stratigraphic architecture of the B600 high-
frequency, depositional sequence, which is a highstand systems tract in the transgressive sequence set of the lower,
third-order composite sequence.

primary distribution and thickness patterns in the
most distal (downdip) parts of the Burgan incised-
valley systems.
The TSTs correspond to retrogradational succes-
sions of coarsening-upward strata and display a
systematic downdip to updip change from marine
carbonates, to marginal-marine mudstones and sand-
stones, to mudstone-prone alluvial- and coastal-plain
deposits.
Highstand systems tracts correspond to progada-
tional successions of coarsening-upward, dominantly
marginal-marine sandstones and mudstones down-
dip and alluvial- and coastal-plain mudstones and
sandstones updip (Figure 14). The shorelines in these
highstands trend northwest– southeast and are best
developed in northern Kuwait.
Composite Sequences
The 19 HFSs identified within the Burgan Forma-
tion define four distinct sequence sets that form parts
of two composite sequences. The basal four HFSs
(B900, B850, B800, and B750) stack in an overall ag-
gradational succession and form the LSS of the
lower Burgan composite sequence (Figure 3). This
aggradational sequence set is dominated by braided
fluvial deposits that contain few mudstone breaks
230 / Strohmenger et al.
(Figure 12). The basal aggradational sequence set is
overlain by six HFSs (B725, B700, B650, B600, B550,
and B500) that stack in an overall retrogradational
pattern, which constitutes the TSS of the lower
Burgan composite sequence (Figure 3). In the Raud-
hatain and Sabiriyah field area in northern Kuwait,
the basal HFSs in this retrogradational succession are
dominated by nonmarine facies, whereas the upper-
most sequences contain increasing marginal-marine
components (Figure 12). The next four HFSs (B450,
B400, B350, and B300) stack in an overall prograda-
tional succession and comprise the HSS of the lower
Burgan composite sequence (Figure 3). In the north-
ern fields area, these sequences are dominated by
marginal-marine highstand deposits (Figure 12). The
uppermost five sequences (B250, B200, B150, B125,
and the LST of B100) stack in an aggradational pattern
and form the LSS of the upper Burgan and Mauddud
upper composite sequence (Figure 3). In the northern
fields area, these HFSs consist of alternating non-
marine LSTs and marginal-marine TSTs and HSTs
(Figure 12).
Based on this delineation of sequence sets, a com-
posite sequence boundary B900_SB is placed at the
base of the Burgan Formation (Figure 3). In general,
the overlying aggradational, retrogradational, and
progradational sequence sets correspond to the LSS,
TSS, and HSS of the lower composite sequence. It
should be noted, however, that the high-frequency
transgressive surface B725_TS and the high-frequency
maximum flooding surface B500_MFS are used as
composite TS and MFS, respectively, for the lower
composite sequence (Figures 3, 12). The uppermost
aggradational sequence set is interpreted as a second
LSS. The composite transgressive surface B100_TS is
interpreted as the TS of this composite sequence, with
strata in the overlying Mauddud Formation, form-
ing the TSS and the HSS of the younger composite
sequence.
In general, the various sequence sets are thicker
and more sandstone prone in the southwest area of
Kuwait and thinner and more mudstone prone in the
northeast (Figures 12 – 14).
Reservoir Quality
Reservoir quality of the Burgan Formation is closely
related to the interpreted depositional environ-
ments. Fluvial-dominated sandstones, which include
lithofacies 1A and 1C, possess the best reservoir-
quality attributes. These lithofacies types have av-
erage porosity and permeability values of 25% and
1600 md, respectively. Tidal-dominated sandstones,
which encompass lithofacies 1B, 2A, 2B, 2C, and 2F,
exhibit moderately good reservoir-quality properties.
They have average porosity and permeability values of
23% and 270 md, respectively. The marginal-marine
deposits, which include lithofacies 2A, 2E, and 3B,
generally display poorer reservoir quality as a result
of variable extents of bioturbation. Average porosity
and permeability values for these lithofacies are 19%
and 10 md, respectively.
Reservoir Quality Distribution
Stratal distributions, thickness variations, and re-
gional facies architecture of the Burgan systems tracts
can be interpreted from isopach and paleogeograph-
ic maps of these intervals (Figures 13, 14). High-
frequency lowstands of the Burgan Formation consist
of fluvial and tidal deposits that filled incised valleys.
In general, the incised valleys become thinner, nar-
rower, and more tidal influenced downdip to the
northeast (Figure 13). Conversely, the valleys become
thicker, more widespread, and more fluvially influ-
enced updip to the south and southwest (Figure 13).
High-frequency highstands of the Burgan consist of
shoreline deposits, which, in general, trend north-
west–southeast, with more marine facies to the north-
east and more nonmarine facies toward the south-
west (Figure 14).
The sequence-stratigraphic framework presented
in this paper provides an improved understanding of
the distribution of sandstone-prone lowstands and
marginal-marine highstands in each of the 19 HFSs
identified within the Burgan Formation (Figure 12).
Furthermore, some additional new play concepts were
identified. Within the context of the new Mauddud–
Burgan boundary, distinct isolated incised-valley sys-
tems in the clastic member of the Mauddud Forma-
tion can be defined. Within the Burgan proper, the
possibility exists that marginal-marine shorelines in
the TSSs may also contain hydrocarbons off the flank
of structures. This combined structural-stratigraphic
trap would depend on mudstone-prone coastal-plain
deposits in each sequence acting as the updip lateral
seal and marine shales in the overlying sequence
acting as the top seal.
MAUDDUD FORMATION
A sequence-stratigraphic framework for the Maud-
dud Formation has been established using all avail-
able well and core data from Kuwait. The Mauddud
Formation can be described by the TSS and the HSS
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 231
FIGURE 15. Schematic facies model of the Mauddud transgressive sequence set (upper third-order composite
sequence).
as a third-order composite sequence that is composed
of eight high-frequency, depositional sequences. The
top (MAU100_TS) and base (B100_TS) of the Maud-
dud Formation are marked by TSS (Strohmenger et al.,
2002; Demko et al., 2003) (Figure 3).
The lower Mauddud (Mauddud TSS) shows a lat-
eral change in lithology from limestone in northern
Kuwait at the Raudhatain and Sabiriyah fields area to
siliciclastics in southern (Greater Burgan field area;
Figure 15) and southeastern Kuwait (Minagish field
area). The upper Mauddud (Mauddud HSS) is mostly
eroded in southern and southwestern Kuwait (Stroh-
menger et al., 2002; Demko et al., 2003) (Figure 16).
The Mauddud Formation in Kuwait comprises
eight carbonate lithofacies and three siliciclastic litho-
facies. Carbonate lithofacies were deposited in inner
to lower ramp, normal to slightly restricted environ-
ments. Clastic lithofacies were deposited in inner-
ramp (deeper lagoon), nearshore marine, and offshore
marine environments (Strohmenger et al., 2002; Demko
et al., 2003) (Figures 15, 16), similar to those de-
scribed for the Burgan Formation.
The Cenomanian Wara Formation overlies the
Mauddud Formation and provides a regional top seal
(Figure 3). It characteristically is a slightly calcareous
to noncalcareous marine shale in northern Kuwait
(Raudhatain and Sabiriyah fields area). In southern
(Greater Burgan field area) and southwestern Kuwait
(Minagish field area), glauconite-rich, deltaic sand-
stones are more common.
Lithofacies Types and Depositional Model
Sediments of the Mauddud Formation were de-
posited along a gently northward- and northeast-
ward-dipping homoclinal ramp (Strohmenger et al.,
2002; Demko et al., 2003).
Lithofacies and lithology of the Mauddud Forma-
tion of Kuwait vary by geographic area and time. In
the Mauddud transgressive sequence set, siliciclastic
deposits of southern (Greater Burgan field area) and
southwestern Kuwait (Minagish field area) grade into
carbonates toward the north and northeast (Abdali,
Raudhatain, and Sabiriyah fields area; Figure 15).
The Mauddud Formation can be described by
means of 11 lithofacies types (F1 – F11). The facies
scheme follows, in great parts, the one established by
I. Goodall, N. Cross, and D. Payne (1996, personal
communication), C. Hollis, N. Cross, T. Needham, and
B. Jones (1998, personal communication), and N. Cross,
R. Heath, and G. Paintal (1999, personal communi-
cation), with some modifications. Lithofacies types
range from moderate- to high-energy upper-ramp
232 / Strohmenger et al.
FIGURE 16. Schematic facies model of the Mauddud highstand sequence set (upper third-order composite sequence).
deposits (F1, F2, and F3) through medium-energy,
upper- to middle-ramp (F4), and lagoonal deposits
(F4, F5 and F6) to low-energy, lower-ramp (F7, F8,
and F11), and deeper lagoonal deposits (F7, F8, F9,
F10, and F11; Figures 15, 16).
The thickness of the Mauddud Formation increases
toward the north and northeast and decreases to-
ward the south and southwest. The thinning is the
result of reduced accommodation during the lower
Mauddud TSS (facies change from carbonates to silici-
clastics; Figure 15) and pronounced erosion of the
upper Mauddud HSS in southern (Greater Burgan
field area) and southwestern Kuwait (Minagish field
area) (Figure 16).
Lithofacies 1 (F1): Rudist Floatstone to Rudstone
Lithofacies type F1 (Figure 17A) is rich in rudists or
rudist fragments. Miliolids, conical orbitolinids,
other benthic foraminifera, skeletal fragments, and
echinoderms are common. Discoidal orbitolinids are
uncommon. Nonskeletal grains are peloids. Biotur-
bation is moderate.
This lithofacies type commonly grades into the
high-energy skeletal-peloidal grainstone (lithofacies
F2), as well as into the restricted lagoon nodular,
miliolid-bearing packstone-wackestone (lithofacies
F5). The grain composition (occurrence of miliolids),
as well as the facies-stacking pattern, suggests the
rudist floatstone-rudstone represents a ramp-crest to
inner-ramp, moderate to high-water-energy sediment,
deposited laterally or mostly lagoonward of lithofa-
cies F2 (skeletal-peloidal grainstone).
Relatively thick accumulations (as much as 60 ft
[18 m]) of rudist floatstone-rudstone occur within
Mauddud sequences MAU450 and MAU200. Espe-
cially within Mauddud sequence MAU450, rudist
floatstone-rudstone forms very porous intervals that
are easily identified by low gamma-ray- and high-
resistivity-log responses.
Cementation is minor. Dominant cement type is
blocky calcite cement, partly to completely filling the
molds of dissolved rudist shells.
Dominant porosity types are moldic and vuggy
porosity.
The average porosity is 18%, and the typical per-
meability is 5 md.
Lithofacies 2 (L2): Skeletal and Peloidal Grainstone
Lithofacies type F2 (Figure 17B) is rich in conical
orbitolinids, skeletal fragments, and echinoderms.
Other benthic foraminifera, red algae, and green algae
are common. Discoidal orbitolinids and gastropods
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 233

FIGURE 17. Carbonate lithofacies types, slabbed core photographs. (A) Lithofacies F1: rudist floatstone-rudstone.
(B) Lithofacies F2: skeletal-peloidal grainstone showing Glossifungites (Gl) burrows as well as fenestral structures
(keystone vugs). (C) Lithofacies F3: skeletal-peloidal mud-lean packstone showing bioturbation. (D) Lithofacies F4:
bioturbated skeletal-peloidal packstone.

are sparse. Nonskeletal grains are coated grains and
peloids. Bioturbation is low.
This lithofacies type frequently shows Glossifun-
gites burrows and/or, less common, desiccation cracks
(Figures 17B, 18A, B), as well as, also more uncom-
monly, karstification (Figure 18C). Sedimentary
structures indicating intertidal conditions, such as
trough – cross-bedding, fenestral structures (keystone
vugs and sheet cracks; Figure 17B), geopedal fillings
(vadose silt), and circumgranular cracks, are general-
ly limited to lithofacies type F2.
This lithofacies generally occurs at the top of the
Mauddud parasequences and HFSs. The lack of ma-
trix and the overall facies-stacking pattern suggest
the skeletal-peloidal grainstone to represent a high-
energy, shoal (ramp-crest) to upper-ramp deposits.
Cementation varies between minor and extensive.
Typical cement types are neomorphosed isopachous
rim cement, blocky calcite cement, and syntaxial cal-
cite cement. Glossifungites burrows as well as karst
fillings are commonly dolomitized.
Dominant porosity type is microporosity. Inter-
particle, intraparticle, and moldic porosity are minor
porosity types.
The average porosity is 15%, and the typical per-
meability is 2 md.
234 / Strohmenger et al.
Lithofacies 3 (F3): Skeletal and Peloidal
Mud-lean Packstone
Lithofacies type F3 (Figure 17C) is rich in conical
orbitolinids. Skeletal fragments and echinoderms are
common. Discoidal orbitolinids, other benthic fora-
minifera, and gastropods are sparse. Nonskeletal grains
are peloids. Bioturbation varies from moderate to
high.
This lithofacies type commonly alternates with
lithofacies type F2 (skeletal-peloidal grainstone). The
relatively low mud content and the overall facies-
stacking pattern suggests the skeletal-peloidal mud-
lean packstone to represent a moderate-energy, upper-
ramp deposit.
Like lithofacies type F2 (skeletal-peloidal grain-
stone), the skeletal-peloidal mud-lean packstone lo-
cally is karstified.
FIGURE 18. Sedimentary struc-
tures, slabbed core photographs.
(A) Erosional surface with Glossi-
fungites burrows. (B) Glossifungites
burrows and/or desiccation cracks
with younger sediment infill.
(C) Microkarst with sediment
infill.
Cementation generally is mi-
nor. Cement types are blocky cal-
cite and syntaxial calcite cements.
Dominant porosity type is mi-
croporosity. Intraparticle and
moldic porosity are minor po-
rosity types.
The average porosity is 15%,
and the typical permeability is
2 md.
Lithofacies 4 (F4): Bioturbated
Skeletal and Peloidal Packstone
Lithofacies type F4 (Figure 17D)
is rich in conical and discoidal
orbitolinids and echinoderms.
Skeletal fragments are common.
Other benthic foraminifera, ru-
dist fragments, and thin-shelled
bivalves are sparse. Nonskeletal
grains are peloids. Bioturbation
is high.
This lithofacies typically un-
derlies lithofacies type F3 (skeletal-
peloidal mud-lean packstone)
and, less commonly, lithofacies type F5 (nodular,
miliolid-bearing packstone-wackestone). It commonly
overlies lithofacies type F2 (skeletal-peloidal grain-
stone) and lithofacies type F3 (skeletal-peloidal mud-
lean packstone). Grain composition, texture, and the
overall facies-stacking pattern suggest that the bio-
turbated skeletal-peloidal packstone represents a low-
to moderate-energy, upper- to middle-ramp deposit. It
may also be present behind the high-energy bar de-
posits (lithofacies F1, F2, and F3) in an inner-ramp,
protected lagoonal environment.
Cementation is minor to extensive. Dominant ce-
ment type is blocky calcite cement.
Dominant porosity type is microporosity. Intrapar-
ticle and moldic porosity are minor porosity types.
The average porosity is 12%, and the typical per-
meability is 0.5 md.
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 235

FIGURE 19. Carbonate lithofacies types, slabbed core photographs. (A) Lithofacies F5: nodular, miliolid-bearing packstone-
wackestone. (B) Lithofacies F6: skeletal wackestone-mudstone (C) Lithofacies F7: clay-rich, bioturbated skeletal wacke-
stone showing Thalassinoides (Th) and Planolites (Pl) burrows. (D) Lithofacies F8: bioturbated glauconitic packstone.

Lithofacies 5 (F5): Nodular, Miliolid-bearing of miliolids), as well as the facies-stacking pattern,

Packstone to Wackestone
Lithofacies type F5 (Figure 19A) commonly con-
tains miliolids, conical orbitolinids, other benthic fo-
raminifera, rudist fragments, skeletal fragments, and
thin-shelled bivalves. Sponge spicules are common
and only occur within this lithofacies type. Discoidal
orbitolinids and echinoderms are uncommon. Non-
skeletal grains are peloids. Bioturbation and burrowing
is high, most probably causing its nodular appearance.
This lithofacies type is restricted to the upper
Mauddud (Mauddud HSS; Figure 16) and commonly
over- and underlies the rudist floatstone-rudstone
(lithofacies F1), predominantly within Mauddud se-
quence MAU200. The grain composition (occurrence
suggests the nodular, miliolid-bearing packstone-
wackestone to represent a low- to moderate-energy,
inner-ramp, restricted lagoonal deposit.
Cementation is minor, and the dominant cement
type is blocky calcite cement.
The dominant porosity type is microporosity. Intra-
particle and moldic porosity are minor porosity types.
The average porosity is 19%, and the typical per-
meability is 0.5 md.
Lithofacies 6 (F6): Skeletal Wackestone to Mudstone
Lithofacies type F6 (Figure 19B) is mud rich, light
gray, with sparse skeletal fragments and peloids. Biotur-
bation is very low. Pyrite framboids are quite frequent.
236 / Strohmenger et al.
This lithofacies type is present only in the upper
Mauddud HSS (Mauddud sequence MAU300; Figure 16).
We interpret this lithofacies type to represent a low-
energy, inner-ramp, protected lagoonal deposit.
Cementation (fine-crystalline, blocky calcite filling
the pore space of dissolved skeletal fragments) is minor.
The dominant porosity type is microporosity, and
intercrystalline porosity is present.
Average porosity and typical permeability (only four
samples analyzed) is within the range of lithofacies F5
(nodular, miliolid-bearing packstone-wackestone).
Lithofacies 7 (F7): Clay-rich, Bioturbated
Skeletal Wackestone
Lithofacies type F7 (Figure 19C) is rich in discoidal
orbitolinids and echinoderms, as well as in trace fossils,
including Thalassinoides and Planolites. Skeletal frag-
ments, thin-shelled bivalves, and pelagic foraminif-
era are common. Nonskeletal grains are peloids. Bio-
turbation is very high (solution-seam rich). Pyrite is
common.
This lithofacies typically underlies lithofacies type
F4 (bioturbated skeletal-peloidal packstone) or over-
lies lithofacies type F2 (skeletal-peloidal grainstone)
and lithofacies type F3 (skeletal-peloidal mud-lean
packstone). Grain composition, texture, and the over-
all facies-stacking pattern suggest the clay-rich, biotur-
bated skeletal wackestone to represent a low-energy,
middle- to lower-ramp deposit. It may, however, also
be present behind the high-energy bar deposits (litho-
facies F1, F2, and F3) in the inner-ramp, protected
lagoon environment.
Cementation is minor. The dominant cement type
is fine-crystalline, blocky calcite cement. Dolomiti-
zation occurs along solution seams.
The dominant porosity type is microporosity, and
intraparticle and intercrystalline porosity are minor
porosity types.
The average porosity is 4%, and the typical per-
meability is 0.01 md.
Lithofacies 8 (F8): Bioturbated
Glauconitic Packstone
Lithofacies type F8 (Figure 19D) commonly con-
tains discoidal orbitolinids, echinoderms, and skeletal
fragments. Conical orbitolinids, thin-shelled bivalves,
and rudist fragments are uncommon. Nonskeletal
grains are glauconite (glauconitized peloids), peloids,
quartz, and pyrite. Bioturbation is high.
This lithofacies type typically overlies sequence
boundaries on top of lithofacies types F2 (skeletal-
peloidal grainstone) and F3 (skeletal-peloidal mud-
lean packstone). We interpret this lithofacies type as
representing periods of low sedimentation rates during
rapid sea-level rise. It dominantly occurs within the
Mauddud TSS. The depositional environment ranges
from inner ramp, lagoonal (Mauddud TSS; Figure 15)
to middle to lower ramp (Mauddud HSS; Figure 16).
Cementation is minor to extensive. Dominant ce-
ment types are blocky calcite, ferroan dolomite, and
ferroan calcite cement. Ferroan cements are most com-
mon in southern Kuwait (Greater Burgan field area).
Dominant porosity type is microporosity. Intrapar-
ticle and moldic porosity are minor porosity types.
The average porosity is 7%, and the typical per-
meability is 0.05 md.
Lithofacies 9 (F9): Bioturbated
Glauconitic Sandstone
Lithofacies type F9 (Figure 20A) is quartz rich and
contains sparse discoidal orbitolinids and skeletal frag-
ments. It is rich in trace fossils, including Teichichnus,
Asterosoma and Terebellina (uncommon). Grain types
are detrital quartz, glauconite, pyrite, and siderite. Bio-
turbation is very high. This lithofacies is similar to
lithofacies 2D (glauconitic sandstone) of the Burgan
Formation but has a higher glauconite content.
Like lithofacies type F8 (bioturbated glauconitic
packstone), this lithofacies typically overlies sequence
boundaries on top of lithofacies types F2 (skeletal-
peloidal grainstone), F3 (skeletal-peloidal mud-lean
packstone), and F10 (bioturbated mud-rich sand-
stone). It dominantly occurs within the Mauddud
TSS (Figure 15). The bioturbated glauconitic sandstone
is interpreted to represent an inner-ramp, nearshore-
marine, lower-shoreface deposit.
Cementation is minor to extensive. Dominant ce-
ment types are blocky calcite, ferroan dolomite, and
ferroan calcite cement.
Predominant porosity type is microporosity, and in-
traparticle and intergranular porosity types are present.
The average porosity is 10%, and the typical per-
meability is 0.3 md.
Lithofacies 10 (F10): Bioturbated
Mud-rich Sandstone
Lithofacies type F10 (Figure 20B) is intensely bio-
turbated and rich in trace fossils, including Teichichnus,
Asterosoma, Thalassinoides, Planolites, and Chondrites.
Lithofacies type F10 occurs within the Mauddud
TSS (Figure 15). Relatively thick intercalations (>60 ft;
>18 m) occur within the Mauddud sequence MAU500.
This lithofacies is similar to lithofacies 3F (calcareous,
bioturbated sandstone) of the Burgan Formation.
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 237
The bioturbated mud-rich sandstone is interpreted to
represent an inner-ramp, nearshore-marine, lower-
shoreface deposit.
Cementation is minor; dominant porosity type is
intergranular.
The average porosity is 12%, and typical perme-
ability is 11 md.
Lithofacies 11 (F11): Dark-gray Shale
and Mudstone
Lithofacies F11 (Figure 20C) shows only sparse
skeletal fragments, thin-shelled bivalves, planktonic
foraminifera, and nannofossils. This lithofacies is sim-
ilar to lithofacies 5A (laminated gray shale) of the
Burgan Formation. Millimeter-scale laminae indi-
cate low bioturbation caused by stressed (lagoon) or
deeper marine, offshore (lower-ramp) environment
of deposition.
Dominant lithofacies type of the Mauddud trans-
gressive sequence set (Figure 15); deposited largely in
an inner-ramp, deeper lagoonal setting, juxtaposed
to lithofacies types F7 (clay-rich, bioturbated wacke-
stone), F8 (bioturbated glauconitic packstone), and F9
FIGURE 20. Siliciclastic lithofa-
cies types, slabbed core pho-
tographs. (A) Lithofacies F9:
bioturbated glauconitic sand-
stone showing Teichichnus (Te)
burrows. (B) Lithofacies F10:
bioturbated mud-rich sand-
stone showing Teichichnus (Te)
and Thalassinoides (Th) bur-
rows. (C) Lithofacies F11: dark-
gray shale and mudstone.
(bioturbated glauconitic sand-
stone; Figure 15). The fact that
it is not highly bioturbated as
it would be expected for off-
shore shales underlying low-
er shoreface deposits (litho-
facies F9 and F10) indicates
a high-stress environment.
The dark-gray shale and mud-
stone are interpreted to rep-
resent a low-energy, deeper
marine, lower-ramp to basin-
al deposit during the Maud-
dud HSS (Figure 16) and the
Wara Formation.
Predominant porosity type
is microporosity.
The average porosity is 5%, and typical perme-
ability is 0.03 md.
High-frequency Sequences
Eight high-frequency, depositional sequences were
interpreted within the Mauddud Formation. From bot-
tom to top, these are B100 (TST and HST), MAU600,
MAU500, MAU450, MAU400, MAU350, MAU300,
and MAU200 (Figure 3).
Eight SBs, two TSs, three MFSs, and two FSs were
identified. All surfaces, except for one flooding sur-
face (MAU600_FS), can be correlated throughout all
of Kuwait.
Sequence boundaries typically are characterized by
an abrupt change in texture and lithofacies. Glossi-
fungites burrows and/or desiccation cracks are common
in the upper part of each high-frequency sequence
(Figures 18A, B; 21). Karstification is seen locally
(Figure 18C). Evidence of at least periodic exposure,
such as fenestral structures (keystone vugs and sheet
cracks; Figure 17B) and circumgranular cracks, is also
present in the upper part of each high-frequency se-
quence. Sandstone or packstone, rich in glauconite,
238 / Strohmenger et al.

FIGURE 21. Mauddud Formation idealized carbonate parasequence (upper part of a high-frequency sequence)
showing shallowing-upward trend of facies from base to top: bioturbated skeletal-peloidal packstone (F4, blue
color), skeletal-peloidal mud-lean packstone (F3, red color), and skeletal-peloidal grainstone (F2, orange color). Flooding
surface/sequence boundary (FS/SB) is interpreted by Glossifugites burrows or, uncommonly, karstification (microkarst)
and is overlain by bioturbated glauconitic packstone (F8) or bioturbated glauconitic sandstone (F9, green color).

typically immediately overlies each high-frequency
sequence boundary (Figure 21), recording periods of
rapid flooding and resulting in low sedimentation rate.
The identified high-frequency sequence boundaries
correspond to combined flooding surface and se-
quence boundaries (FS/SB, Figure 21).
Maximum flooding surfaces are interpreted at major
marine incursions marked by a change from retrograd-
ing to prograding stacking of marine parasequences.
Flooding surfaces are interpreted on top of shallowing-
upward parasequences, commonly corresponding to
Glossifungites surfaces (Figure 21).
Mauddud Sequence B100
(B100_SB to MAU600_SB)
In northern Kuwait (Raudhatain and Sabiriyah
fields area; Figure 22), this sequence is shale and
mud rich at the base of the TST, grading upward, to
grain rich at the top of the HST. Dominantly tidal-
flat deposits occur in southern Kuwait (Greater
Burgan field area; Figure 23), and dominantly coastal-
plain deposits occur in southwestern Kuwait (Min-
agish field area; Figure 23). The sequence is partly
eroded by Mauddud sequence boundary MAU600_SB
in southern Kuwait (Figure 23).
The sequence consists of lithofacies F11 (dark-gray
shale and mudstone) and lithofacies F9 (bioturbated
glauconitic sandstone) grading upward into lithofa-
cies F7 (clay-rich, bioturbated skeletal wackestone),
lithofacies F8 (bioturbated glauconitic packstone),
and lithofacies F4 (bioturbated skeletal-peloidal pack-
stone) in northern Kuwait (Raudhatain and Sabiriyah
fields area; Figures 22, 23).
Mauddud Sequence MAU600
(MAU600_SB to MAU500_SB)
This sequence is composed of mud-dominated car-
bonates in northern Kuwait (Raudhatain and Sabir-
iyah fields area; Figure 22), dominantly coastal-plain
and tidal-influenced deposits in southern Kuwait
FIGURE 22. Field-scale cross section (northwest – southeast) showing facies distribution and sequence-stratigraphic framework of the Mauddud Formation
throughout the northern Kuwait Raudhatain and Sabiriyah fields. Mauddud transgressive sequence set: B100_TS to MAU400_MFS. Mauddud highstand sequence
set: MAU400_MFS to MAU100_SB. Mauddud lowstand systems tract: MAU100_SB to MAU100_TS. GR = gamma-ray log; FAC = Mauddud carbonate and siliciclastic facies
and Burgan deposits (based on core); MD = measured depth (feet); DT = sonic log; NEU = neutron porosity log; DENS = density log.
Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 239
 240 / Strohmenger et al.

FIGURE 23. Regional cross section oriented north – south and east – west showing facies distribution and sequence-stratigraphic framework of the Mauddud
Formation. Note that chronostratigraphic boundaries (time lines) crosscut the lithostratigraphic boundary between the Mauddud carbonate member (Mauddud
Formation) and the Mauddud clastic member (Burgan Formation, time equivalent to Mauddud Formation). The proposed sequence-stratigraphic correlation
is supported by age dating and palynofacies analyses (T. D. Davies and T. C. Huang, 2000, personal communication). Stratigraphy-diagnostic palynofacies
assemblages are shown as colored dots. Blue dots: only found above Mauddud transgressive surface MAU100_TS (in Wara Formation). Yellow dot: only found between
Mauddud sequence boundary MAU100_SB and Mauddud trangressive surface MAU100_TS (Mauddud lowstand systems tract, onlapping on Mauddud sequence
boundary MAU100_SB in southern and southwestern Kuwait). Green dots: only found between Burgan transgressive surface B100_TS and Mauddud maximum
flooding surface MAU400_MFS (Mauddud transgressive sequence set: lower part of Mauddud carbonate member in northern Kuwait and Mauddud clastic member in
southern and southwestern Kuwait). Red dots: only found below Burgan transgressive surface B100_TS (Burgan lowstand sequence set). No stratigraphy-diagnostic
palynofacies assemblage was found in grain-dominated, shallow-water carbonates between Mauddud maximum flooding surface MAU400_MFS and Mauddud
sequence boundary MAU100_SB (Mauddud highstand sequence set). Color codes for Mauddud carbonate and siliciclastic facies as well as for Burgan siliciclastic
deposits are shown in Figure 22. GR = gamma-ray log; FAC = Mauddud carbonate and siliciclastic facies and Burgan deposits (based on core); MD = measured depth
(feet); DT = sonic log; NEU = neutron porosity log; DENS = density log.
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 241
(Greater Burgan field area; Figure 23), and dominantly
coastal-plain and tidal-flat deposits in southwestern
Kuwait (Minagish field area; Figure 23).
Dominantly lithofacies F11 (dark gray shale and
mudstone) and minor lithofacies F8 (bioturbated
glauconitic sandstone) and lithofacies F7 (clay-rich,
bioturbated skeletal wackestone) occur in the north-
ern fields area. The succession may show upward in-
crease of porosity and permeability and shallowing-
upward lithofacies trend, as well as coarsening-upward
texture in the northern fields area (Figures 21, 22).
Mauddud Sequence MAU500
(MAU500_SB to MAU450_SB)
Marginal-marine (distal lower shoreface) siliciclas-
tics occur in northern Kuwait (Raudhatain and Sabi-
riyah fields area) and were not eroded by sequence
boundary MAU450_SB (Figure 22). Dominantly coastal-
plain deposits occur in southern Kuwait (Greater Burgan
field area; Figure 23) and dominantly tidal-influenced
to fluvial deposits occur in southwestern Kuwait (Mi-
nagish field area; Figure 23).
Mauddud Sequence MAU450
(MAU450_SB to MAU400_SB)
This sequence is composed of grain-dominated
carbonates in northern Kuwait (Raudhatain and
Sabiriyah fields area; Figure 22) and dominated by
marginal-marine (distal lower shoreface) deposits in
southern (Greater Burgan field area; Figure 23) and
southwestern Kuwait (Minagish field area; Figure 23).
The sequence is composed predominantly of litho-
facies F2 (skeletal-peloidal grainstone), with minor
lithofacies F3 (skeletal-peloidal mud-lean packstone)
and lithofacies F4 (bioturbated skeletal-peloidal pack-
stone). Relatively thick accumulations (as much as
60 ft [18 m]) of highly porous lithofacies F1 (rudist
floatstone-rudstone) occur as aerially restricted inter-
calations throughout the northern fields area. This
sequence may show an upward increase of porosity
and permeability and a shallowing-upward lithofacies
trend, as well as coarsening-upward texture (northern
fields area; Figures 21, 22).
Mauddud Sequence MAU400
(MAU400_SB to MAU350_SB)
This sequence is composed of grain-dominated car-
bonates in northern Kuwait (Raudhatain and Sabiri-
yah fields area), with lithofacies 8 (bioturbated glau-
conitic packstone) commonly overlying Mauddud
sequence boundary MAU400_SB (Figure 22). Litho-
facies F8 (bioturbated glauconitic packstone) overlies
Mauddud sequence boundary MAU400_SB in south-
ern Kuwait (Greater Burgan field area; Figure 23).
Lithofacies F9 (bioturbated glauconitic sandstone)
overlies Mauddud sequence boundary MAU400_SB in
southwestern Kuwait (Minagish field area; Figure 23).
This sequence is dominated by lithofacies F4 (bio-
turbated skeletal-peloidal packstone) in southern
Kuwait (Greater Burgan field area; Figure 23) and
dominated by marginal-marine siliciclastics with
thin, karstified lithofacies F3 (skeletal-peloidal mud-
lean packstone) at the top in southwestern Kuwait
(Minagish field area; Figure 23). It is partly erod-
ed in southwestern Kuwait (Minagish field area;
Figure 23).
The sequence is composed predominantly of litho-
facies F2 (skeletal-peloidal grainstone) and litho-
facies F3 (skeletal-peloidal mud-lean packstone), as
well as minor lithofacies F4 (bioturbated skeletal-
peloidal packstone) and lithofacies F7 (clay-rich, bio-
turbated skeletal wackestone) in the northern fields
area (Figures 22, 23).
Mauddud Sequence MAU350
(MAU350_SB to MAU300_SB)
This sequence is composed predominantly of litho-
facies F2 (skeletal-peloidal grainstone) and litho-
facies F3 (skeletal-peloidal mud-lean packstone), as
well as minor lithofacies F4 (bioturbated skeletal-
peloidal packstone) and intercalations of lithofacies F1
(rudist floatstone-rudstone) in northern Kuwait (Raud-
hatain and Sabiriyah fields area; Figure 22). It is partly
eroded in the Greater Burgan field area (lithofacies F4:
skeletal-peloidal packstone; Figure 23) and eroded
(not present) in southwestern Kuwait (Minagish field
area; Figure 23).
This sequence may show an upward increase of
porosity and permeability and shallowing-upward
lithofacies trend as well as coarsening-upward tex-
ture in the northern fields area (Figures 21, 22).
Mauddud Sequence MAU300
(MAU300_SB to MAU200_SB)
This sequence is composed predominantly of litho-
facies F2 (skeletal-peloidal grainstone) and lithofa-
cies F3 (skeletal-peloidal mud-lean packstone), as
well as minor lithofacies F4 (bioturbated skeletal-
peloidal packstone) and lithofacies F1 (rudist float-
stone-rudstone) in northern Kuwait (Raudhatain and
Sabiriyah fields area; Figure 22). The sequence is eroded
(not present) in southern (Greater Burgan field area;
Figure 23) and southwestern Kuwait (Minagish field
area; Figure 23).
242 / Strohmenger et al.
Mauddud Sequence MAU200
(MAU200_SB to MAU100_SB)
This sequence is composed predominantly of
lithofacies F5 (nodular, miliolid-bearing packstone-
wackestone) and lithofacies F3 (skeletal-peloidal
mud-lean packstone), as well as minor lithofacies F4
(bioturbated skeletal-peloidal packstone) and litho-
facies F1 (rudist floatstone-rudstone) in northern Ku-
wait (Raudhatain and Sabiriyah fields area; Figure 22).
The sequence is eroded (not present) in southern
(Greater Burgan field area; Figure 23) and southwest-
ern Kuwait (Minagish field area; Figure 23).
Mauddud Interval MAU100_SB to MAU100_TS
This interval is composed predominantly of litho-
facies F11 (gray shale and mudstones) and lithofacies
F7 (clay-rich, bioturbated skeletal wackestone).
The interval was deposited during the third-order
relative sea-level rise above Mauddud sequence bound-
ary MAU100_SB (Figures 22, 23) and is not present
in southern (Greater Burgan field area; Figure 23) and
southwestern Kuwait (Minagish field area; Figure 23).
Composite Sequence
The TSS of the Mauddud composite sequence starts
at the base of Mauddud transgressive surface B100_TS
and is bounded on the top by Mauddud maximum
flooding surface MAU400_MFS of the Mauddud high-
frequency sequence MAU400. The overlying HSS
is bounded on top by Mauddud sequence boundary
MAU100_SB. The top of the Mauddud Formation is
marked by the transgressive surface MAU100_TS
(Figures 22, 23) that merges with the Mauddud se-
quence boundary MAU100_SB toward southern and
southwestern Kuwait (Figure 23).
The Mauddud TSS shows a lateral change in litho-
logy from limestone in northern Kuwait (Raudhatain
and Sabiriyah fields area; Figure 22) to siliciclastics in
southern (Greater Burgan field area; Figure 23) and
southwestern Kuwait (Minagish field area; Figure 23).
An overall shoaling-upward or progradational sig-
nature characterizes the Mauddud HSS. The HSS is
interpreted to be eroded down to and below sequence
boundary MAU350_SB in southern (Greater Burgan
field area; Figure 23) and southwestern Kuwait (Mi-
nagish field area; Figure 23).
Our sequence-stratigraphic interpretation of the
Mauddud Formation contrasts with the previously
used lithostratigraphic correlation. Lithostratigraphic
correlation places the boundary between the Maud-
dud Formation and the underlying Burgan Forma-
tion at the base of the deepest limestone bed. In con-
trast, the sequence-stratigraphic boundary between the
Mauddud Formation and the underlying Burgan For-
mation occurs approximately 100 ft (30 m) deeper in
the clastic section of the Mauddud Formation. The
lithostratigraphic boundary between the Mauddud
carbonate member and the Mauddud clastic member
is highly diachronous (Figure 23). The thickness rela-
tionships of the two Mauddud Members are recipro-
cal. The Mauddud carbonate member becomes very
thin in locations where the Mauddud clastic member
is thick. Toward the north (Raudhatain and Sabiriyah
fields area), the Mauddud clastic member becomes
very thin to absent, whereas the Mauddud carbonate
member is thinning toward southern Kuwait (Great-
er Burgan field area; Figure 23) and southwestern
Kuwait (Minagish field area; Figure 23).
The sequence-stratigraphic correlation of the
Upper Burgan and the Mauddud presented here is
supported by biostratigraphic and palynofacies data
(T. D. Davies and T.-C. Huang, 2000, personal com-
munication) (Figure 23).
Reservoir Quality Distribution
In general, most Mauddud lithofacies have mod-
erate porosity and overall low permeability. Litho-
facies types F2, F3, F4, F5, and F6 have essentially the
same porosity-permeability distributions. Only litho-
facies type F1 (rudist floatstone-rudstone) and litho-
facies F10 (bioturbated, mud-rich sandstone) show
higher porosity and/or permeability values. Intervals
that show enhanced reservoir quality are related to
fracturing and faulting (enhanced permeability), the
occurrence of microkarst (Figure 18C), the occurrence
of rudist floatstone-rudstone (lithofacies F1: Maud-
dud high-frequency sequence MAU450; Figure 17A),
or the occurrence of the bioturbated, mud-rich sand-
stone (lithofacies F10: Mauddud high-frequency se-
quence MAU500; Figure 20B).
CONCLUSIONS
Sequence-stratigraphic and biostratigraphic analy-
ses indicate that the traditional lithostratigraphic
Burgan-Mauddud contact is actually a time-transgres-
sive boundary (Figure 23). The uppermost Burgan and
overlying Mauddud formations form a third-order
composite sequence that becomes more siliciclastic
prone toward the south (in the Greater Burgan field
area) and toward the southwest (in the Minagish field
area; Figure 23) and carbonate prone to the north
and northeast of Kuwait near the Raudhatain and
Sabiriyah fields (Figures 22, 23). A lower, third-order
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 243
composite sequence encompasses the remainder of
the underlying Burgan Formation (Figure 12). It com-
prises marginal-marine deposits to the northeast
and nonmarine deposits to the southwest of Kuwait
(Figure 12).
Within the Burgan Formation, 19 high-frequency,
depositional sequences are recognized and stack to
form one third-order composite sequence and the
LSS of a younger third-order composite sequence
(Figures 3, 12). Lowstand systems tracts consist pri-
marily of fluvial, incised-valley-fill deposits in southern
and western Kuwait that become thinner, laterally
more discontinuous, and more tidal-influenced in a
downdip direction to the northeast (Figure 13). The
TSTs display a systematic updip to downdip change
from mud-prone alluvial- and coastal-plain strata to
marginal-marine mudstones and sandstones to ma-
rine carbonates. The HSTs are dominated by updip
alluvial- and coastal-plain mudstones and sandstones
and marginal-marine sandstones and mudstones
downdip (Figure 14).
Facies analysis of the Burgan Formation reveals the
presence of 22 lithofacies, which are delineated based
on unique physical criteria, including grain size, com-
position, sorting, grading, and physical and biogenic
sedimentary structures. These lithofacies describe six
distinct facies, which form the basic building blocks
of two facies associations (Figure 4).
The primary reservoirs in the Burgan Formation
are fluvial and tidal deposits of the incised-valley-fill
facies association (Figure 4). Minor potential reser-
voirs are shoreline sandstones of the TSTs and HSTs
(Figure 4). In addition, the possibility exists that
marginal-marine shorelines in the TSSs may contain
hydrocarbons off the flanks of present-day structures.
This combination of structural and stratigraphic trap
would be dependent on mud-prone coastal-plain sedi-
ments in each sequence acting as the updip lateral
seal and marine shales in the overlying sequence
serving as a (very high-risk) top seal.
The Mauddud Formation is sequence stratigraphi-
cally subdivided into eight high-frequency deposi-
tional sequences (Figures 3, 22, 23). The identified
high-frequency SBs, TSs, MFSs, and FS can be cor-
related regionally throughout Kuwait. The Mauddud
transgressive sequence set displays a lateral change in
lithology from limestone in northern Kuwait (Raud-
hatain and Sabiriyah fields area; Figure 22) to silici-
clastics in southern (Greater Burgan field area) and
southwestern Kuwait (Minagish field area; Figures 15,
23), interfingering with what has traditionally been
recognized as the Burgan Formation (Figures 3, 23).
The Mauddud highstand sequence set is carbonate
prone and is mostly eroded in southern and south-
western Kuwait, resulting in a significant southward
and southwestward thinning (Figures 16, 23).
A total of eight carbonate lithofacies and three
siliciclastic lithofacies types were identified within
the Mauddud Formation. Carbonate lithofacies were
deposited in inner- to lower-ramp, normal- to slightly
restricted-marine environments. Siliclastic lithofa-
cies were deposited in a fairly wide range of environ-
ments, including deeper lagoon, nearshore-marine,
and offshore-marine (Mauddud carbonate member),
as well as coastal-plain, fluvial, and tidal, (Mauddud
clastic member) environments.
In general, most Mauddud lithofacies have mod-
erate porosity and low permeability, with micropo-
rosity as the dominant pore type. Intervals with en-
hanced reservoir quality can be related to fracturing
and faulting, the occurrence of microkarst (Figure 18C),
the presence of rudist floatstone-rudstone (Figure 17A),
and the distribution of intercalated mud-rich sand-
stone (Figure 18A). Within the context of the new
Mauddud – Burgan chronostratigraphic boundary,
distinct isolated incised-valley systems can be defined
within the Mauddud clastic member. These IVF are
expected to have reservoir characteristics similar to
those of the underlying Burgan Formation.
The proposed sequence-stratigraphic framework
and the sequence-stratigraphy-keyed facies scheme
result in a predictable distribution of reservoir and
seal facies and allow for a better prediction of the
vertical and lateral distribution of reservoir quality
and reservoir continuity at both field scale and re-
gional scale.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the manage-
ments of ExxonMobil Exploration Company, Exxon-
Mobil Upstream Research Company, and Kuwait Oil
Company (KOC) for their permission to publish this
paper. For valuable discussions, we thank Linda W.
Corwin (ExxonMobil), Daniel H. Cassiani (Exxon-
Mobil), Kathleen M. McManus (ExxonMobil), Menahi
Al-Anzi (KOC), Ahmed Al-Eidan (KOC), and Moham-
med Al-Ajmi (KOC). We extend our thanks to David
Awwiller (ExxonMobil) for his detailed reservoir-
quality analyses and to Tom D. Davies and Ting-
Chang Huang for performing the biostratigraphic
and palynofacies analyses. Dolores A. Claxton (Exxon-
Mobil) is thanked for drafting the figures. We extend
special thanks to Arthur D. Donovan (BP; formerly
244 / Strohmenger et al.
with ExxonMobil) for his valuable contributions to
this study. The authors greatly appreciate the thor-
ough and thoughtful reviews of J. R. (Rick) Sarg and
James McGovney.
REFERENCES CITED
Adasani, M., 1965, The Greater Burgan field: Fifth Arab
Petroleum Congress, Cairo, p. 7 – 27.
Adasani, M., 1967, The northern Kuwait oil fields: Sixth
Arab Petroleum Congress, Baghdad, p. 1 – 39.
Al-Anzi, M., 1995, Stratigraphy and structure of the Bahra
field Kuwait, in M. I. Al-Husseini, ed., The Middle East
petroleum geosciences (Geo94): Selected Middle East
papers from the Middle East Geoscience Conference:
Bahrain, Gulf PetroLink, v. 1, p. 53 – 64.
Al-Eidan, A. J., W. B. Wethington, and R. B. Davies, 2001,
Upper Burgan reservoir distribution, northern Kuwait:
Impact on reservoir development: GeoArabia, v. 6,
no. 2, p. 179 – 208.
Al-Rawi, M. M., 1981, Geological interpretation of the oil
entrapment in the Zubair Formation, Raudhatain field:
Society of Petroleum Engineers, Middle East Oil Tech-
nical Conference, Bahrain, SPE Paper 9591, p. 149–159.
Alsharhan, A. S., and A. E. M. Nairn, 1993, Carbonate plat-
form models of Arabian Cretaceous reservoirs, in J. A. T.
Simo, R. W. Scott, and J.-P. Masse, eds., Cretaceous
carbonate platforms: AAPG Memoir 56, p. 173 – 184.
Alsharhan, A. S., and A. E. M. Nairn, 1997, Sedimentary
basins and petroleum geology of the Middle East:
Amsterdam, Elsevier, 843 p.
Bou-Rabee, F., 1996, Geologic and tectonic history of
Kuwait as inferred from seismic data: Journal of Pe-
troleum Science and Engineering, v. 16, p. 151 – 167.
Brennan, P., 1990a, Raudhatain field, Kuwait, Arabian
basin, in E. A. Beaumont and N. H. Foster, eds., AAPG
Treatise of Petroleum Geology, Atlas of Oil and Gas
Fields: Structural Traps 1, p. 187 – 210.
Brennan, P., 1990b, Greater Burgan field, Kuwait, Arabian
basin, in E. A. Beaumont and N. H. Foster, eds., AAPG
Treatise of Petroleum Geology, Atlas of Oil and Gas
Fields: Structural Traps 1, p. 103 – 128.
Carman, G. J., 1996, Structural elements of onshore Kuwait:
GeoArabia, v. 1, no. 2, p. 239 – 266.
Christian, L., 1997, Cretaceous subsurface geology of the
Middle East region: GeoArabia, v. 2, no. 3, p. 239 – 256.
Demko, T. M., P. E. Patterson, H. R. Feldman, C. J. Stroh-
menger, J. C. Mitchell, P. J. Lehmann, G. Alsahlan,
H. Al-Enezi, and M. Al-Anezi, 2003, Sequence stratig-
raphy and reservoir architecture of the Burgan and
Mauddud formations (Lower Cretaceous), Kuwait (abs.):
AAPG Annual Meeting Program, v. 5, p. A39.
Fox, A. F., 1961, The development of the southeastern
Kuwait oil fields: Institute of Petroleum Review, v. 15,
no. 180, p. 373 – 379.
Haq, B. U., 1991, Sequence stratigraphy, sea-level change,
and significance for the deep sea, in D. I. M. Mac-
donald, ed., Sedimentation, tectonics and eustasy: Sea-
level changes at active margins: International Asso-
ciation of Sedimentologists Special Publication 12,
p. 3 – 39.
Haq, B. U., J. Hardenbol, and P. R. Vail, 1987, Chronology
of fluctuating sea levels since the Triassic: Science,
v. 235, p. 1156 – 1167.
Haq, B. U., J. Hardenbol, and P. R. Vail, 1988, Mesozoic
and Cenozoic chronostratigraphy and cycles of sea-
level change, in C. K. Wilgus, B. S. Hastings, C. G. St. C.
Kendall, H. W. Posamentier, C. A. Ross, and J. C. Van
Wagoner, eds., Sea-level changes: An integrated
approach: SEPM Special Publication 42, p. 71 – 108.
Kaufman, R. L., H. Dashti, C. S. Kabir, J. M. Pederson, M. S.
Moon, R. Quttainah, and H. Al-Wael, 1997, Charac-
terizing the Greater Burgan field: Use of geochemistry
and oil fingerprinting: Society of Petroleum Engineers
10th SPE Middle East Oil Show and Conference,
Bahrain, SPE Paper 37803, p. 385 – 394.
Milton, D. I., and C. C. S. Davies, 1965, Exploration and
development of the Raudhatain field: Journal of the
Institute of Petroleum, v. 51, no. 493, p. 17 – 28.
Mitchum Jr., R. M., 1977, Seismic stratigraphy and global
changes in sea level: Part 11 — Glossary of terms used
in seismic stratigraphy, in C. E. Payton, ed., Seismic
stratigraphy — Applications to hydrocarbon explora-
tion: AAPG Memoir 26, p. 205 – 212.
Mitchum Jr., R. M., and J. C. Van Wagoner, 1991, High-
frequency sequences and their stacking patterns:
Sequence-stratigraphic evidence of high-frequency
eustatic cycles, in K. T. Biddle and W. Schlager, eds.,
The record of sea-level fluctuations: Sedimentary Ge-
ology, v. 70, p. 131 – 160.
Mitchum Jr., R. M., P. R. Vail, and S. Thompson III, 1977,
Seismic stratigraphy and global changes of sea level:
Part 2 — The depositional sequence as a basic unit for
stratigraphic analysis, in C. E. Payton, ed., Seismic
stratigraphy — Applications to hydrocarbon explora-
tion: AAPG Memoir 26, p. 53 – 62.
Pemberton, G. S., J. A. MacEachern, and R. W. Frey, 1992a,
Trace fossil models; environmental and allostrati-
graphic significance, in R. G. Walker and J. P. Noel,
eds., Facies models; response to sea level change: Geo-
logical Association of Canada, p. 47 – 72.
Pemberton, G. S., J. C. Van Wagoner, and G. D. Wach,
1992b, Ichnofacies of a wave-dominated shoreline:
SEPM Core Workshop 17, p. 339 – 382.
Sarg, J. F., 1988, Carbonate sequence stratigraphy, in C. K.
Wilgus, B. S. Hastings, C. G. St. C. Kendall, H. W.
Posamentier, C. A. Ross, and J. C. Van Wagoner, eds.,
Sea-level changes: An integrated approach: SEPM
Special Publication 42, p. 155 – 181.
Sarg, J. F., J. R. Markello, and L. J. Weber, 1999, The second-
order cycle, carbonate-platform growth, and reservoir,
source, and trap prediction, in P. M. Harris, A. H. Saller,
and J. A. T. Simo, eds., Advances in carbonate sequence
stratigraphy: Application to reservoirs, outcrops and
models: SEPM Special Publication 63, p. 11 – 34.
 Sequence Stratigraphy and Reservoir Architecture of Burgan and Mauddud Formations / 245
Strohmenger, C. J., T. M. Demko, J. C. Mitchell, P. J.
Lehmann, H. R. Feldman, A. Douban, A. J. Al-Eidan,
G. Alsahlan, and H. Al-Enezi, 2002, Regional sequence
stratigraphic framework for the Burgan and Mauddud
formations (Lower Cretaceous, Kuwait): Implications
for reservoir distribution and quality (abs.): GeoArabia,
v. 7 , no. 2, p. 304.
Vail, P. R., 1987, Seismic stratigraphy interpretation using
sequence stratigraphy: Part 1 — Seismic stratigraphy
interpretation procedure, in A. W. Bally, ed., Atlas of
seismic stratigraphy, v. 1: AAPG Studies in Geology 27,
p. 1 – 10.
Vail, P. R., R. M. Mitchum Jr., R. G. Todd, J. M. Widmier,
S. Thompson III, J. B. Sangree, J. N. Bubb, and W. G.
Hatlelid, 1977, Seismic stratigraphy and global changes
of sea level: Part 1— Overview, in C. E. Payton, ed.,
Seismic stratigraphy— Applications to hydrocarbon ex-
ploration: AAPG Memoir 26, p. 49– 212.
Vail, P. R., F. Audemard, S. A. Bowman, P. N. Eisner, and
C. Perez-Cruz, 1991, The stratigraphic signatures of tec-
tonics, eustasy and sedimentology — An overview, in
G. Einsele, W. Ricken, and A. Seilacher, eds., Cycles and
events in stratigraphy: Berlin, Springer, p. 617 – 659.
Van Wagoner, J. C., R. M. Mitchum Jr., H. W. Posamentier,
and P. R. Vail, 1987, Seismic stratigraphy interpreta-
tion using sequence stratigraphy: Part 2 — Key defini-
tions of sequence stratigraphy, in A. W. Bally, ed., Atlas
of seismic stratigraphy, v. 1: AAPG Studies in Geology 27,
p. 11–14.
Van Wagoner, J. C., H. W. Posamentier, R. M. Mitchum Jr.,
P. R. Vail, J. F. Sarg, T. S. Loutit, and J. Hardenbol, 1988,
An overview of the fundamentals of sequence stratig-
raphy and key definitions, in C. K. Wilgus, B. S. Hastings,
C. G. St. C. Kendall, H. W. Posamentier, C. A. Ross,
and J. C. Van Wagoner, eds., Sea-level changes: An in-
tegrated approach: SEPM Special Publication 42, p. 39 –
46.
Van Wagoner, J. C., R. M. Mitchum, K. M. Campion, and
V. D. Rahmanian, 1990, Siliciclastic sequence stratig-
raphy in well logs, cores, and outcrop: AAPG Methods
in Exploration Series 7, 55 p.
Walker, R. G., and A. G. Plint, 1992, Wave- and storm-
dominated shallow marine systems, in R. G. Walker
and N. P. James, eds., Facies models — Response to
sea-level change: Geological Association of Canada,
p. 219 – 238.
 7
Dull, D. W., R. A. Garber, and W. S. Meddaugh, 2006, The sequence
stratigraphy of the Maastrichtian (Upper Cretaceous) reservoir at Wafra
field, Partitioned Neutral Zone, Saudi Arabia and Kuwait: Key to reservoir
modeling and assessment, in P. M. Harris and L. J. Weber, eds., Giant
hydrocarbon reservoirs of the world: From rocks to reservoir characteriza-
tion and modeling: AAPG Memoir 88/SEPM Special Publication, p. 247 –279.

The Sequence Stratigraphy of the

Maastrichtian (Upper Cretaceous)
Reservoir at Wafra Field, Partitioned
Neutral Zone, Saudi Arabia and
Kuwait: Key to Reservoir Modeling
and Assessment
Dennis W. Dull
Chevron Energy Technology Company, Houston, Texas, U.S.A.

Raymond A. Garber
Chevron Energy Technology Company, Houston, Texas, U.S.A.

W. Scott Meddaugh
Chevron Energy Technology Company, Houston, Texas, U.S.A.

ABSTRACT

T
he Maastrichtian (Upper Cretaceous) reservoir is one of five prolific oil res-
ervoirs in the giant Wafra oil field. Although discovered and first produced
in 1959, the reservoir is currently in early development because of its low
but variable oil gravity, high sulfur content, relatively high water cut, and appar-
ent compartmentalization. This made it a much less attractive resource than other
productive intervals at Wafra field. Less than 1% of the original oil in place in the
Maastrichtian has been produced.
The Maastrichtian oil production is largely from subtidal dolomites at an
average depth of 760 m (2500 ft). Carbonate deposition occurred on a very gently
dipping, shallow, arid, and restricted ramp setting that transitioned between
normal-marine conditions to restricted lagoonal environments. The average po-
rosity of the reservoir interval is about 15%, although productive zones have po-
rosity values as much as 30–45%. The average permeability of the reservoir in-
terval is about 30 md; individual core plugs have measured permeability as much

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215879M883030

247
248 / Dull et al.

as 1200 md. This study was undertaken to (1) determine reservoir volumetrics,
(2) understand the areal and stratigraphic distribution of intervals likely to yield
higher volumes of better quality oil, and (3) provide a reservoir property model for
use in fluid-flow simulation.
The key to modeling the reservoir was the construction of an appropriately
detailed sequence-stratigraphic framework for use in building a geostatistical res-
ervoir model with high-quality descriptions from five cored wells in the reservoir.
The geostatistical model of the Maastrichtian reservoir demonstrates the
layered and compartmentalized nature of the reservoir and clearly shows that
the location of the reservoir facies in the Maastrichtian is controlled by the
original depositional fabric and subsequent dolomitization, both of which have
been influenced by the paleotopography. Such understanding is critical to ef-
ficiently develop the 1.5 billion bbl of Maastrichtian oil at Wafra field.

INTRODUCTION
The Wafra field is located in the Partitioned Neutral
Zone (PNZ) between Kuwait and Saudi Arabia. Figure 1
shows the location of the major oil fields in the PNZ.
Figure 2 provides a generalized stratigraphic column
for the PNZ showing the five producing horizons at
Wafra field. Oil was discovered in the Upper Cretaceous
Maastrichtian in the PNZ at the South Fuwaris field lo-
cated south and west of Wafra field in March 1959. The
first Maastrichtian production at Wafra field occurred
in November 1959. The field has produced about
20 million bbl of oil primarily from two wells. Figure 3
provides a structure map for the Maastrichtian show-
ing the location of the Maastrichtian producing wells,
five wells with core data used to build the sequence-
stratigraphic interpretation, and the 123 wells used
to generate the geostatistical reservoir model.

FIGURE 1. Location of the Maastrichtian reservoir, Wafra field, Partitioned Neutral Zone (PNZ). S = south. Also shown are
the other producing fields, S. Umm Gudair, S. Fuwaris, and Humma in the PNZ.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 249

 the integration of the

cores with the well logs
for extrapolating the high-
frequency sequence (HFS)
correlations beyond the
cores in constructing the
reservoir model
 the impact of the data
integration in a three-
dimensional (3-D) Earth
model to provide insight
into possible varying oil-
water contacts and reser-
voir compartmentalization
 the importance of se-
quence stratigraphy and
3-D modeling in identify-
ing the reservoir potential
of the Wafra field Maas-
trichtian reservoir

REGIONAL
STRATIGRAPHY

The Maastrichtian inter-

val in the Wafra field con-
sists of the Upper Cretaceous
Tayarat Formation, which is
the uppermost part of the
Aruma Group and was de-
posited on a stable platform
(Figure 4). The geological set-
ting for the Maastrichtian
carbonates of the Arabian
shield includes (1) the Arabo-
Nubian shield, situated in
Saudi Arabia, covering sev-
eral hundreds of kilometers
in width; (2) a zone of clastic
sediments that rim the east-
ern margin of the shield con-
FIGURE 2. Stratigraphic column for the PNZ showing the five producing reservoirs of
taining both continental and
Wafra field.
shallow-marine siliciclastics,
sourced from the shield; and
The objectives of this chapter are to demonstrate (3) a carbonate shelf lying seaward of the clastic belt,
which was the site of shallow shelf carbonate sedi-
 the importance of core in defining the sequence- mentation (Harris et al., 1984). This shelf contains
stratigraphic framework and depositional intrashelf basins, which were filled with shales and
environment deep-water limestones.
 the methodology used to define the sequence- The Upper Cretaceous sequence was deposited dur-
stratigraphic framework ing a period of mild tectonic activity and is variable in
250 / Dull et al.

FIGURE 3. Structure map on top of high-frequency sequence (HFS) M00 (first Maastrichtian shale) constructed from
3-D seismic data. The contour interval is 50 ft (15 m). The map shows production from 35 wells at an average depth of
2700 ft (822 m). Map shows well locations used in the modeling, Maastrichtian producing wells, and cored well
locations.

regional thickness and facies. Most of the tectonism
affected areas to the east of the PNZ, particularly in the
northern Oman Mountains (Harris et al., 1984). A re-
gional unconformity is observed at the top of the Cre-
taceous (upper Maastrichtian), and moldic porosity is
commonly developed at this contact with the overly-
ing Tertiary. In some places, the upper Maastrichtian
is characterized by large collapse breccias formed by
evaporite solution (Harris et al., 1984). A lithostrati-
graphic correlation chart is found in Figure 5.
In outcrop, the Aruma is treated as a formation
and is divided into three members from base to top:
Khanasir Limestone Member, Hajajah Limestone
Member, and Lina Shale Member (El-Nakhal and El-
Naggar, 1994; Philip et al., 2002). The lower two mem-
bers are believed to be of Maastrichtian age, whereas
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 251

FIGURE 4. Map of the Arabian

Peninsula showing location of
the Partitioned Neutral Zone
(PNZ) in relation to major
structural features.

Gulf (El-Nakhal and El-Naggar,

1994) (Figure 6). In wells in
the Wafra field, the Tayarat
Formation is about 400 m
(1300 ft) thick, whereas in out-
crop, the Tayarat equivalent
is about 160 m (524 ft) thick.
The Aruma Formation un-
conformably overlies sand-
stones of the Wasia Forma-
tion and is overlain by dense
argillaceous limestone and
thin pyritic shales of the Umm
er Radhuma Formation.
As the Aruma succession
thickens from outcrop to the
subsurface to the east around
the Persian Gulf, the type lo-
the uppermost member is believed to be of Paleocene cality is not complete (Alsharhan and Nairn, 1990).
age based on biostratigraphic analysis. In the subsur- In addition, different formation names have been
face, the Aruma is given group rank, and in Kuwait, it used, and it has been subdivided differently in dif-
is divided into three formations (Gudair, Bahrah, and ferent places. Consequently, the Aruma is typically
Tayarat) (El-Nakhal and El-Naggar, 1994). The Aruma referred to as a group and may include anywhere from
increases in thickness from the outcrop areas in the two to six formations (Alsharhan and Nairn, 1990). As
west into the subsurface at the edges of the Persian noted earlier, the Maastrichtian interval in the Wafra

FIGURE 5. Lithostratigraphic correlation of the Upper Cretaceous in the Arabian Peninsula (Alsharhan and Nairn, 1990).
252 / Dull et al.
FIGURE 6. Schematic cross section showing Aruma surface to subsurface relationships (El-Nakhal and El-Naggar, 1994).
field in the PNZ is found within the Tayarat Forma-
tion, a part of the Aruma Group.
In Abu Dhabi, the Shah field produces minor oil
and gas-condensate from the Upper Cretaceous Aruma
Group, which consists of Coniacian to Maastrichtian
lithologies (Alsharhan and Kendall, 1995). There, the
Aruma Group consists of basal Laffan Formation (Co-
niacian) containing mainly slightly calcareous shales,
which is overlain by the Halul Formation (Santonian),
which consists of an argillaceous limestone. This is
overlain by the Fiqa Formation (Campanian), which
consists of alternating calcareous shales, marls, and
argillaceous limestone. These are capped by the Maas-
trichtian Simsima Formation containing fossilifer-
ous and shaly limestone and dolomitic limestone rim-
ming the Omani foredeep (Alsharhan and Kendall,
1995).
The Simsima Formation, equivalent to the Aruma
Formation in Oman, is interpreted to have been de-
posited on a shallow carbonate platform. In the
Shah field, it consists of two large-scale deposition-
al cycles. The lower cycle consists of a coarsening-
upward wackestone-mudstone and rudistid packstone-
grainstone, with finely crystalline dolomite and
minor anhydrite deposited in intertidal to supra-
tidal environments. The upper unit is composed of a
vertical succession of fining-upward sequences con-
taining packstone, wackestone, shale, and very finely
crystalline dolomite deposited in a subtidal to inter-
tidal setting. It becomes more restricted at the top.
The subsurface Simsima commonly contains inter-
particle, moldic, and vuggy porosity with locally good
intercrystalline porosity in the dolomites (Alsharhan
and Kendall, 1995).
During the Maastrichtian, the Arabian Peninsula
was located about 128 north of the equator on a rel-
atively stable continental margin on the African plate
as it moved north toward the Asian plate (Golonka
et al., 1994; Morris et al., 2002). The Neotethys fore-
deep narrowed during the Late Cretaceous, and the
paleogeographic reconstruction by Golonka et al.
(1994) shows the development of a land mass in the
northeast in part of the present-day Arabian Pen-
insula caused by the continued northward drift of
the Arabian plate (Sharland et al., 2001; Morris et al.,
2002). The work of Sayed and Mersal (1998) in Jebel
Rawdah in the northern Oman Mountains (a west-
northwest- to east-southeast-trending, postobduction
fold) shows that the Oman Mountains were emer-
gent in the early Maastrichtian. Smith et al. (1995)
placed the ophiolite in the Oman Mountains as latest
Campanian to Maastrichtian, whereas later work by
Schreurs and Immenhauser (1999) showed that the
ophiolite was deposited in late Maastrichtian to Pa-
leocene. The proximity of a landmass to the north-
east of the PNZ provides a likely source of the shale
in the second Maastrichtian interval. Alsharhan and
Nasir (1996), in their research in the western Oman
Mountains, showed that the Late Cretaceous Qalah
formation, which uncomformably onlaps the Sim-
sima Formation, indicates that the obducted Semail
ophiolite was extensively weathered in a tropical en-
vironment. Skelton et al. (1990), in their study of the
Qalah and Simsima Formations, suggest the Semail
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 253
ophiolite was emergent, and that a more open shelf
environment exists. The absence of evaporites in the
second Maastrichtian interval is consistent with de-
position on a shallow, gently dipping ramp under
humid conditions.
The obducted complex was strongly emergent at
the start of the Maastrichtian, the Aruma Sea trans-
gressed onto its western flank, and the ophiolite
subsided through that stage, forming a relatively
open shelf in the north, becoming broader and more
restricted farther south (Skelton et al., 1990). The con-
tinued narrowing of the Neotethys could also pos-
sibly explain the restriction in the marine circu-
lation that resulted in the lagoonal sedimentation
during sea level lowstands in the upper Maastrichtian
interval. Overall, the abundant evidence of subaer-
ial exposure, thin beds, dolomitization, and anhy-
drite suggests that deposition during the first Maas-
trichtian occurred on a very gently dipping, shallow,
arid, and restricted ramp setting that transitioned
between normal-marine conditions to restricted la-
goonal environments.
RESERVOIR DESCRIPTION
The Maastrichtian reservoir is composed mainly of
dolomite and limestone with thin beds of organic-
rich shale. Individual shale layers ranges from less
than 2 cm to 1 m (0.78 in. to 3.3 ft) or more. The
second Maastrichtian shale separates the Maastricht-
ian into an upper and lower portion with signifi-
cantly different depositional characteristics. The
type log given in Figure 7 shows the vertical extent
of the first and second Maastrichtian intervals. The
upper part of the Maastrichtian is also known as the
first Maastrichtian reservoir. The lower portion is com-
monly referred to as the second Maastrichtian res-
ervoir. This chapter will use upper and lower Maas-
trichtian, instead of first and second Maastrichtian
because second Maastrichtian shale does not extend
across the entire modeled area.
The lower Maastrichtian is composed largely of
dolomite with local limestone and shale, whereas
the upper Maastrichtian is composed largely of lime-
stone with locally abundant dolomite and shale. The
shale in the upper Maastrichtian is organic rich, with
total organic content as much as nearly 50% of the
rock volume and it is largely nonmarine. The shale in
the lower Maastrichtian contains very little organic
matter, less than 1% of the rock volume, and is largely
marine in origin, which is supported by the palyno-
logical interpretation as summarized in Table 1. Anhy-
drite occurs throughout the reservoir interval as nod-
ules (Figure 8), void and fracture filling (Figures 9, 10),
and, occasionally, as thick intervals of coalesced nod-
ules (Figure 11).
Oil production from the upper Maastrichtian res-
ervoir production occurs in the dolomitized, mud-
dominated peloidal packstones, skeletal-peloidal
grain-dominated packstones, and rudist rudstones.
The oil production from the lower Maastrichtian in-
terval occurs in dolomitized peloidal grainstones and
packstones.
The field has produced less than 1% of the origi-
nal oil in place (OOIP) based on an estimate of nearly
1.5 billion bbl obtained from the reservoir model
completed as part of this study. Production is highly
variable; initial rates range between several hundred
barrels of oil per day (BOPD) to more than 3000 BOPD.
Some wells are exceptional producers; one well has
produced more than 5 million bbl of oil and some
water out very quickly. Although representing a very
significant oil resource, the Maastrichtian reservoir
provides several challenges that this study addresses,
including (1) the presence of multiple barriers and
baffles that compartmentalize the reservoir areally
and vertically; (2) the areal variation of oil gravity
(13 –218 API); (3) the absence of a well-defined oil-
water contact; (4) a water zone in the uppermost part
of the reservoir interval; and (5) the distribution of
possible fractures and faults that may provide an ex-
planation for why some wells water out very quickly
whereas other wells produce at very low water cuts
for many years.
Prior published reports on the Maastrichtian res-
ervoir at Wafra field are very limited. Brief mention
is made by Nelson (1968) and Alsharhan and Nairn
(1997). Danielli (1988) provides some information
about the shallow reservoirs at Wafra field, includ-
ing the Maastrichtian.
SEQUENCE-STRATIGRAPHIC FRAMEWORK
OF THE MAASTRICHTIAN
The Maastrichtian reservoir represents a shallow-
ing-upward, lower order sequence culminating in a
regionally correlative subaerial exposure surface that
marks the top of the Maastrichtian and is interpreted
to be equivalent to the top of tectonostratigraphic
megasequence (TMS) AP9 of Sharland et al. (2001).
Abdel-Kireem et al. (1994) described an unconformi-
ty at base of the Paleocene and the top of the late
254 / Dull et al.

FIGURE 7. Type log cored well 3 for the Maastrichtian showing the reservoir interval, high-frequency sequence-
stratigraphic picks, and lithologic core description. The well-log curves are track 1 (CGR = corrected gamma ray;
SGR = spectral gamma ray; VCLGR = clay volume from gamma ray); track 2 (core porosity; PHIE = effective porosity);
track 3 (PHIE; BVO = bulk volume oil; BVW = bulk volume water); and track 4 (facies). Depth in feet.

Maastrichtian in northeastern Iraq. The TMS AP9 face is overlain by a marine incursion that is the base
marks the end of the Cretaceous ophiolite obduction of the Paleocene-age Umm Er Radhuma Formation.
on the northern Arabian plate, resulting in widespread A stratigraphic cross section through the five cored
regression (Beydoun, 1991, 1993). This exposure sur- wells showing the identified evidence for subaerial
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 255
Table 1. Interpreted environments of deposi-
tion from the shale palynology study of the
Maastrichtian.
Well Depth Interpretation HFS
(ft)
5 2567 Marine M00
5 2619 No palynomorphs M00
4 2743 Possible marine M10
3 2784 Marsh and embayment M20
5 2780 No palynomorphs M20
4 2796 Marine M20
3 2854 Restricted lagoon and fresh water M30
4 2851 Possible marine M40
3 2932 Distal lagoon, marine influence M50
5 2851 Marine M50
4 2885 Probable marine M50
exposure as hardgrounds and karst, interpreted envi-
ronments of deposition from the palynology, and in-
terpreted HFSs that were used to define the sequence-
stratigraphic framework of the Maastrichtian is given
in Figure 12. The Maastrichtian is an overall regressive
sequence composed of a transgressive systems tract
(TST) and a highstand systems tract (HST). Figures 8–11
and 13–31 show the core slab photos for the HFS M10
to M80 interval for cored well 3. The location of the
cored wells and cross section location are shown on
the map of the field in Figure 3.
The establishment of an appropriately detailed
sequence-stratigraphic framework provides the foun-
dation for the reservoir models as the stratal relation-
ships, and the genetically related sequences largely
control the spatial distribution of the facies and pe-
trophysical properties of the reservoir (Kerans and
Tinker, 1997; Deutsch, 2002). The basic framework
of the Maastrichtian is the HFS. The HFS, as defined
by Kerans and Tinker (1997), is bounded by a base-
level-fall to base-level-rise turnaround and is com-
posed of genetically related cycles and cycle sets. A
cycle is defined as the smallest set of genetically
related lithofacies representing a single base-level rise
and fall. This is comparable to the parasequence of
Van Wagoner et al. (1988, 1990). The cycle set is a
bundle of cycles that show a consistent trend of ag-
gradation or progradation (Kerans and Tinker, 1997).
The HFS was chosen as the basic building block for
correlation in the Maastrichtian because it could be
correlated with a great deal of confidence. Individual
cycle correlation in the Maastrichtian is hampered by
large distances between wells, abundance of expo-
sure surfaces, and thin beds related to low accommo-
dation space, making it uncertain if a facies change
was caused by eustatic sea level changes or topogra-
phy. The sequence-stratigraphic modeling for the
Maastrichtian was based on 413 m (1356 ft) of core
from five wells and well logs from 123 wells. The core
description and the correlation of the five Maastricht-
ian cores have revealed what is interpreted as nine
HFSs of an overall regressive sequence. Figure 7 is the
type log for the Maastrichtian showing nine of the
HFSs and the lithofacies from the core description.
With only five cored wells, it was essential to use the
well logs from the other wells to construct the 3-D
stratigraphic framework for the full Maastrichtian
reservoir interval. As documented below, a synthetic
lithofacies curve was generated from the petrophysi-
cal data to assist in the fieldwide stratigraphic inter-
pretation and correlation.
The HFS for the Maastrichtian was identified using
established bounding surface criteria that indicate a
base-level-fall to base-level-rise turnaround. The bound-
ing surfaces of the HFS were identified most com-
monly by subaerial exposure surfaces or unconfor-
mities expressed as hardgrounds that show burrow
and root casts filled with anhydrite commonly capped
by organic terrestrial lagoonal shale (Figure 32). Less
commonly, the bounding surfaces were identified by
karst breccia with solution-enhanced open fractures
and void-filling anhydrite occasionally associated with
subaerial hardgrounds (Figure 33). Uncommonly, the
HFS were identified by (1) tidal flat caps of dolomud-
stone and dolowackestone, with pseudomorphs of
anhydrite after gypsum (Figure 34), (2) lithofacies tract
offset from high-energy peloidal dolograinstones to
subtidal-marine dolowackestones and mud-rich dolo-
packstones (Figure 35), or (3) rudist and skeletal dolo-
rudstone with tidal-flat caps (Figure 36).
Fieldwide Correlation
The fieldwide sequence-stratigraphic correlation
of the 10 HFSs of the Maastrichtian identified in the
cored wells was extended to all the wells based on a
synthetic facies curve (FACIESG) and clay volume from
gamma-ray curve (VCLGR). Both curves were used to
assist in the fieldwide correlation.
The correlation between well-log response and res-
ervoir quality was essential to predicting facies on
noncored wells with well logs. A review of the core
material, measured porosity, and measured perme-
ability indicated seven lithofacies or rock types. In
the course of core review, it became necessary to dis-
tinguish the mud-lean dolopackstones from the mud-
rich dolopackstones and the dolorudstone from the
256 / Dull et al.

FIGURE 8. Core slab photos 2905 – 2920 ft (855 – 890 m)

FIGURE 10. Core slab photos 2938 – 2956 ft (895 – 900 m)
for cored well 3.
for cored well 3.

FIGURE 9. Core slab photos 2887 – 2905 ft (879 – 855 m) FIGURE 11. Core slab photos 2869 – 2887 ft (874 – 879 m)
for cored well 3. for cored well 3.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 257

FIGURE 12. Stratigraphic cross section through the five cored wells showing the identified evidence for subaerial
exposure, interpreted environments of deposition from the palynology, and interpreted HFSs that were used to define
the sequence-stratigraphic framework of the Maastrichtian. The Maastrichtian is an overall regressive sequence
composed of a transgressive system tract and a highstand system tract. Depth in feet.

dolofloatstones. In addition, the cemented hardground
or dolostone and dolomudstone were combined be-
cause they were indistinguishable on the well logs
and poor reservoir rocks. Figure 37 is a graph of po-
rosity and permeability of the dolopackstones show-
ing the difference between the mud-rich and mud-lean
dolopackstones. The dolorudstones and dolofloatstones
shown in Figure 38 also show a significant difference
in porosity and permeability. Figure 39 is a graph of
core porosity and permeability for the seven litho-
facies indicating the grain-rich facies; dolograinstone,
mud-lean dolopackstone, dolorudstone, and mud-rich
dolopackstone are the primary reservoir facies (litho-
facies 4–7). Lithofacies 1–3, dolomudstone, dolowacke-
stone, and dolorudstone are of low reservoir quality
and potential barriers or baffles to flow.
A method of predicting lithofacies in noncored wells
was developed based on the multiple cross-product,
nonlinear method as documented by Lawrence et al.
(1997). The FACIESG curve was calculated using the
core lithofacies data from cored wells 2– 5.
Seven distinct carbonate lithofacies and shale were
identified from the core description that could be dis-
tinguished using well-log curves. The pore system is
all secondary in nature and is primarily intercrystal-
line because of the fabric-destructive dolomitization.
The original depositional textures appear to have in-
fluenced the development of the intercrystalline
pores. The mud-dominated lithofacies (lime mud-
stone, dolomudstone, dolowackestone, and dolo-
floatstone) (Figure 40a–c) are of poor reservoir quality
because of the predominance of original fine lime
micrite matrix or dolomitized muddy matrix resulting
in connected pore system-dominated very fine inter-
crystalline micropores. Figure 40d–g are thin-section
photomicrographs of the more grain-dominated litho-
facies (mud-rich dolopacktstone, dolorudstone, mud-
lean dolopackstone, and dolograinstones). These rock
258 / Dull et al.

FIGURE 15. Core slab photos 2753 – 2771 ft (839 – 844 m)

FIGURE 13. Core slab photos 2722 – 2740 ft (829 – 835 m) for cored well 3.
for cored well 3.

FIGURE 14. Core slab photos 2740 – 2753 ft (835 – 839 m) FIGURE 16. Core slab photos 2771 – 2784 ft (844 – 848 m)
for cored well 3. for cored well 3.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 259

FIGURE 17. Core slab photos 2784 – 2802 ft (848 – 854 m) FIGURE 19. Core slab photos 2815 – 2833 ft (858 – 863 m)
for cored well 3. for cored well 3.

FIGURE 18. Core slab photos 2802 – 2815 ft (854 – 858 m) FIGURE 20. Core slab photos 2833 – 2851 ft (863 – 868 m)
for cored well 3. for cored well 3.
260 / Dull et al.

FIGURE 21. Core slab photos 2851 – 2869 ft (868 – 874 m) FIGURE 23. Core slab photos 2956 – 2974 ft (900 – 906 m)
for cored well 3. for cored well 3.

FIGURE 22. Core slab photos 2920 – 2938 ft (890 – 895 m) FIGURE 24. Core slab photos 2974 – 2989 ft (906 – 911 m)
for cored well 3. for cored well 3.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 261

FIGURE 25. Core slab photos 2989 – 3007 ft (911 – 916 m) FIGURE 27. Core slab photos 3025 – 3040 ft (922 – 926 m)
for cored well 3. for cored well 3.

FIGURE 26. Core slab photos 3007 – 3025 ft (916 – 922 m) FIGURE 28. Core slab photos 3040 – 3058 ft (926 – 932 m)
for cored well 3. for cored well 3.
262 / Dull et al.

FIGURE 29. Core slab photos 3058 – 3076 ft (932 – 937 m)

for cored well 3.

FIGURE 30. Core slab photos 3076 – 3094 ft (937 – 943 m) FIGURE 31. Core slab photos 3094 – 3202 ft (943 – 975 m)
for cored well 3. for cored well 3.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 263

FIGURE 32. Core slab photo from cored well 3 (2857 ft;
870 m) showing uneven and eroded hardground with
burrows and root casts filled with anhydrite overlain by
organic-rich lagoonal shale.

fabrics are of better reservoir quality because of the

predominance of more coarsely crystalline dolomite,
producing the larger, better connected poor system and
lack of microporosity associated with a muddy matrix.
In addition, the permeability of the grain-dominated
lithofacies is higher because of common fine interpar-
ticle pores and associated late-stage etching. Table 2
shows the mean porosity and permeability of the car-
bonate lithofacies ranked by reservoir quality.
Lithofacies data from well 1 were not used because
of difficulties ensuring a reasonable depth match
given the very limited amount of core data. The well-
log curves PHIE (effective porosity), Sw (water satura-
tion), VCLGR (clay volume from corrected gamma- FIGURE 33. Subaerial hardground overlying karst breccia
ray curve), CGR (corrected gamma ray for uranium), with solution-enhanced open fractures and void-filling
and SGR (spectral gamma ray) were used to calculate anhydrite. Cored well 5 at depth of 2765 ft (842 m).
264 / Dull et al.
FIGURE 34. Dolomudstone with anhydrite after gypsum
crystals (cored well 4 at 2854 ft [869 m]).
the FACIESG curve. The correlation coefficient varied
between 0.41 and 0.79. A relatively low correlation
between actual facies and predicted facies (FACIESG)
was found for wells 2 and 4 and was attributed to
having less core data as compared to the other two
wells and are thus likely to be less precisely depth
matched than the other cored wells.
The FACIESG curve was critical to the fieldwide
correlation for the Maastrichtian because it provided
lithofacies curves to assist in identifying the cycle hi-
erarchy. The cycle hierarchy and the HFSs identified
in the core description, coupled with the FACIESG
curves on the noncored wells, were used for field-
wide correlation and to construct the 3-D sequence-
stratigraphic framework.
The VCLGR curve generated from the corrected
CGR was used in the sequence-stratigraphic correla-
tion, particularly in the upper Maastrichtian HFS M40
to M00. The core description revealed that many of the
hardground surfaces were capped with organic-rich
shale (Figure 41). The hardgrounds are consistently
identified by their low porosity and permeability and
appear as mudstones on the FACIESG curves. The shale
can easily be distinguished with the VCLGR curve as
shown on the type log given previously in Figure 7.
FIGURE 35. Core slab photographs from cored well 3
showing cycle top of HFS M70 of high-energy peloidal
dolograinstones capped with subtidal-marine low-energy
mud-rich peloidal dolopackstones and dolowackestones
(3007 – 3025 ft; 916 – 922 m).
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 265
FIGURE 36. Core slab photographs (2878 – 2905 ft; 877 –
885 m) from cored well 5 showing dolofloatstone capped
with intertidal dolomudstone, karst, and hardground
defining the top of HFS M40.
Lower Maastrichtian HFS M90 to M50:
Transgressive Systems Tract
The deposition in the lower Maastrichtian reser-
voir from the HFS M90 to HFS M50 interval repre-
sents a marine transgression with a maximum flood
back in the middle Maastrichtian. The maximum
flooding surface (MFS) in HFS M50 at Wafra field
is interpreted to be equivalent to the MFS K180 of
Sharland et al. (2001). The maximum flooding event
resulted in the deposition of the second Maastrich-
tian shale across most but not all of the Wafra field
area.
The TST is composed of five HFS. The HFS M90 to
M60 represents overall shallowing-upward sequences
composed of cycles from subtidal, peloidal, and bio-
clastic dolowackestone to dolopackstone or dolo-
grainstone and, less commonly, subtidal mudstone
to peloidal dolograinstone cycles. The uppermost
HFS records the maximum transgression and is com-
posed of marine shale, argillaceous mudstone, and
peloidal wackestone. The existence of the peloidal
grainstones in the lower Maastrichtian in the M70
interval in cored well 3 seems to indicate a less-
restricted and higher energy environment of deposi-
tion, with cycles that are generally thicker, indicat-
ing a greater accommodation space. The deepening
HFS M50 includes an MFS that is marked by the
deposition of the second Maastrichtian shale. The
second Maastrichtian was probably deposited in more
open-marine conditions as evidenced by the lack of
organic matter as compared to the upper Maastrich-
tian and as suggested from the palynological inter-
pretation from HFS M50 in wells 3, 4, and 5 as given
in Table 1.
FIGURE 37. Graph of porosity
and permeability for lithofa-
cies showing the difference
on reservoir quality between
mud-rich (4) and mud-lean
dolopackstone (6).
266 / Dull et al.
Upper Maastrichtian HFS M40 to M00:
Highstand Systems Tract
The upper Maastrichtian HST is composed of five
shallowing-upward HFS sequences commonly capped
with hardgrounds and organic-rich shale. Flugel (2004,
p. 206) describes hardgrounds as subtidal cementa-
tion at the water-sediment interface as ‘‘related to a
FIGURE 38. Graph of porosity
and permeability for lithofacies
showing the difference on reser-
voir quality between dolofloat-
stone (3) and dolorudstone (5).
combination of nondeposition or low accumulation
rates, and condensation.’’ Hardgrounds are surfaces
that are cemented before the next sedimentation
event. The core data show that the HST is typically
characterized by thin cycles, indicating less accom-
modation space, and also exhibits a greater variabil-
ity in facies with relative sea level changes. The lack
FIGURE 39. Graph of porosity
and permeability for the seven
lithofacies showing that the
more grain-dominated rocks
have higher permeability. The
best reservoir rock is the grain-
dominated lithofacies shown
in red (7 = dolograinstone; 6 =
mud-lean dolopackstone) and
green (5 = dolorudstone; 4 =
mud-rich dolopackstone). The
poor reservoir rock and poten-
tial baffles or barriers to flow
are shown in blue (3 = dolo-
floatstone; 2 = dolowacke-
stone; 1 = dolomudstone).
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 267
of accommodation space and resulting thin cycles in
the upper Maastrichtian is reflected in the number
and types of hardgrounds.
Two types of hardgrounds dominate the upper
Maastrichtian: a subtidal firmground to incipient hard-
ground and an intertidal to supratidal hardground
as described by Hillgartner (1998). Hillgartner (1998),
p. 1095, described the subtidal firmground as ‘‘low ac-
cumulation rate in a lagoonal environment, which
favors consolidation and incipient cementation of
the sediment at the water-sediment interface.’’ The
upper Maastrichtian firmgrounds show characteris-
tics similar to those described by Hillgartner (1998):
highly bioturbated with overlying sediment infilling
the burrows and an irregular surface probably caused
by weak erosion and/or differential cementation. The
overlying sediment infilling the burrows in this case
is organic-rich lagoonal shale as shown in Figure 41.
The intertidal to supratidal hardgrounds tend to be
flat with a great deal more of associated anhydrite-
infilling burrows, fenestral pores, and evidence of
exposure such as tension gashes. The intertidal to sup-
ratidal hardgrounds are occasionally found above
microkarsts.
The recognition of hardgrounds of low porosity
and permeability in the upper Maastrichtian is im-
portant not only for cycle correlation but for po-
tential impact on reservoir compartmentalization.
The incipient subtidal and intertidal hardgrounds as
described by Hillgartner (1998) were found to have
lateral extents from 0.1 to 1 km (0.06 to 0.6 mi). Some
of the hardground-capped cycles cannot be correlat-
ed in the upper Maastrichtian, suggesting that the
hardgrounds may not act as continuous barriers to
flow throughout the entire reservoir.
Floatstones and rudstones dominate the middle
part of the HST from the M50 to the M20 marker.
The floatstones-rudstones are composed of broken
to intact rudists (Figure 42), bivalves, and gastropods.
The floatstones commonly contain the large benthic
foraminifer, Loftusia (Figure 43). The floatstone-
rudstone less commonly contains corals, stromato-
poroids, and bryozoans. The deposition of the rudist
and skeletal debris is interpreted to be biostromal,
with no evidence of mounding. An abrupt facies
change is indicated from the HFS M20 to M10 from
restricted lagoon to more open-marine deposition of
peloidal and bioclastic packstones.
A subareal exposure surface occurs at the top of
HFS M10 (Figure 12) that is porosity destructive and
provides the seal on the oil-productive interval. This
is shown on the cores as low porosity and permeabil-
ity and on the well logs as a sharp change from low
to high water saturation (Figure 44). The HFS M10
surface separates the oil-productive interval from
known water production from the HFS M00. A ma-
rine transgression across the top of the Maastrichtian
formation overlies the unconformity surface M00
that marks the top of the Maastrichtian. HFS M00 in
cored well 5 shows a karsted interval capped with an
intertidal hardground that is overlain with marine
shale (Figure 45).
The palynological and x-ray diffraction data from
the shale for the HFS M50 to HFS M00 suggest that
the shale deposition transitioned from a quiet la-
goonal setting to marine settings that are interpreted
as localized marine incursions that resulted from eu-
static sea level rise.
GEOSTATISTICAL RESERVOIR MODEL
A geostatistical model for the Maastrichtian res-
ervoir was generated to assess OOIP and reservoir
development potential. Figure 3 shows the area mod-
eled and the distribution of wells with appropri-
ate well-log data (e.g., the synthetic facies curve
FACIESG described above along with porosity and Sw
curves). The modeled area covers 144 km2 (55.6 mi2).
The modeled interval includes the entire Maastricht-
ian reservoir and the top of the Maastrichtian (HM00)
to the top of Hartha. The overall model grid dimen-
sions are 106  136  622 cells (8.97 million cells
total). The areal cell size is 100  100 m (330  330 ft),
and the vertical layering is nominally 0.3 m (1 ft)
between the HM00 and HM90 markers. The nominal
vertical layering is 0.6 m (2 ft) from the HM90 marker
to the base of the model.
The geostatistical property model was generated
using an industry standard workflow (Deutsch, 2002;
Kelkar and Perez, 2002; Goovaerts, 1997). Porosity was
distributed by sequential Gaussian simulation (sGs)
constrained by sequence-stratigraphic layers. Sw was
distributed using a colocated cokriging with sGs al-
gorithm also constrained by sequence-stratigraphic
layers. Tables 3 and 4 summarize the input data and
semivariograms used to build the reservoir property
model. Although not used to constrain porosity or
Sw distribution, several additional reservoir proper-
ties, such as FACIESG and Pe (photoelectric effect),
were also distributed using sGs constrained by a strat-
igraphic layer to assist in understanding the develop-
ment potential of the reservoir, particularly the se-
quences below M20.
268 / Dull et al.

FIGURE 40. Thin-section photomicrographs illustrating the seven dolomite lithofacies that compose the Maastrichtian
arranged by reservoir quality. (a) Argillaceous lime mudstone with a few bioclasts and composed primarily of silt-size
micrite and micrite that has been neomorphosed to microspar then etched, creating the microporosity. Organic matter
is present as fine filamentous algal matter. Porosity occurs as micropores and reduced molds. The permeability is poor
because the connected pore system is dominated by the very fine intercrystalline pores (Phi = 17.9%, k = 0.16 md).
(b) Bioclastic dolowackestone or grain-poor argillaceous and organic dolomite composed of fine-grained dolomite
matrix cut by coarse-grained dolomite laminae or burrows. The matrix is composed of organic matter and detrital clays
but the dolomite crystals form the framework. Organic matter is common as pellets and flattened tabular grains.
Porosity occurs in micropores, and permeability is poor because of the replacement of the fine carbonate matrix with
tightly interlocking dolomite crystals (Phi = 4.5%, k < 0.10 md). (c) Bioclastic dolofloatstone with the most common grains
are benthonic foraminifera (Rotalids) and Loftusia. Other bioclasts include ostracods, bivalve debris, gastropods, and
echinoderm plates. Organic matter is found as reddish brown pellets and possibly bitumen. Uncommon angular fragments
of bones are also present. The original matrix is not preserved and has been replaced or etched. Dolomite is the major
mineral. Porosity is a combination of molds and secondary interparticle pores. Permeability is low as result of the
variations in porosity and the frequency of organic laminae (Phi = 14.4%, k < 2.21 md). (d) Mud-rich dolopackstone in
which the fabric-destructive dolomitization has largely destroyed the original depositional fabric. No evidence of
carbonate mud matrix is preserved, and the organic matter is preserved as pellets. The finer dolomite crystal laminae
suggest that the original fabric contained some lime mud. Porosity occurs as secondary intercrystalline pores with
some local intracrystalline pores. Permeability is low because of the generally interlocking mosaic of the dolomite
crystals. Larger isolated pores exist, but are connected by finer pore throats (Phi = 17.6%, k < 5.00 md).
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 269

FIGURE 40. (cont.). (e) Dolorudstone is a mixture of fine-to medium-grained dolomite that has partly replaced grains
and relatively coarse dolomite, destroying much of the original rock fabric. The fabric suggests an original bioclastic
grainstone that was overpacked. No matrix appears to have been present. The framework grains are commonly rudist
debris, bivalve debris fragments of Orbitoides with possible stromatoproids and/or Loftusia. The porosity is all secondary
as intercrystalline pores. Calcite cement has reduced some of the finer and more isolated pores. Permeability is higher
because of the occurrence of fine interparticle pores (Phi = 19.8%, k < 80.0 md). (f ) Mud-lean peloidal dolopackstone
(Phi = 15.3%, k < 9.5 md) that has undergone fabric-destructive dolomitization and dissolution of the original rock. Large
tabular molds may have been molluskan debris. Only ghosts of what appear to be medium-grain-size peloids, ooids,
or other rounded grains are found. Organic matter is present as fibrous or brown amorphous pellets. Porosity is
composed exclusively of secondary pores, primarily intercrystalline pores, although reduced molds are also found.
Some intracrystalline pores exist but are uncommon and ineffective. Permeability is good probably because of
the uneven distribution of pores and appears to be associated with carbonaceous laminae, suggesting that the original
fabric may have been a poorly sorted grainstone. (g) Peloidal dolograinstone with a mixture of fine- to medium-
crystalline dolomite that has partly replaced grains and coarsely crystalline dolomite, which has largely destroyed the
rock fabric. Extensive pores created by leaching of undolomitized limestone have been extensively cemented
by calcite. No matrix appears to have been present. Partly mimetic-replaced bioclasts suggest small benthonic fora-
minifera, molluskan debris, algae, and fragments of large foraminifera. Very uncommonly are grain outlines com-
plete and almost always rounded or tabular. Porosity occurs as secondary intercrystalline pores. Permeability is
high because of late-stage etching and an open dolomite framework (Phi = 19.9%, k = 1108 md).
270 / Dull et al.

Table 2. Average core porosity and permeability locations as well as select exploration targets in the
by lithofacies.
Maastrichtian reservoir’s deeper portions.
Lithofacies* Mean Permeability Mean Figure 48 shows the variation in Pe and, hence,
(md) Porosity limestone and dolomite throughout the reservoir.
Note that the dolomite is much more abundant in the
1 0.022 0.086 more productive, upper parts of the Maastrichtian.
2 1.445 0.080
3 20.550 0.114
4 11.477 0.134
MAASTRICHTIAN DEVELOPMENT POTENTIAL
5 81.037 0.182
6 114.817 0.200
The Maastrichtian reservoir contains more than
7 294.868 0.202
1 billion bbl of oil in place. The degree of compart-
*1 = dolomudstone; 2 = dolowackestone; 3 = dolofloatstone; mentalization of the Maastrichtian can be shown by
4 = mud-rich dolopackstone; 5 = dolorudstone; 6 = mud-lean
dolopackstone; 7 = dolograinstone.

As shown in Table 3 and Figures 46 and 47, a gen-

eral increase in porosity with depth is present, along
with an increase in Sw. Porosity is very irregularly
distributed in the upper parts of the reservoir, but is
generally best developed in the M10 and M20 inter-
vals. The Sw distribution shown in Figure 47 provides
the best depiction of the distribution of productive
and potentially productive parts of the Maastrichtian
reservoir at Wafra. The middle parts of the Maas-
trichtian show isolated regions of high porosity and
low Sw that may be near-term exploration targets.
Developing a proper sequence-stratigraphic frame-
work and using it to constrain 3-D reservoir models is
key if the 3-D models are to be used to prioritize infill

FIGURE 41. Core slab photo from cored well 2 of black FIGURE 42. Core slab photo of rudist that appears to be
coaly shale in the upper Maastrichtian deposited on an intact, center filled with blue anhydrite (cored well 5,
uneven subaerial hardground. (12816 ft [858 m]). 2722 ft [829 m]).
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 271
FIGURE 43. Core slab photo of Loftusia dolowackestone
(cored well 4, 2826 ft [861 m]).
maps of hydrocarbon pore volume (HCPV) for the
HFS M10 and M40 shown in Figures 49 and 50. The
trend of the highest HCPV for HFS M10 is coinci-
dent with present-day structural trends and repre-
sents a peloidal dolopackstone shoal complex of the
late HST. The M10 is the most extensive and pro-
ductive interval of the Maastrichtian. Whereas, the
HFS M40 shows a northeast to southwest linear trend
that likely parallels paleotopography. The northeast
area of the model also shows significant HCPV de-
velopment in the M40 that trends northeast to
southwest along the paleohigh. The northeast trend
of prospective reservoir development appears to
continue outside the model area beyond well con-
trol. No Maastrichtian well completions exist in this
area, and the model suggests significant exploration
potential to the north of the model area along the
Maastrichtian paleohigh and along paleodeposition-
al strike. Note that if the reservoir property models
used to generate the HCPV distributions were not
constrained by stratigraphy, but instead by lithology,
the trend shown in Figure 50 probably would not
have been captured, much less be used, to assist devel-
opment and exploration planning. Additional stud-
ies, incorporating seismic attributes, are in progress
to further refine the 3-D reservoir property maps and
to enhance the usefulness of the models to assist res-
ervoir development.
SUMMARY
The methodology for developing the sequence-
stratigraphic framework for the Maastrichtian at
Wafra field included a workflow to integrate core
and well-log data enabling the construction of the
3-D sequence-stratigraphic framework. The work-
flow for the Maastrichtian consisted of
 using the five cores to define the depositional
environment and cycle hierarchy
 correlating the five cores to define the ten HFSs
 developing synthetic facies curves from the well-
log data that enabled the use of the 123 well logs
with FACIESG and VCLGR curves to construct
the 3-D stratigraphic framework
The sequence-stratigraphic framework is essential
to the geostatistical modeling and to reservoir simu-
lation because it controls the spatial distribution of
petrophysical properties and identifies the chrono-
stratigraphic surfaces that act as barriers and baffles
to flow that compartmentalize the reservoir.
The modeling of the Maastrichtian reservoir has
revealed
 a large oil resource of more than 1 billion bbl of
oil in place
 that entrapment of hydrocarbons is primarily
stratigraphic
 a highly compartmentalized reservoir
272 / Dull et al.

FIGURE 44. A stratigraphic cross section of the cored wells shows the varying water saturation observed in the reservoir
interval and the grain-dominated lithofacies that typically have low water saturation. The M00 at the top of the
Maastrichtian from production tests is known to be water productive. Besides the M00, the M90 interval is the only
Maastrichtian sequence that shows nearly 100% water saturation. Depth in feet.

 multiple oil-water transition zones
 no single definitive oil-water contact
 a water zone above and below the oil reservoir
 additional development potential associated
with paleohighs and facies trends
and vertical hydrocarbon distribution of the Maas-
trichtian that will have a substantial impact on the
reservoir development.
ACKNOWLEDGMENTS

The sequence-stratigraphic framework and geo- The authors thank the Ministry of Petroleum and
statistical modeling have quantified a very signifi- Mineral Resources, Kingdom of Saudi Arabia; Saudi
cant asset and provided great insight into the areal Arabian Texaco; and Chevron Energy Technology
The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 273

Table 3. Summary of the well-log porosity and

water-saturation data used to construct the
reservoir property models listed by HFS.

Stratigraphic Average Average

Interval Porosity Sw

M00 0.166 0.795

M10 0.150 0.531
M20 0.137 0.635
M30 0.131 0.611
M40 0.130 0.670
M50 0.125 0.761
M60 0.127 0.607
M70 0.135 0.676
M80 0.165 0.779
M90 0.202 0.896

FIGURE 45. Core slab photo of top of HFS M00 in cored

well 5 (2567 – 2568 ft; 782.4 – 782.7 m) showing hard-
ground capped with shaly dolomudstone and a dark
organic-rich marine shale marking a marine incursion in
the late Maastrichtian. Note the large lithoclast above
the irregular hardground surface.
274 / Dull et al.

Table 4. Summary of the variograms used in the property distribution of the geostatistical models.
Interval Range x (m) Range y (m) Azimuth Range z (m) Nugget Sill Form

M00 2460 1350 N1258E 6.0 0 1 Exponential

M10 1660 1220 N608E 2.7 0 1 Exponential
M20 820 660 N458E 4.6 0 1 Exponential
M30 1830 1260 N258E 4.4 0 1 Exponential
M40 1600 1110 N258E 3.4 0 1 Exponential
M50 1100 710 N408E 2.6 0 1 Exponential
M60 2060 1140 N108E 4.3 0 1 Exponential
M70 1720 1380 N358E 5.5 0 1 Exponential
M80 1760 1020 N408E 6.6 0 1 Exponential
M90 1860 1000 N358E 9.8 0 1 Exponential

FIGURE 46. Perspective view of Wafra Maastrichtian reservoir model showing porosity distribution. Note that porosity
distribution is very discontinuous, even in the generally higher porosity, lowermost stratigraphic zones. Within the
uppermost stratigraphic zones, only the M10 and M20 intervals show consistent high-porosity distribution across
large areas of the field. Vertical exaggeration = 5. Map shows the structure on top of the Hartha Formation (base
of Maastrichtian).
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 275

FIGURE 47. Perspective view of Wafra Maastrichtian reservoir model showing Sw distribution. Note the general increase
in Sw with depth. Also note that some of the middle stratigraphic zones (e.g., M40) have locally low Sw regions. Vertical
exaggeration = 5. Map shows structure on the top of the Hartha Formation (base of Maastrichtian).

FIGURE 48. Perspective view of Wafra Maastrichtian reservoir model showing Pe (photoelectric effect) distribution. Note
the general increase in Pe with depth showing the irregular distribution of limestone ( yellow) and dolomite (blue).
Note, however, that in general, there is much more dolomite in the upper parts of the Maastrichtian. Vertical
exaggeration = 5. Map shows structure on the top of the Hartha Formation (base of Maastrichtian).
276 / Dull et al.

FIGURE 49. Hydrocarbon pore volume

(HCPV) map for the M10 to the M20
interval. The area shown in orange is
greater than 10 ft (3 m) of HCPV. The
wells that have been completed in the
Maastrichtian are also shown. Contour
interval is 2 ft (0.6 m).
The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 277

FIGURE 50. Map of hydrocarbon pore

volume (HCPV) for M40. The greatest
of amount of HCPV is coincident with
the maximum dolograinstone thickness.
This observation strongly supports the
premise that the hydrocarbon trapping
is primarily stratigraphic. The contour
interval for M40 HCPV map is 1 ft (0.3 m).
278 / Dull et al.
Company for their permission to share the results
of this study. We also thank the reviewers for their
insight and guidance to this study: Steven L. Bachtel,
ConocoPhillips Company; Sherry Becker, ExxonMobil
Upstream Research Company; and P. Mitch Harris,
Chevron Energy Technology Company.
REFERENCES CITED
Abdel-Kireem, M. R., A. M. Samir, and H. P. Luterbacher,
1994, Planktonic foraminifera from the Kolosh For-
mation (Paleogene) of the Sulaimaniah – Dokan re-
gion, northeastern Iraq; Tuebingen University Neues
Jarhrbuch Geol.: Paleontology Monatsh, v. 9, p. 517 –
527.
Alsharhan, A. S., and S. J. Y. Nasir, 1996, Sedimentological
and geochemical interpretation of a transgressive se-
quence: The Late Cretaceous Qahlah formation in the
western Oman Mountains: United Arab Emirates Uni-
versity Sedimentary Geology, v. 101, no. 3 – 4, p. 227 –
242.
Alsharhan, A. S., and C. G. St. C. Kendall, 1995, Facies var-
iation, depositional setting and hydrocarbon poten-
tial of the Upper Cretaceous rocks in the United Arab
Emirates: Cretaceous Research, v. 16, p. 435 – 449.
Alsharhan, A. S., and A. E. M. Nairn, 1990, A review of the
Cretaceous formations in the Arabian Peninsula and
Gulf: Part III. Upper Cretaceous (Aruma Group) stratig-
raphy and paleontology: Journal of Petroleum Ge-
ology, v. 13, p. 247 – 266.
Alsharhan, A. S., and A. E. M. Nairn, 1997, Sedimentary
basins and petroleum geology of the Middle East: New
York, Elsevier, p. 843.
Beydoun, Z. R., 1991, Arabian plate hydrocarbon, geology
and potential — A plate tectonic approach: AAPG Stud-
ies in Geology 33, 77 p.
Beydoun, Z. R., 1993, Evolution of the northeastern
Arabian plate margin and shelf: Hydrocarbon habitat
and conceptual future potential: Revue de l’Institut
Français du Pétrole, v. 48, p. 311 – 345.
Danielli, H. M. C. D., 1988, The Eocene reservoirs of Wafra
field, in A. J. Lomando and P. M. Harris, eds., Giant
oil and gas fields — A core workshop: SEPM Core
Workshop 12, p. 119 – 154.
Deutsch, C. V., 2002, Geostatistical reservoir modeling:
New York, Oxford University Press, 376 p.
El-Nakhal, H. A., and Z. R. El-Naggar, 1994, Review of the
biostratigraphy of the Aruma Group (Upper Cretaceous)
in the Arabian Peninsula and surrounding regions:
Cretaceous Research, v. 15, p. 401 – 416.
Flugel, E., 2004, Microfacies of carbonate rocks: Analysis,
interpretation and application: Berlin, Spring-Verlag,
p. 976.
Golonka, J., M. I. Ross, and C. R. Scotese, 1994, Phan-
erozoic paleogeographic and paleoclimatic modeling
maps, in A. F. Embry, B. Beauchamp, and D. J. Glass,
eds., Pangea global environments and resources: Cana-
dian Society of Petroleum Geologists Memoir 17, p. 1 –
47.
Goovaerts, P., 1997, Geostatistics for natural resources
estimation: New York, Oxford University Press, 483 p.
Harris, P. M., S. H. Frost, G. A. Seiglie, and N. Schneider-
mann, 1984, Regional unconformities and deposition-
al cycles, Cretaceous of the Arabian Peninsula, in J. S.
Schlee, ed., Interregional unconformities and hydro-
carbon accumulation: AAPG Memoir 36, p. 67 – 80.
Hillgartner, H., 1998, Discontinuity surfaces on a shallow-
marine carbonate platform (Berriasian, Valanginian,
France and Switzerland): Journal of Sedimentary Re-
search, Section B: Stratigraphy and Global Studies,
v. 68, no. 6, p. 1093 – 1108.
Kelkar, M., and G. Perez, 2002, Applied geostatistics for
reservoir characterization: Society of Petroleum Engi-
neers, 264 p.
Kerans, C., and W. S. Tinker, 1997, Sequence stratigraphy
and characterization of carbonate reservoirs: SEPM
Short Course Notes 40, 130 p.
Lawrence, T. D., A. Kohar, B. Sukamto, and H. Pramono,
1997, Sonic density well log data editing with pseudo-
curve generation — Indonesian examples using a mul-
tiple crossproduct non-linear method: World Petro-
leum Congress, p. 163 – 167.
Morris, A., M. W. Anderson, A. H. F. Robertson, and K. Al-
Riyami, 2002, Extreme tectonic rotations within an
eastern Mediterranean ophiolite (Baer-Bassit, Syria):
Earth and Planetary Science Letters, v. 202, no. 2,
p. 247 – 261.
Nelson, P. H., 1968, Wafra field— Kuwait-Saudi Arabia neu-
tral Zone. Second Regional Technical Symposium, So-
ciety of Petroleum Engineers, Saudi Arabia Section,
Dhahran, 16 p.
Philip, J. M., J. Roger, D. Vaslet, F. Cecca, S. Gardin, and
A. M. S. Memesh, 2002, Sequence stratigraphy, bio-
stratigraphy and paleontology of the Maastrichtian –
Paleocene Aruma Formation in outcrop in Saudi Arabia:
GeoArabia, v. 7, no. 4, p. 699–718.
Sayed, M. G. S. A., and M. A. Mersal, 1998, Surface geology
of Jebel Rawdah, Oman Mountains: GeoArabia, v. 3,
no. 3, p. 401 – 414.
Schreurs, G., and A. Immenhauser, 1999, West-northwest
directed obduction of the Batain group on the east-
ern Oman continental margin at the Cretaceous –
Tertiary boundary: Tectonics, v. 18, no. 1, p. 148 –
160.
Sharland, P. R., R. Archer, D. M. Casey, S. H. Hall, A. P.
Heward, A. D. Horbury, and M. D. Simmons, 2001,
Arabian plate sequence stratigraphy: GeoArabia Spe-
cial Publication 2, 371 p.
Skelton, P. W., S. C. Nolan, and R. W. Scott, 1990, The
Maastrichtian transgression onto the northwestern
flank of the Proto-Oman Mountains: Sequences of
rudist-bearing beach to open shelf facies, in A. H. F.
Robertson, M. P. Searle, and A. C. Ries, eds., The ge-
ology and tectonics of the Oman region: Geological
Society (London) Special Publication 49, p. 521 – 547.
 The Sequence Stratigraphy of the Maastrichtian Reservoir at Wafra Field / 279
Smith, A. B., N. J. Morris, A. S. Gale, and W. J. Kennedy,
1995, Late Cretaceous carbonate platform faunas of
the United Arab Emirates – Oman border region: Ox-
ford University Bulletin of the Natural History Museum
(Geology Series), v. 51, no. 2, p. 91–119.
Van Wagoner, J. C., H. W. Posamentier, R. M. Mitchum,
P. R. Vail, J. F. Sarg, T. S. Loutit, and J. Hardenbol,
1988, An overview of the fundamentals of sequence
stratigraphy and key definitions, in C. K. Wilgus, B. J.
Hastings, H. Posamentier, J. C. Van Wagoner, C. A. Ross,
and C. G. St. C. Kendall, eds., Sea-level change: An in-
tegrated approach: SEPM Special Publication 42, p. 39 –
46.
Van Wagoner, J. C., R. M. Mitchum, K. M. Campion, and
V. D. Rahmanian, 1990, Siliclastic sequence stratigra-
phy in well logs, cores, and outcrops: Concepts for
high-resolution correlation of time and facies: AAPG
Methods in Exploration Series 7, 55 p.
 Porter, M. L., A. R. G. Sprague, M. D. Sullivan, D. C. Jennette, R. T. Beaubouef,
T. R. Garfield, C. Rossen, D. K. Sickafoose, G. N. Jensen, S. J. Friedmann,

8
and D. C. Mohrig, 2006, Stratigraphic organization and predictability of
mixed coarse- and fine-grained lithofacies successions in a lower Miocene
deep-water slope-channel system, Angola Block 15, in P. M. Harris and L. J.
Weber, eds., Giant hydrocarbon reservoirs of the world: From Rocks to
reservoir characterization and modeling: AAPG Memoir 88/SEPM Special
Publication, p. 281 – 305.

Stratigraphic Organization and

Predictability of Mixed Coarse- and
Fine-grained Lithofacies Successions
in a Lower Miocene Deep-water
Slope-channel System,
Angola Block 15
M. L. Porter C. Rossen
ExxonMobil Production Company, Houston, ExxonMobil Development Company,
Texas, U.S.A. Houston, Texas, U.S.A.

A. R. G. Sprague D. K. Sickafoose
ExxonMobil Upstream Research Company, ExxonMobil Exploration Company,
Houston, Texas, U.S.A. Houston, Texas, U.S.A.

M. D. Sullivan G. N. Jensen
Chevron Energy Technology Company, ExxonMobil Production Company,
Houston, Texas, U.S.A. Houston, Texas, U.S.A.

D. C. Jennette S. J. Friedmann
Apache Oil, Houston, Texas, U.S.A. Lawrence Livermore Laboratory, Livermore,
California, U.S.A.
R. T. Beaubouef
ExxonMobil Exploration Company, D. C. Mohrig
Houston, Texas, U.S.A. Department of Earth, Atmospheric
and Planetary Sciences, Massachusetts
T. R. Garfield Institute of Technology, Boston,
ExxonMobil Production Company, Massachusetts, U.S.A.
Houston, Texas, U.S.A.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215880M883273

281
282 / Porter et al.

ABSTRACT

R
egional seismic mapping identified lower and middle Miocene slope
channels as significant exploration targets for Angola Block 15. Seventeen
exploration wells, followed by four appraisal wells, established these
slope-channel complexes as a world-class development opportunity. ExxonMo-
bil’s current development activity targets stacked turbidite-dominated reservoirs
in long-reach, high-angle wellbores tied back to tension leg platform (TLP) and
close-moored floating production, storage, and offshore loading facilities.
One of the major development targets in Block 15 is the Burdigalian-aged
(Bur1) slope-channel reservoirs. The Bur1 slope-channel system was one of many
lower Miocene sediment fairways that provided a mechanism for the delivery of
coarse-grained turbidites and mixed muddy and sandy debrites into the Lower
Congo basin. This slope-channel system traverses across the block in an east–west
direction and can be continuously mapped on adjacent seismic data sets across a
30–40-km (18–25-mi) reach. Three conventional cores and 28 well penetrations
calibrate updip to downdip changes in lithofacies type and channel architecture.
Map patterns of the Bur1 slope-channel system show distinctive changes in
sinuosity, channel confinement, and degree of amalgamation broadly related to
concurrent growth of salt-related structures. Channel-complex confinement is
more pronounced, and vertical amalgamation is better developed in segments
that cross structural highs. The Bur1 channel system shows weaker lateral
amalgamation, greater sinuosity, and less erosional confinement in structural
lows.
The episodic fill of the Bur1 slope-channel system can be better understood
by a hierarchical arrangement of unconformity-bounded stratal units. Within
these unconformity-bounded channel sets, nested channels form composite
channel complexes that show distinctive trends in lithofacies type and vertical
facies succession. Compared to other offshore Angola slope-channel systems,
this Bur1 system is noteworthy because of the relatively coarse granule to cobble
grain sizes encountered. Well logs and high-resolution seismic data calibrated to
conventional cores show that the lower parts of the channel complexes are
dominated by sandy-muddy debrites, slumps, and injected sandstones. These
facies are typically overlain by coarse-grained, gravelly, and well-amalgamated
sandy turbidites. The overlying facies succession is more variable, but commonly
consists of interbedded sandy and muddy turbidites, injected sandstones, and a
range of both muddy and sandy debrites.

BURDIGALIAN DEEP-WATER slope-channel systems (Bouchet et al., 2005), and the

CHANNEL SYSTEMS
Introduction
The past 5 yr have seen a remarkable expansion of
offshore exploration and development activity in
Angola. The highly productive hydrocarbon system
in the Lower Congo basin, coupled with large struc-
tures, well-established sediment fairways, and excel-
lent seismic coverage has made this part of west
Africa a world-class development opportunity. Total’s
Girassol field development showed the huge poten-
tial that exists for deep-water production from Angola’s
growth of ExxonMobil’s Kizomba development to
more than 550 thousand bbl/day of production in 3 yr
shows the prolific nature of deep-water depositional
systems that exist in Block 15 (Figure 1).
Seismic stratigraphy applied to conventional and
high-resolution three-dimensional (3-D) data sets of-
fered a compelling method to understanding the
Tertiary history of the Angolan continental margin
(Beaubouef et al., 1998; Schollnberger and Vail, 1999;
Gottschalk, 2002). Recent studies of the Zaire Fan
show a compelling view of successive fan development
driven by erosion, sediment bypass, and punctuated
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 283

FIGURE 1. Location map of the Bur1 slope-channel system, offshore Angola. The Bur1 channel system forms a part of
the greater Kizomba development. Water depth ranges from 750 to more than 1300 m (2460 to more than 4265 ft)
in the study area. A 50-km (31-mi) transect of the Bur1 slope-channel system is possible because of the contiguous
3-D seismic coverage in the 400-km2 (154-mi2) study area. The distribution of oil and gas fields in the Lower Congo basin is
shown by green polygons. The inset photo shows one of the two close-moored tension leg platform (TLP) and floating
production, storage, and offloading vessel (FPSO) facilities producing hydrocarbons from the Bur1 system.

deposition in the updip submarine canyon and levee-
channel system (Babonneau et al., 2002; Droz et al.,
2003). The Pleistocene to Holocene history of the
Zaire Fan shows a linked shelf-slope-basin deposi-
tional system and provides an analog for interpreting
the genetic packaging of complex deep-water depo-
sitional fan elements through sequence-stratigraphic
concepts (e.g., Garfield et al., 1998; Raposo and Sykes,
1998; Goulding et al., 2000; Temple and Broucke,
2004). For reservoir-scale studies, placement of sub-
seismic-scale stratigraphy in sinuous deep-water chan-
nel fills is critical in developing a depositional model
suitable for development drilling (e.g., Kolla et al.,
2001). A hierarchical approach to the physical stra-
tigraphy of deep-water systems allows the smallest
stratigraphic elements (beds) to be placed in in-
creasing larger composite stratal packages (complex
system sets) commonly mapped with seismic and
sequence-stratigraphic methods (Sprague et al., 2002,
2005).
This chapter shows the seismic expression, map
trends, and reservoir elements of a major lower Mio-
cene slope-channel system in offshore Angola. De-
velopment drilling reveals a complex, multistage
lithofacies succession of coarse- and fine-grained
turbidites and mixed debrites. These lithofacies form
predictable fining-upward channel fills in larger,
unconformity-bounded channel-complex sets that
can be mapped across tens of kilometers. Regional map-
ping of the Bur1 system shows the position and extent
of reservoir unit changes along depositional dip, and
these variations can be attributed to sea-floor gradient
changes associated with salt movement. A deposi-
tional model is developed from the superposition of
284 / Porter et al.

FIGURE 2. General stratigraphy of the Lower Congo basin. The Bur1 slope-channel system is contained within the
Oligocene – Miocene Malembo formation. The Malembo and the underlying Landana Formation contain the bulk of
postrift, sag-phase sedimentation in the Lower Congo basin. The lower-right diagram shows a generalized depositional
model of a linked shelf-slope-basin depositional system (e.g., Garfield et al., 1998). The Bur1 reservoirs discussed in
this chapter represent only a small segment of a very large, deep-water depositional system. The study area is
approximately located at cross section AA0 on this profile and represents a deep-water confined channel complex
system in a midslope depositional setting.

unconformity-bounded packages that shows the rela-
tionships of turbidites and debrites in a nested stratal
hierarchy of deep-water channel complexes.
GEOLOGIC SETTING
Deep-water slope-channel systems are one of the
important reservoir types in the hydrocarbon system
of the Lower Congo basin. The Lower Congo basin is
a Cretaceous to Holocene sedimentary basin devel-
oped on the west African continental margin. The
early history of the basin is represented by a thick,
Aptian-aged salt succession that formed over rift
basin deposits associated with the Early Cretaceous
breakup of Gondwana (Scrutton and Dingle, 1976;
Brice et al., 1982; Emery and Uchupi, 1984; Burke,
1996) (Figure 2). A postrift, early sag phase filling of
the basin is represented by marine deposits of the
Cretaceous Pinda Formation. Source rocks are mostly
in the Upper Cretaceous Iabe Formation, which over-
lies the Pinda, and these shale-dominated rocks ac-
cumulated under low sedimentation rates. Fully ma-
rine conditions developed with continued sea-floor
spreading of the south Atlantic, and this early Tertiary
drift phase is represented by marine shales, sandstones,
and carbonates of the Landana Formation. Middle to
late Tertiary basin fill is contained in the Oligocene
to Miocene Malembo formation. The Malembo for-
mation is a 6-km (3.7-mi)-thick, time-transgressive
lithostratigraphic unit that contains strata deposited
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 285

FIGURE 3. Seismic traverse through the Bur1 slope-channel system. The upper panel shows the seismic expression
and structural configuration of the lower and middle Miocene along BB0. The Bur1 channel system is marked by dashed
horizons along a 55-km (34-mi) transect. Some of the well penetrations in the Bur1 are shown with corresponding
log responses. The lower panel is an interval amplitude extraction of the Bur1 slope-channel system. Note the confinement
of the system near or at structural highs and more weakly confined, sinuous map pattern developed in the intervening
synclines. Well penetrations, including cored wells, are shown by symbols on the seismic amplitude map.

in nearshore to deep-water slope and basinal envi-
ronments (Temple and Broucke, 2004).
At least three major episodes of clastic input across
the slope and into the basin are recognized in the
Oligocene and Miocene of the basin (Ardill et al.,
2002). The Bur1 deep-water slope system discussed in
this chapter forms a part of the second major influx of
coarse-grained clastic sediments in the lower Mio-
cene. Biostratigraphic analysis of cores and cuttings
in exploration wells shows that this slope-channel
system is early Burdigalian in age. Nannofossil as-
semblages in bracketing hemipelagic slope mudstone
successions, tied to aged-dated European reference
sections, show the Bur1 slope-channel system devel-
oped ca. 20 Ma. Internal age subdivision of the Bur1
channel system is problematic because of a lack of
age-diagnostic faunal assemblages and multiple epi-
sodes of erosion, bypass, and deposition. Correlated
ages of capping mudstone sequences that bracket the
overall slope-channel system suggest about 2 m.y. of
deposition, ending at about 18 Ma.
Regional tectonic studies and structural analysis
of the Lower Congo basin have established a linked
extensional-contractional structural style for the An-
golan continental margin (Larson and Ladd, 1973;
Rabinowitz and LaBreque, 1979; Duval et al., 1992;
Hartman et al., 1998; Gottschalk, 2002). Structures
have developed as a result of salt-related deformation
associated with detachment on the Aptian salt. Updip
parts of the basin-margin slope are characterized by
extensional faults and tilted fault blocks, and downdip
sections show a full suite of salt structures, compres-
sional anticlines, and thrust faults (e.g., Gottschalk
et al., 2004).
THE BUR1 SLOPE-CHANNEL SYSTEM
Sinuosity and Width
A seismic amplitude extraction across the reservoir
interval shows the large-scale map pattern of the Bur1
slope-channel system (Figure 3). The Bur1 system shows
low to moderate sinuosity across a 40-km (25-mi)
286 / Porter et al.
reach and a well-developed east-to-west orientation.
This depositional trend is remarkably consistent, pre-
sumably related to the regional slope gradient into
the Lower Congo basin during deposition. Salt dia-
pirs, extensional faults, and compressional folds seg-
ment the slope profile into several large, north–
south-oriented anticlines and intervening synclines.
These structural elements, early in their growth his-
tory, clearly influenced the width of this deep-water
channel system. The Bur1 channel system is more
confined, narrow, and straight as it courses over struc-
tural highs (Figure 3). Widths of the confined channel
system are 700 –2200 m (2300 – 7200 ft) across posi-
tive structural elements. Channel system sinuosities,
measured from interpreted thalweg segments across
three anticlines in the study area, range between 1.06
and 1.15 (n = 5). In each intervening syncline, the
channel system shows less erosional confinement
and widths three to six times greater than over struc-
tural highs. Higher channel sinuosities (range 1.12 –
1.73, n = 5) are developed in these laterally expanded
parts of the Bur1 channel system (Figure 3).
System Depth and Relief of the Thalweg
High-resolution 3-D seismic data show that the
composite fill thickness of the Bur1 slope system
varies along a 40-km (25-mi) depositional fairway in
Block 15. This variation is a product of erosional and
depositional confinement of the channel complexes,
proximity to growing salt structures, and accommo-
dation provided by extensional fault networks. Av-
erage Bur1 thickness in the Kizomba development is
about 120 m (390 ft), with a range between 70 and
more than 220 m (230 and more than 720 ft). Paleo-
relief that developed on channel margins above the
channel thalweg can be estimated from the thickness
of channel fill, stratal onlap positions, dimensions of
rotated slump blocks and their related displacement
lengths, and cross sections from well-imaged, indi-
vidual channel profiles. These aggregate data types
suggest that the relief of Bur1 channels during depo-
sition never exceeded 50 m (164 ft), and in some of
the more weakly confined parts of the system, relief
may have been as little as 5–10 m (16–33 ft). Thus,
the Bur1 system is not a submarine canyon fill or the
product of a single erosional event followed by depo-
sition. Instead, the composite Bur1 sediment fill is a
product of multiple episodes of sediment gravity flow
scour, headward erosion, and channel-margin slump-
ing, alternating with episodic sedimentation by tur-
bidity and debris flows. Updip to downdip variations
in Bur1 channel system fill were controlled by vari-
able sediment influx to the basin, local sea-floor gra-
dients associated with active salt movement, and a
regional depositional gradient that crosses linked ex-
tensional and contractional tectonic domains of a
rapidly prograding shelf-slope system.
Levee Confinement
The Bur1 channel system is bounded by moderate-
to low-amplitude seismic reflections that dip away at
very low angles (<18) from the reservoir-prone axis
(Figure 4A). Downlap positions of these dipping strata
can be kilometers away from the trend of the channel
system (M. Grove, 2005, personal communication).
Well penetrations show that these seismic facies are
low-net-to-gross silty mudstones interpreted here as
flanking levees. Relief from the thalweg to levee crest
is on the order of 10–120 m (33–393 ft). Multiple epi-
sodes of levee growth and erosion are evident from the
stratal truncations observed on seismic traverses and
comparison of levee thickness with the laterally ad-
jacent, more erosionally bounded channel complexes.
Flanking levees are best developed at the base of the
channel system. Levee geometries are laterally more
restricted in the younger parts of the system and
show classic geometries well documented in Pleisto-
cene successions from the Amazon and Gulf of Mexico
(Pirmez et al., 1997).
Changes in Slope-channel System Confinement
across Structural Highs and Synclinal Lows
Erosional confinement of the Bur1 channel system
is most pronounced near or across the current posi-
tive structural elements. Channel segments that cross
positive structural elements are relatively straight in
plan view, and this straight geometry sharply con-
trasts with the more highly sinuous channel patterns
in the intervening synclines. Significant positive re-
lief at the sea floor was not a factor during Bur1 de-
position, as the Bur1 system crosses each of the salt-
cored anticlines or north – south extensional fault
systems at right angles, without channel diversion
or deflection. Net erosion of the sea floor by high-
energy turbidity currents and levee growth was suf-
ficient to maintain the channel system’s position and
gradient across growing positive tectonic elements.
Factors Controlling the External Form of the
Confined Channel System
Extensive 3-D seismic coverage shows that lower
Miocene slope-channel systems are large-scale de-
positional features that encompass more area than
the well-imaged, high-amplitude channel packages
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 287

FIGURE 4. Seismic cross section through the Bur1 slope-channel system. (A) The seismic expression of the Bur1 is varied
because of multiple rock types and a stratigraphy built up through multiple episodes of erosion and deposition. The axial
parts of the system are seismically characterized as multicyclic, high-amplitude, semicontinuous seismic facies. Off-axis
and channel-margin successions show moderate- to low-amplitude response, and these strata onlap the erosional boundary
of the system. An Earth model of mixed-impedance channel fills and nonchannelized depositional elements (B) is
convolved with a seismic wavelet to produce a modeled seismic response (C) that is similar to the actual seismic traverse.

commonly illustrated in reservoir studies. Lateral di-
mensions are large, on the order of 2 – 20 km (1.2 –
12 mi). These low-amplitude seismic facies represent
an integral component of the depositional system
and are fundamentally related to the genesis of the
associated channel belts. Establishing the exact chro-
nostratigraphic relationships between these deposits
and those of the channel fills is difficult, but it is clear
that they contain a stratigraphic record of protracted
overbank processes. Thickness patterns and geom-
etries suggest that these are net depositional units
formed via aggradation adjacent to evolving channel
complexes that were deposited from numerous tur-
bidity currents with strong lateral gradients in veloc-
ity, concentration, and depositional rates. Recogni-
tion of these units is important in understanding the
evolution of the slope-channel system because these
widely dispersed, low-net facies establish the long-
term channel system trend across the slope and into
the basin. The paucity of coarse-grained lithofacies
(i.e., coarse-grained sandstones, gravels, and conglom-
erates) in these flanking levee packages suggests large-
scale sediment bypass down the channel system fair-
way. Flow stripping of fine-gained suspension clouds,
sea-floor gradients modified by mass transport com-
plex emplacement, and shelf-edge control of sediment
delivery access points determine the offset spacing
of the channel system fairways.
KIZOMBA FIELD DEVELOPMENT
Field History
The early geoscience work and drilling history
leading up to the Kizomba development is presented
by Reeckmann et al. (2001). The phased development
288 / Porter et al.

FIGURE 5. Reservoir subdivision of the Bur1 slope-channel system. The Bur1 channel system can be subdivided into
four unconformity-bounded channel sets. This reservoir subdivision is based on the physical relationships of strata in
each channel set and contains lithofacies deposited in both channel and intrachannel areas. This approach avoids
overestimation of reservoir connectivity in assuming that all sandstones are time equivalent and recognizes lateral facies
changes into more shale-prone strata that baffle the reservoirs. Depth in meters.

consists of Kizomba A, a combination 36-slot TLP with
a close-moored, 2.2-MMBO capacity FPSO (floating
production, storage, and offloading vessel) (Figure 1).
First oil was produced in August 2004, and the facility
in mid 2006 is producing more than 250,000 bbl/day.
Drilling activity in Kizomba A continues, with pres-
sure support wells from five subsea drill centers. The
Kizomba B development, a virtual twin of Kizomba A,
began production in July 2005. In addition to res-
ervoirs accessible from the TLP, Kizomba B takes
production from the Dikanza subsea development
10 km (6 mi) to the northwest. Current production
from Kizomba B exceeds 250,000 bbl/day. Production-
focused activities in the Kizomba field include dril-
ling pressure support wells, establishing production
in unpenetrated compartments, and shooting a four-
dimensional, high-resolution seismic survey to man-
age field depletion.
Bur1 Well Penetrations in Block 15
A total of 28 wells penetrate or are completed in
the Bur1 deep-water channel system (Figure 3). More
than half of the wells target multitiered, gravelly
sandstone reservoirs in the Bur1 slope-channel sys-
tem. The high-angle, long-reach well design used in
Kizomba is optimized by volume-based interpre-
tation workflows in high-resolution 3-D seismic
cubes.
Distribution of Cored and Uncored Wells
Four conventional cores (554 m [1817 ft] total)
establish the vertical lithofacies arrangement in the
Bur1 channel-complex system. Three of the cored
wells establish a 27-km (16-mi) updip to downdip
transect of channel system sets in the Kizomba devel-
opment (Figure 5). The cores are described in detail,
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 289

FIGURE 6. Gravel-prone lithofacies in the Bur1 cores. Gravel-prone lithofacies in the Bur1 channel system are typically
thick bedded, moderate to poorly sorted bedsets bounded by sharp, erosive boundaries. The coarsest examples are R1
or R3 types that contain greater than 30% granule, pebble, or small cobble clasts. S1 and S3 bedsets contain 5 – 30%
granule to pebble clasts. Low-angle cross-stratification and clast imbrication are well developed in the R1 and S1 bedsets.
Color bar is 20 cm (8 in.) long.

and core results are used for seismic facies calibra-
tion, rock property studies, and inputs to geologic
modeling. Forty kilometers (25 mi) to the north, wells
in Xikomba field produce from a time-equivalent
but geographically distinct Burdigalian deep-water
channel system, and these cores show a comparative
channel-fill succession of turbidites and debrites.
DEEP-WATER LITHOFACIES IN CORED WELLS
Four facies associations are observed in cores from
the Bur1 channel-complex system. These facies asso-
ciations are distinct lithofacies successions of geneti-
cally related bedsets, each characterized by distinctive
sedimentary textures, bed thicknesses, and petrologic
constituents. All the deep-water facies associations in
the Bur1 slope system are common to other confined
channel systems studied in subsurface and outcrop
data sets (Raposo and Sykes, 1998; Sykes et al., 1998;
Campion et al., 2000).
Facies Association 1
This is a gravel-prone facies association consisting of
clast-supported sandy gravels that display massive,
ungraded to normally graded clast fabrics (Figures 6, 7).
Description
Bed thickness averages 1.3 m (4.2 ft), with a range
between 0.3 and 8 m (1 and 26 ft) thick. Gravels and
sandy conglomerates show a framework-supported
texture and contain greater than 30% granule- to
pebble-size clasts. Elongate pebble- to cobble-size
clasts show well-developed planar to inclined, imbri-
cated clasts between sharply erosional bed boundaries.
290 / Porter et al.

FIGURE 7. Gravel-prone, traction-dominated lithofacies associations. A vertical progression of lithofacies in gravel-prone,

traction-deposited bedsets consists of a lower part of very large shale rip-up clasts that are overlain by clast-supported
gravels and very coarse sandstones. These lithofacies associations are vertically well amalgamated and form the bases
of thick channel complexes as represented by a blocky well-log response. Core segment is 5 m (16 ft) long. Gamma-
ray (GR) and deep induction resistivity (ILD) log curves are depicted on the well log.

Very coarse sandstones are massive in texture and
marked by thin, basal accumulations of granule con-
glomerate with 5–30% granule and pebble clasts. Nor-
mal grading is common, and overall sorting is char-
acterized as poor to very poor. More than 80% of clasts
described in this facies association are well-rounded,
quartzofeldspathic, igneous, and schistose metamor-
phic clasts, with minor amounts of well-indurated
sandstone and metavolcanic lithologies. Isolated
calcite-cemented concretions show convex upper and
lower boundaries (Figure 8). Bed thickness, textures,
and grain-size trends compare favorably with coarse-
grained deposits categorized as R1 and R3 bed types in
Lowe (1982).
Interpretation
The gravel-rich beds in this facies association rep-
resent bed-load –dominated deposition from high-
concentration turbidity currents. The imbricated clast
fabric, mapped distribution in high-resolution seismic
data, and relation to channel-defined erosional cuts
suggest deposition as large gravelly bedforms (e.g.,
Winn and Dott, 1979). Gravel units developed during
maximum sediment throughput in active channel
thalwegs by the accumulation of coarse bed load from
multiple flow events.
Facies Association 2
This is a sand-prone facies association of thick-
bedded, well- and poor-sorted structureless to rippled
sandstones (Figure 9).
Description
Structureless sandstones are medium to very fine
grained, and bed thickness ranges between 0.06 and
0.6 m (0.2 and 2 ft) thick, with an average bed thick-
ness of 0.2 m (0.6 ft). Lower bed contacts are planar
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 291

FIGURE 8. Gravel-prone, traction-dominated lithofacies associations. Pebble- to granule-size lithofacies show massive
to weakly stratified textures in medium to thick (0.3 – 2-m; 1 – 6.6-ft) bedsets. This lithofacies association may contain
discrete, 0.5 – 1-m (1.6 – 3.3-ft) intervals of calcite cements. Convex, curved upper and lower boundaries defined by
the cementation front suggest an ellipsoidal, concretionary shape to these cements. Well-log response is blocky. Core
segment is 5 m (16 ft) long. Gamma-ray (GR) and deep induction resistivity (ILD) log curves are depicted on the well log.

and erosional; upper bed contacts are planar to gently
curved. Sedimentary fabric is structureless and well
sorted, although sparse swirled lamination is pres-
ent in thicker bedded examples. These beds typically
overlie the gravel-prone lithofacies. Current-rippled
sandstones include fine- to very fine-grained sand-
stone and interbedded silty mudstone. Bed thickness
ranges from 0.06 to 0.5 m (0.2 to 1.6 ft) and averages
0.2 m (0.6 ft) thick. This facies association commonly
caps the massive sandy turbidites, and contacts are
conformable and generally planar in core. Sedimen-
tary structures are dominated by 1–5-cm (0.4–2-in.)-
thick laminasets of current ripples. Minor amounts of
bioturbation and soft-sediment deformation are pres-
ent in the muddy sandstone beds.
A distinctive suite of more poorly sorted sand-
stones with a silty or muddy matrix is also developed
in the sand-prone facies association. These medium
to fine sandstones are 0.3 –1.5 m (1– 5 ft) thick, and
bed-planar bed contacts are well defined by erosion-
al truncation of dewatering and fluid-escape struc-
tures (Figure 9). Grain fabric is typically massive and
homogeneous. Most of these thick-bedded sand-
stones show a slight fining-upward grain-size trend.
Pebble-size floating shale clasts, swirled flow fabric,
and small granules are common components of these
beds. The macroscopic presence of clay-size matrix
is inferred by the abundance of vertical dewatering
pipes, swirled matrix flow structures, and floating,
matrix-supported clasts. Particle size analysis shows
1 – 8% clay content in the sandy matrix of these
bedsets.
Interpretation
The vertical facies transition from underlying sandy
conglomerates to these thick-bedded, well-sorted
sandstones suggest rapid suspension deposition by
sand-rich turbidity currents in a channel-axis setting.
Rapid deposition is suggested by the thick-bedding,
292 / Porter et al.

FIGURE 9. Sand-prone, suspension-dominated lithofacies associations. Well-amalgamated sandstones show massive

to normal grading and sharp, conformable bed contacts. Minor amounts of granule or shale clast-rich lags define the
bases of many bedsets in this lithofacies association. Muddy sandstones of debrite origin are developed locally in
the core suite. Core segment is 5 m (16 ft) long. Gamma-ray (GR) and deep induction resistivity (ILD) log curves are
depicted on the well log.

conformable bed contacts, and uniform grain-size dis- Facies Association 3

tribution (Arnott and Hand, 1989). Massive sandstones
are comparable to the Ta/Tb turbidite subdivision of
Bouma (1962).
In contrast to the well-sorted sandstones, mixed
grain sizes and sedimentary structures in more poorly
sorted, clay-bearing sandstone beds suggest a sandy
debris-flow origin. The dispersed clast fabric, meter-
scale bedding, and abundant fluid-escape structures
suggest that matrix strength was a profound control
on the movement and deposition of these sandy de-
brites (e.g., Middleton and Hampton, 1976; Iverson,
1997). The matrix strength imparted by the small
amount of clay dampened fluid turbulence and hin-
dered flow-related textural size segregation. Numeric
and physical process modeling suggests that some
thick-bedded sandstones previously described as the
product of amalgamated Ta beds may, in fact, be the
product of sandy debrite deposition (Angevine et al.,
1990; Shanmugam, 1996).
This is an interbedded sand and mud lithofacies
association. Massive, current-rippled, and laminat-
ed bed types are identified on the basis of inter-
nal sedimentary structures and presence of grading
(Figures 10–12).
Description
Interbedded sandy and muddy facies consist of
medium- to very fine-grained sandstone, siltstone,
and mudstone in fining-upward, 0.1 – 0.3-m (0.3 –
1-ft) beds. These beds are comparable to classic Bouma
(1962) bed types and include current-rippled sand-
stone, wavy to parallel-laminated sandy siltstone,
and parallel-laminated mudstone (Figure 10). Sand-
stones are typically very well sorted.
Structureless mudstones occur in 0.1 – 4-m (0.3 –
13-ft)-thick beds with planar to highly contorted bed
contacts. The matrix has a swirled or deformed ap-
pearance and commonly shows shear banding and
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 293

FIGURE 10. Spectrum of sandy and interbedded sandy-muddy lithofacies in Bur1 cores. This lithofacies suite spans a
complete range of sand-to-mud ratios and is best described as classic turbidites. Bed thickness is strongly related to sand
content. Well-amalgamated Ta and Tb beds are medium to thick bedded, whereas Tc , Tcd, and Tde are mud dominated
and thin bedded. Color bar is 20 cm (8 in.) long.

internal fabric variations on a centimeter scale. Float-
ing sand-size quartzose grains and larger organic detri-
tus are common in these mudstones (Figures 13, 14).
Interpretation
The current-rippled, heterolithic nature and di-
versity of lamina types suggest that these mud-prone
beds were deposited from suspension from turbidity
currents (Arnott and Hand, 1989). Bed fabric in muddy
turbidites succession is comparable to Bouma Tde beds
(e.g., Bouma, 1962). Rippled sandy turbidites were
deposited under waning-flow conditions in channel-
margin and proximal overbank settings. Dilute sedi-
ment plumes associated with each turbidity current
spread laterally from the axial parts of deep-water
channels. Accumulation and preservation of these
fine sandstones and mudstones were enhanced by
their off-axis position, removed from the erosive ef-
fects of subsequent high-concentration flows in the
channel thalwegs. It is likely that sediment gravity
flows varied substantially in strength, sediment vol-
ume, and texture during the active phase of channel-
complex development. This shale-prone succession
accumulated in the waning stages of channel fill and
abandonment, when channel avulsions likely changed
the locus of channel deposition to another part of the
slope complex.
Facies Association 4
This is an injected and slumped facies association.
294 / Porter et al.

FIGURE 11. Interbedded sandy-muddy lithofacies associations. Mud-prone lithofacies show parallel to wavy lamina sets
and a shaly log response in the intervening intervals between channelized reservoir packages. The core segment is
5 m (16 ft) long. Gamma-ray (GR) and deep induction resistivity (ILD) log curves are depicted on the well log.

Description clasts commonly show internal bedding or lamination

Concordant (sills) and discordant (dikes) sand-
stones occur in a wide variety of shapes, thicknesses,
and geometries in the Bur1 cores (Figure 15). Sand-
stones are well sorted, and grain size is very restricted,
in the very fine- to fine-sand range. The internal struc-
ture is commonly massive, and some of the thicker
discordant beds have small shale rip-up clasts at the
upper and lower contacts (Figure 16). Injected sand-
stone thickness ranges from millimeter scale to 0.75 m
(2.5 ft) in dikes, and 0.4–1 m (1.3–3.3 ft) thick in sills.
In some sections, more or less isolated sills of sand-
stone pass upsection into a complex network of ver-
tical and horizontal injected sandstones throughout
a 10–15-m (33–49-ft) interval.
Stratigraphically below coarse-grained reservoir sec-
tions, many cored intervals contain 0.5–5-m (1.6–
16-ft) blocks of folded and contorted sandstone and
mudstone blocks. Internal fabrics are always deformed
and never parallel to master bedding surfaces. Large
to be truncated at block margins. The blocks are em-
bedded in sandy or muddy matrix.
Interpretation
The injected sandstones are locally remobilized
sands derived from well-sorted, sandy bedforms. Sand-
stone dikes and sills in interchannel shales may be
related to sediment loading under and adjacent to
rapidly deposited gravelly and sandy bedforms. Later
sand remobilization into overbank and capping aban-
donment shale successions likely occurred during shal-
low burial of the sandy channel complexes. Lithology-
related permeability contrast between the sandy
channel fills and mud-prone overbank facies with
compaction led to overpressure in porous sands. Hydro-
static pressure release forced these well-sorted sands
into confining mudstones as a series of dikes and sills.
The origin of the slumped strata is more problem-
atic and may represent more than one depositional
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 295

FIGURE 12. Interbedded sandy and muddy lithofacies associations. Muddy debrites form the upper part of a well-
developed fining-upward well-log motif (arrow). The muddy debrites contain 2 – 20% sand in the clay-rich matrix. These
muddy debrites are overlain by thin-bedded Tc , Tcd, and Tde turbidites. Thickness of this lithofacies association is
commonly 2 – 8 m (6.6 – 26 ft). The core segment is 5 m (16 ft) long. Gamma-ray (GR) and deep induction resistivity (ILD)
log curves are depicted on the well log.

process and episode of deposition. Proximity of slump
deposits to major erosional surfaces implies channel
instability and undercutting by energetic turbidity
currents. Crosscutting relations of deformed strata
with injected sandstones suggest that slump blocks
failed from oversteepened cut banks and were buried
prior to compaction by deposition of bed-load –
dominated bar deposits.
FACIES ARCHITECTURE OF THE BUR1
DEEP-WATER CHANNEL SYSTEM
Vertical and Lateral Facies Stacking Arrangements
High-resolution seismic data sets and numerous
well penetrations establish some general trends of
vertical and lateral facies relationships in the Bur1
slope-channel system. In most areas where current
development activities are focused, gravel-prone litho-
facies are common in mapped channel-axis posi-
tions. The lower one third to one half of the channel
fill consists of thick-bedded gravels and very coarse
sandstones (Figure 17). Upper parts of the channel
fills consist of interbedded sand-prone turbidites and
mixed muddy and sandy debrites, thin-bedded muddy
turbidites, and injected sandstones. Multiple, stacked
channel-fill deposits can be grouped with laterally
associated channel-margin and overbank deposits to
form genetically distinct channel complexes. Seismic
mapping of erosion surfaces, together with corre-
lation of low-net facies related to channel-complex
abandonment and lateral migration, provides a four-
fold, unconformity-bounded subdivision of the Bur1
into channel-complex sets (e.g., Sprague et al., 2002)
(Figures 5, 18).
Coarse-grained Gravels in Confined
Channel Complexes
The gravel-prone facies succession shows a high-
impedance seismic character that allows detailed geo-
body extractions in the Bur1 channel system (Figure 19).
296 / Porter et al.

FIGURE 13. Spectrum of debrite types in the Bur1 cores. A complete spectrum of lithofacies with floating, outsized
lithoclasts, swirled or deformed primary bedding geometries, dewatering pipes, and interstitial clay is present in the
Bur1 system. Completely sand-prone or completely mud-prone examples occupy the ends of a true spectrum of bed
types and are interpreted here as sandy debrites and muddy debrites. Most of the debrites cored in the Bur1 fit into the
middle part of the bed-type spectrum and are pragmatically separated into sandy-muddy debrites (poorer reservoir
quality) and muddy-sandy debrites (better reservoir quality) on the basis of sand in the matrix, water saturation,
permeability values, and bed thickness. Color bar is 20 cm (8 in.) long.

Seismic facies mapping in high-resolution 3-D cubes
indicates these geobodies to be an excellent proxy
for the size, distribution, and shape of both traction-
deposited gravel bedforms and intervening intra-
channel units, such as thick- to thin-bedded turbidites,
and mixed sandy-muddy debrites (e.g., Goulding et al.,
2000; Beaubouef, 2004).
Horizon-keyed amplitude extractions along the
erosional base of the Bur1 channel system show in-
terpreted gravel bars and coarse-grained thalweg de-
posits to be elongate and paddle shaped (Figure 19).
Downflow-directed lengths are 5– 10 times longer
than corresponding widths. The long-axis orientation
of the gravel bars is never parallel to the thalweg po-
sition as determined from isochron maps. Instead,
gravel bars splay 30 –808 off the downcurrent flow
direction. Interpretation of seismic amplitudes in the
channel system suggests that gravel-prone units are
not uniformly deposited across the width of the Bur1
system. Strong stratigraphic control of these units
by channel migration and incision is inferred by
sinuous map patterns and cross-sectional geometries
in channel complexes. Although late-stage incision
or unrelated subsequent erosion by turbidity currents
cannot be discounted, the oblique orientation of the
preserved gravelly bars to the overall trend of the
sinuous channel suggests a primary, preserved depo-
sitional extent to these deposits.
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 297

FIGURE 14. Mud-prone lithofacies associations. Mixed muddy debrites form highly varied successions of thick- and
thin-bedded, contorted facies successions. Log response is serrate to bell shaped. Thin, interbedded Tc and Tcd turbidites
commonly punctuate the thicker debrite-rich intervals. Permeability contrast between turbidite-deposited sandstones
and muddy debrites is indicated by the degree of hydrocarbon saturation. The core segment is 5 m (16 ft) long. Gamma-
ray (GR) and deep induction resistivity (ILD) log curves are depicted on the well log.

Implications of the Gravel-prone Facies Successions
The abundance of gravel-prone lithofacies in chan-
nel complexes of the Bur1 system illustrates some im-
portant geologic features and processes that operate
within large slope-channel systems. These gravelly
sediments are clearly associated with channel erosion
and fill, organized in bedforms, and do not represent
eroded coarse-grained remnants of older strata. Frame-
work clast petrology is essentially that of Proterozoic
igneous and metamorphic rock suites of the Angolan
shield, and this implies an updip, direct shelfal input
of gravels or cannibalization of older, coarse-grained
shelf-margin fluvial-mouth bar deposits by shelf-
edge failure and headward erosion up the channel
systems. A relatively narrow shelf and Tertiary-aged
regional tectonic uplift were important factors in
keeping the lower Miocene slope-channel systems of
Angola charged with coarse-grained sediments (Brice
et al., 1982).
The lower Miocene shelf margin, identified on re-
gional seismic lines, was at least 40–50 km (25–31 mi)
to the east of the study area. The competency of trans-
porting sediment gravity flows was considerable.
Clast sizes of as much as 45 cm (17 in.) in diameter
are present in the cored suite, and pebbles to small
cobbles comprise most of the gravel-prone bedsets.
Larger framework clast sizes likely exist in these de-
posits but are not represented in the cored lithofacies
because of the mechanical difficulty of coring in largely
unconsolidated reservoirs by extended-reach wells.
Low-angle cross-stratification, imbricate clast fab-
rics, and grain-size trends suggest that the gravelly
units were deposited as traction carpets by high-
concentration turbidites (e.g., Winn and Dott, 1977;
Lowe, 1982). Effective transport of such coarse-grained
lithofacies more than 50 km (31 mi) from the time-
equivalent shelf margin implies tremendously energetic
sediment gravity flows into the Lower Congo basin.
298 / Porter et al.

FIGURE 15. Mud-prone lithofacies associations. Thick successions of muddy debrites are uniformly gray in color and
may show discrete, well-sorted disconformable intervals of fine-grained sandstone. Crosscutting relations of the
sandstone with the bedding geometry of muddy debrites suggest a postdepositional injection origin of the sandstones.
The shaly log response is dependent on the amount of injected sandstone; those intervals with abundant sandstone
injections show a highly serrate log pattern. Core segment is 5 m (16 ft) long. Gamma-ray (GR) and deep induction
resistivity (ILD) log curves are depicted on the well log.

Deposition of the gravel units was focused in the
deeper thalweg part of channel complexes. Erosional
bedset boundaries suggest episodic development of
the barforms, and numerous high-concentration
flows likely contributed to the volume and arrange-
ment of the gravels preserved in the channel fills.
Because the gravelly units may represent only a
small fraction of the total sediment load in large,
turbulent sediment gravity flows, the preservation of
thick, amalgamated gravelly packages implies that much
greater amounts of sand and mud have completely
bypassed the study area. It is likely that these missing
sediments built downdip depositional elements of the
large, linked deep-water slope and basin depositional
system. It is without a doubt that these depositional
elements of the lower slope and basin deep-water systems
will be better understood as exploration pushes into the
ultradeep parts of the Lower Congo basin.
Depositional History of the Bur1
Slope-channel System
The integration of high-resolution 3-D data, cores,
and well logs demonstrates a composite sedimentary
record for the Bur1 slope-channel system. Cored litho-
facies show both turbidity current and debris-flow
deposits in the composite fill of the channel system
(Figure 20). Multiple episodes of erosion and depo-
sition produced at least four unconformity-bounded
channel-complex sets over the span of 2 m.y. Each of
these channel sets shows both lateral and vertical
connectivity and can be subdivided into an axis
dominated by gravel-rich traction deposits, highly
amalgamated sandstones, and laterally flanking
strata, which contain thin-bedded sandstone and
mudstones, injected sandstones, and slumps depos-
ited in channel-margin, levee, and overbank areas of
the Bur1 slope-channel system.
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 299

FIGURE 16. Injected sandstone and debrite lithofacies associations. This segment shows some of the facies organization
at the base of the Bur1 system. The slope-channel system boundary (white dashed line) can be tied to seismic data
as a regional mapping horizon. The contact is sharp and uncorformably separates the Bur1 channel system from
older slope mudstones (gray lithologies). The lithofacies succession that sits directly on this unconformity is highly
variable. In this core, sandy debrites, slump blocks, and injected sandstones overlie the contact. Core segment is
5 m (16 ft) long. Gamma-ray (GR) and deep induction resistivity (ILD) log curves are depicted on the well log.

The general similarity of channel stacking pattern
and constituent channel-fill elements in a 40-km
(25-mi) reach of the Bur1 channel system suggests
that sediment gravity flow erosion and deposition
extend over long distances (tens to hundreds of kilo-
meters). The Bur1 succession in the Kizomba devel-
opment gives a glimpse into the variations of facies
architecture constructed by highly energetic and com-
petent sediment gravity flows traversing the mid-
slope of the Lower Congo basin.
ACKNOWLEDGMENTS
The authors thank ExxonMobil, together with An-
gola Block 15 coventurers British Petroleum, Statoil,
and ENI for permission to publish our work on this
lower Miocene deep-water system. We also express
our gratitude to Sociedade Nacional de Combustiveis
de Angola (Sonangol) for the opportunity to study
and develop these world-class reservoirs. Our work
includes substantial contributions from numerous
prospect and development geoscientists in the Ex-
ploration and Development companies of ExxonMo-
bil. We thank John Ardill, Dana Butters, Tim Fahrer,
E-Chien Foo, Matt Grove, Randy Kissling, Dave Mason,
Lisa McBee, Jose Sequeira, Mark Rosin, and Bill Tate
for their insight to the geology of Bur1 slope-channel
system. Reviews by Laurence Droz (University of Brest)
and Carlos Pirmez (Shell Oil) greatly improved the
manuscript. The views expressed in this study are
those of the authors and ExxonMobil and do not
necessarily reflect those of the concessionaire or the
Block 15 contractor group.
300 / Porter et al.

FIGURE 17. Depositional environments interpreted from the vertical succession of lithofacies association in the Bur1
channel system. Each of the channel complexes in the Bur1 shows a progression of mixed debrites, turbidites, and
slumps that are overlain by turbidite-dominated facies associations of well-amalgamated gravel- and sand-prone
traction deposits. Because of the admixture of muddy debrites in the channel complexes, log response would suggest a
nonamalgamated, disconnected prediction of channel-fill connectivity. Erosion associated with the gravelly lithofacies
produces lateral and vertical amalgamation of the channel complexes, and dynamic well performance demonstrates
better connectivity than would be inferred by log response. Gamma-ray and Vshale log response of stacked facies
associations is on the leftmost track with core photo annotations; resistivity, porosity, density, and water-saturation
curves are shown on the right log tracks. Depth in meters.
Lower Miocene Slope-channel System in Block 15, Offshore Angola / 301

FIGURE 18. Vertical lithofacies succession and stratal

hierarchy in the Bur1 slope-channel system. Detailed
facies description of Bur1 cores shows the relationship
of bed and bedset elements to larger stratal elements
in a deep-water hierarchy. In this example, yellow
bedsets are gravel- or sand-prone lithofacies deposited
by high- and low-concentration turbidity currents.
Brown-colored units are sandy and muddy sand de-
brites. Gray-colored lithologies are an admixture
of muddy turbidites, debrites, and hemipelagic shales.
The vertical facies log shows the axial expression of
three stacked channel complex sets in this part of the
Bur1 slope-channel system. Leftmost well-log curves
are gamma ray and Vshale ; shallow and deep resistivity
curves are displayed on the right track. The length
of the core description is 110 m (360 ft).
302 / Porter et al.

FIGURE 19. Seismic expression of gravel-prone units in the Bur1 slope-channel system. Four seismic horizon-keyed
extractions of the high-impedance, lower parts of Bur1 channel complexes show the geometry and dimensions of
gravel-rich units. (A) Scrollwork pattern of gravel-rich units deposited within sinuous channels of the Bur1 system. These
units represent the lower half of the composite channel complex fills and are overlain by sand- and mixed sand- and
mud-prone lithofacies. (B) Stratigraphically younger channel complex shows smaller proportion of gravel units in an
overall more sand-prone channel complex set. Scrollwork pattern is developed from laterally migrating channels, but
gravelly units are more isolated than in the channel complex set illustrated in (A). (C) Strongly confined part of the Bur1
system shows the abundance of gravel thalweg units in green. Well penetrations that establish intervening areas are
sandstones deposited by high-concentration turbidity currents. (D) Volume seismic sculpt shows relationship of gravels
with overlying sandstones in a single channel complex set of the Bur1. Well penetrations show blue units as high-
impedance gravels and the red units as lower impedance sandstones.
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 303

FIGURE 20. Depositional model for the Bur1 slope-channel system. The vertical and lateral relationships of turbidite
and debrite deposits are represented in this cross-sectional depositional model of the Bur1 confined slope-channel
system. Core-calibrated logs define the axial facies succession. Debrite and gravelly channel fills define the deepest parts
of erosionally based channel sets, marked by light-blue lines. The overall vertical succession shows most of the gravel-
prone lithofacies in the lower half of the channel system, and the upper fill succession contains more mixed sandstone,
mudstones, and muddy debrites. Seismic facies mapping is used to place depositional elements in the off-axis and
margin areas of the channel system. These lateral facies are thin-bedded mixed sandstones and mudstones deposited
in channel-related, stacked levee margins.
304 / Porter et al.
REFERENCES CITED
Angevine, C. L., P. L. Heller, and C. Paola, 1990, Quan-
titative sedimentary basin modeling: AAPG Continu-
ing Education Course Note Series 32, 133 p.
Ardill, J. A., T. C. Huang, and O. McLaughlin, 2002, The
stratigraphy of the Oligocene to Miocene Malembo
formation of the Lower Congo basin, offshore Angola
(abs.): AAPG Annual Meeting Expanded Abstracts,
p. 9.
Arnott, R. W. C., and B. M. Hand, 1989, Bedforms, pri-
mary structures and grain fabric in the presence of
suspended sediment rain: Journal of Sedimentary
Petrology, v. 59, p. 1062 – 1069.
Babonneau, N., B. Savoye, M. Cremer, and B. Klein, 2002,
Morphology and architecture of the present canyon
and channel system of the Zaire deep-sea fan: Marine
and Petroleum Geology, v. 19, p. 445 – 467.
Beaubouef, R. T., 2004, Deep-water leveed-channel com-
plexes of the Cerro Toro Formation, Upper Cretaceous,
southern Chile: AAPG Bulletin, v. 88, p. 1471 – 1500.
Beaubouef, R. T., T. R. Garfield, and F. J. Goulding, 1998,
Seismic stratigraphy of depositional sequences: High
resolution images from a passive margin slope setting,
offshore west Africa (abs.): AAPG Bulletin, v. 82, p. 1980.
Bouchet, R., B. Levallois, G. Mfonfu, and J.-F. Authier,
2005, Optimizing development of Angola’s Girassol
field: World Oil, v. 226, no. 3, p. 45 – 48.
Bouma, A. H., 1962, Sedimentology of some flysch de-
posits: Amsterdam, Elsevier, 162 p.
Brice, S. E., M. D. Cochran, G. Pardo, and A. D. Edwards,
1982, Tectonics and sedimentation of the south Atlan-
tic rift sequence, Cabinda, Angola, in J. S. Watkins and
C. L. Drake, eds., Studies in continental margin geology:
AAPG Memoir 34, p. 518.
Burke, K., 1996, The African plate: South African Journal
of Geology, v. 99, p. 341 – 409.
Campion, K. M., A. R. Sprague, D. Mohrig, R. W. Lovell,
P. A. Drzewiecki, M. D. Sullivan, J. A. Ardill, G. N.
Jensen, and D. K. Sickafoose, 2000, Outcrop expression
of confined channel complexes: Gulf Coast Section
SEPM Foundation 20th Annual Research Conference,
Houston, Texas, p. 127 – 150.
Droz, L., T. Marsset, H. Ondreas, M. Lopez, B. Savoye, and
F.-L. Spy-Anderson, 2003, Architecture of an active
mud-rich turbidite system: The Zaire Fan (Congo –
Angola margin southeast Atlantic): Results from Zai-
Ango 1 and 2 cruises, AAPG Bulletin, v. 87, p. 1145 –
1168.
Duval, B., C. Cramez, and M. P. A. Jackson, 1992, Raft
tectonics in the Kwanza Basin, Angola: Marine and
Petroleum Geology, v. 9, p. 389 – 404.
Emery, K. O., and E. Uchupi, 1984, The geology of the
Atlantic Ocean: New York, Springer-Verlag, 1050 p.
Garfield, T. R., D. C. Jennette, F. J. Goulding, and D. K.
Sickafoose, 1998, An integrated approach to deep-
water reservoir prediction (abs.): AAPG Bulletin, v. 83,
p. 1314.
Gottschalk, R. R., 2002, The Lower Congo basin, deep-
water Congo and Angola: A kinematically linked
extensional/contractional system (abs.): AAPG Bulle-
tin, v. 86, p. 66.
Gottschalk, R. R., A. V. Anderson, J. D. Walker, and J. C.
Da Silva, 2004, Modes of contractional salt tectonics
in Angola Block 33, Lower Congo basin, west Africa,
in 24th Annual Gulf Coast Section SEPM Research
Conference, Salt-sediment Interactions and Hydro-
carbon Prospectivity: Concepts, Applications, and
Case Studies for the 21st Century, Houston, Texas,
December 5 – 8, 2004, 30 p.
Goulding, F. J., T. R. Garfield, K. W. Rudolph, G. N. Jensen,
and R. T. Beaubouef, 2000, Seismic/sequence stratig-
raphy of deep-water reservoirs: 1. Seismic facies recog-
nition criteria: Past experience and new observations
(abs.): AAPG Annual Meeting Program, v. 9, p. A56.
Hartman, D. A., W. A. Swanson, P. R. Smith, F. J.
Goulding, and C. A. Kelly, 1998, Structural develop-
ment of the continental margin of Congo and north-
ern Angola (abs.): AAPG Bulletin, v. 82, p. 1923.
Iverson, R. M., 1997, The physics of debris flows: Reviews
in Geophysics, v. 35, p. 245 – 296.
Kolla, V., P. Bourges, J. M. Urruty, and P. Safa, 2001, Evo-
lution of deep-water Tertiary sinuous channels off-
shore Angola (west Africa) and implications for reser-
voir architecture: AAPG Bulletin, v. 85, p. 1371 – 1405.
Larson, R. L., and J. W. Ladd, 1973, Evidence from magnetic
lineations for the opening of the South Atlantic in the
Early Cretaceous: Nature, v. 246, p. 209 – 212.
Lowe, D. R., 1982, Sediment gravity flows: II. Depositional
models with special reference to the deposits of high-
concentration turbidity currents: Journal of Sedimen-
tary Petrology, v. 52, p. 279 – 297.
Middleton, G. V., and M. A. Hampton, 1976, Subaqueous
sediment transport and deposition by sediment
gravity flows, in D. J. Stanley and D. J. P. Swift, eds.,
Marine sediment transport and environmental man-
agement: New York, John Wiley and Sons, p. 197 –
218.
Pirmez, C., R. N. Hiscott, and J. D. Kronen Jr., 1997, Sandy
turbidite successions at the base of channel-levee
systems of the Amazon Fan revealed by FMS logs and
cores: Unraveling the facies architecture of large
submarine fans, in R. D. Flood, et al., eds., Proceedings
of the Ocean Drilling Program, Scientific Results,
v. 155, p. 7 – 33.
Rabinowitz, P. D., and J. LaBreque, 1979, The Mesozoic
South Atlantic Ocean and evolution of its continental
margins: Journal of Geophysical Research, v. 84, p. 5973–
6002.
Raposo, A. J. M., and M. A. Sykes, 1998, Exploration for
deep-water reservoirs, offshore Angola (abs.): AAPG
Bulletin, v. 82, p. 1956.
Reeckmann, S. A., D. K. S. Wilkin, and J. Flannery, 2001,
Kizomba, a deep-water giant field, Block 15, Angola,
in M. T. Halbouty, ed., Giant oil and gas fields of the
decade 1990 – 1999: AAPG Memoir 78, p. 227 – 236.
 Lower Miocene Slope-channel System in Block 15, Offshore Angola / 305
Schollnberger, E., and P. R. Vail, 1999, Seismic stratigraphy
of the Lower Congo, Kwanza, and Benguela basins, off-
shore Angola, Africa (abs.): AAPG Bulletin, v. 83,
p. 1338.
Scrutton, R. A., and R. V. Dingle, 1976, Observations on
the processes of sedimentary basin formation at the
margin of southern Africa: Tectonophysics, v. 36, p. 143–
156.
Shanmugam, G., 1996, High density turbidity currents:
Are they sandy debris flows?: Journal of Sedimentary
Research, v. 66, p. 2 – 10.
Sprague, A. R. et al., 2002, The physical stratigraphy of
deep-water strata: A hierarchical approach to the anal-
ysis and genetically related stratigraphic elements for
improved reservoir prediction (abs.): AAPG Annual
Meeting Expanded Abstracts, p. 167.
Sprague, A. R. et al., 2005, Integrated slope channel de-
positional models: The key to successful prediction of
reservoir presence and quality in offshore west Africa:
E-Exitep Proceedings 2005, Veracruz, Mexico, 13 p.
Sykes, M. A., D. Mohrig, C. Rossen, and T. R. Garfield, 1998,
Lithofacies associations within complex slope channel
reservoirs: Debrites and turbidites (abs.): AAPG Bulletin,
v. 82, p. 1973.
Temple, F., and O. Broucke, 2004, Sedimentological
models of the Oligocene and Miocene Malembo for-
mation in offshore Angola (Lower Congo basin) (abs.):
1st Nigerian Association of Petroleum Explorationists–
American Association of Petroleum Geologists West
Africa Deepwater Conference, Abuja, Nigeria, p. A46.
Winn, R. D., and R. H. Dott Jr., 1977, Large-scale traction
produced structures in deep-water fan-channel con-
glomerates in southern Chile, Geology, v. 5, p. 41 – 44.
Winn, R. D., and R. H. Dott Jr., 1979, Deep-water fan-
channel conglomerates of Late Cretaceous age, south-
ern Chile: Sedimentology, v. 26, p. 203 – 228.
 9
Dubois, M. K., A. P. Byrnes, G. C. Bohling, and J. H. Doveton, 2006,
Multiscale geologic and petrophysical modeling of the giant Hugoton
gas field (Permian), Kansas and Oklahoma, U.S.A., in P. M. Harris and
L. J. Weber, eds., Giant hydrocarbon reservoirs of the world: From rocks
to reservoir characterization and modeling: AAPG Memoir 88/SEPM
Special Publication, p. 307 – 353.

Multiscale Geologic and

Petrophysical Modeling of the
Giant Hugoton Gas Field
(Permian), Kansas and
Oklahoma, U.S.A.
Martin K. Dubois, Alan P. Byrnes, Geoffrey C. Bohling, and John H. Doveton
Kansas Geological Survey, University of Kansas, Lawrence, Kansas, U.S.A.

ABSTRACT

R
eservoir characterization and modeling from pore to field scale of the
Hugoton field (central United States) provide a comprehensive view of a
mature giant Permian gas system and aid in defining original gas in place
and the nature and distribution of gas saturation and reservoir properties. Both
the knowledge gained and the techniques and workflow employed have impli-
cations for understanding and modeling reservoir systems worldwide that have
similar geologic age and reservoir architecture (e.g., Gwahar and North fields, Persian
Gulf). The Kansas–Oklahoma part of the field has yielded 34 tcf (963 billion m3)
gas throughout a 70-yr period from more than 12,000 wells. Most remaining gas
is in lower permeability pay zones of the 557-ft (170-m)-thick, differentially de-
pleted, layered reservoir system.
The main pay zones represent 13 shoaling-upward, fourth-order marine-
continental cycles comprising thin-bedded (6.6–33-ft; 2–10-m), marine carbon-
ate mudstone to grainstone and siltstones to very fine sandstones and have re-
markable lateral continuity. The pay zones are separated by eolian and/or sabkha
red beds having low reservoir quality. Petrophysical properties vary among 11 major
lithofacies classes. Neural network procedures, stochastic modeling, and automation
facilitated building a detailed full-field three-dimensional (3-D) 108-million-cell
cellular reservoir model of the 10,000-mi2 (26,000-km2) area using a four-step
workflow: (1) define lithofacies in core and correlate to electric log curves (training
set); (2) train a neural network and predict lithofacies at noncored wells; (3) popu-
late a 3-D cellular model with lithofacies using stochastic methods; and (4) popu-
late model with lithofacies-specific petrophysical properties and fluid saturations.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215881M883274

307
308 / Dubois et al.
INTRODUCTION
The focus of this chapter is the definition and
integration of core-defined lithofacies, core-derived
petrophysical properties, wire-line-log response, and
estimation of reservoir properties to characterize a
giant reservoir system at the core, well, and field scale.
Central to the effort is the use of core-defined litho-
facies to train a neural network to predict lithofacies
at wells without core. We discuss each step of the
workflow, many aspects of which could be applied
in other settings. The results and summary are an early
view of work in progress that is part of an ongoing
collaborative project.
The importance of the Hugoton field study ex-
tends beyond the borders of Kansas and Oklahoma.
Both the knowledge gained and the techniques em-
ployed have implications for understanding and mod-
eling reservoir systems worldwide that have similar
geologic age, reservoir architecture, production char-
acteristics, problems in determining water saturation,
large data sets, split ownership, or maturity. The full-
field model of the 10,000-mi2 (26,000-km2) reservoir
area provides a detailed three-dimensional (3-D) view
of 13 shoaling-upward cycles vertically stacked in a
low-relief shelf setting. The nature of the model and
its construction provides a good analog for similar
thin, stacked-cycles reservoir systems, including the
Aneth field in the Paradox basin (Weber et al., 1994;
Grammer et al., 1996), fields in the prolific Permian
basin of west Texas (Dutton et al., 2005), and the
Khuff Formation in Gwahar and North fields in the
Persian Gulf (McGillivray and Husseini, 1992; Kon-
nert et al., 2001). Fine-scale cellular models are par-
ticularly important for modeling thin-layered, differ-
entially depleted reservoir systems, and methods used
in building the model demonstrate the construction
of a cellular petrophysical model for a giant field.
The study also demonstrates the benefits of pooling
proprietary geologic and engineering data in set-
tings having split ownership (Sorenson, 2005). As the
world’s fields mature, high-resolution modeling at
the full-field scale in data-rich environments will be-
come increasingly important, and we present a large-
scale example for developing such models.
Background
The combined Kansas Hugoton and Panoma, Texas
Hugoton, and West Panhandle fields, with an esti-
mated ultimate recovery of 75 tcf (2.1 trillion m3) gas
(Sorenson, 2005) represent the largest gas field in
North America. Covering southwest Kansas and parts
of the Oklahoma and Texas panhandles, these fields
are situated in the Hugoton embayment of the Ana-
darko basin (Figure 1). Since its discovery in 1922 and
development in the 1950s, 34 tcf gas (963 billion m3)
have been produced from greater than 12,000 wells
across 6200 mi2 (16,000 km3) in the Kansas and Okla-
homa part of the Hugoton field (Figure 2). Unless
otherwise noted, the term ‘‘Hugoton’’ in this chapter
combines the Hugoton (Kansas), Panoma (Kansas and
Oklahoma), and the Guymon-Hugoton (Oklahoma)
fields. Production is from the Lower Permian Chase
Group and Council Grove Group (Figure 3). In most
areas inside the Panoma boundary, the gas column is
continuous between the two stratigraphic intervals
(Pippin, 1970; Parhman and Campbell, 1993) and
reaches a maximum thickness of 500 ft (150 m) in the
west-central part of the study area. One exception
may be in a relatively small part of the field near the
western margin in Morton County, Kansas, that is
described by Olson et al. (1997) as being compart-
mentalized by faults. In Oklahoma, production out-
lined as ‘‘other Council Grove’’ (Figure 1) is from in-
tervals in the Council Grove that are as much as 300 ft
(100 m) below the lowest perforations in the Chase.
The reservoir is shallow, with depth to the top of the
Chase ranging from 2100 to 2800 ft (640 to 850 m)
and lower and upper productive limits, referenced to
sea level, of approximately +100 ft (+30 m) on the east
and +1250 ft (+380 m) on the western updip mar-
gin, respectively. Original wellhead shut-in pressure
in Kansas was 437 psi (3013 kPa) (Hemsell, 1939), sig-
nificantly less than half of a seawater gradient, and
similar, anomalous initial pressures were recorded in
Oklahoma (Sorenson, 2005). Average 72-h wellhead
shut-in pressure in Kansas in 2003 was 32 psi (221 kPa).
Annual production in 2004 was 265 bcf (7.5 billion m3).
Early completions in the Chase were commonly open
hole with a slotted liner followed by a large acid treat-
ment. After 1960, typical completions commonly in-
volve casing, perforation, and acidizing as many as
six zones separately, followed by a large hydraulic sand
fracture treatment, sometimes exceeding 200,000 lb
(91,000 kg) of sand, to the entire perforated interval
(Hecker et al., 1995).
Although much has been published on the Hugo-
ton throughout the 70-yr life of the field, most of the
studies were broad in scope (Hemsell, 1939; Mason,
1968; Pippin, 1970). Sorenson’s (2005) recent article
that presents a paleostructural and pressure history for
the reservoir system stretching from the Texas Pan-
handle to west-central Kansas provides a good recent
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 309

FIGURE 1. Regulatory bound-

aries for Permian (Wolfcam-
pian) gas and oil fields, Kansas,
Oklahoma, and Texas. Wolf-
campian structure in feet, after
Pippin (1970) and Sorenson
(2005).

homa Panhandle. Following

the Kansas Corporation Com-
mission proration order per-
mitting a second well in each
unit, several studies on res-
ervoir characterization (Sei-
mers and Ahr, 1990; Caldwell,
1991; Olson et al., 1997) and
reservoir simulation (Fetko-
vitch et al., 1994; Oberst et al.,
1994) were published. Past
studies by industry have been
generally confined to areas
where they have assets and
data. This chapter provides
details of the most compre-
hensive reservoir characteri-
zation effort to date, both
geographically and strati-
graphically (the entire reser-
voir system, Chase and Coun-
cil Grove groups).
Work related to this study
is part of a collaborative, mul-
tidisciplinary study of the
Hugoton field, supported by
10 industry partners having
gas resources in southwest
Kansas and the Oklahoma
Panhandle (Dubois et al.,
2005). The primary purpose
of the 2-yr study is to devel-
op a comprehensive, field-
wide geologic model that can
be used to quantify and locate
remaining gas for improved
reservoir management. No
overview of the field history and prior work. Detailed field-scale model exists primarily (1) because of the
studies involving reservoir characterization have been immense size of the field and (2) because the most
limited geographically and stratigraphically. For ex- critical petrophysical data were proprietary. The cur-
ample, Seimers and Ahr (1990) investigated the Chase rent project was facilitated by the creation of a co-
in the Oklahoma Panhandle; Olson et al. (1997) operative geologic and engineering data set amassed
studied the Kansas Chase; and Heyer (1999) focused from participating companies, which allowed a com-
on the Council Grove in a small area of the Okla- prehensive global view of the entire reservoir system.
310 / Dubois et al.
FIGURE 2. Gas production from the Wolfcampian (Hugo-
ton and Panoma fields) in the Kansas part of the study
area through 2004. The Oklahoma part of the study area
produced 7 tcf (198 billion m3) Wolfcampian gas in the
same period that 27 tcf (765 billion m3) was produced
in Kansas. The spike in production beginning in the early
1980s was caused by infill drilling the Hugoton field in
Kansas.
Although the field is very mature, individual com-
panies possess excellent modern wire-line-log, core,
and engineering data in their asset areas because of
drilling for deeper production and the infill drilling
program that took place in the late 1980s and early
1990s.
Problem and Approach
A principal goal of the project is to quantify the
nature and distribution of gas saturation and reservoir
properties and original and remaining gas in place.
Determining the gas in place and its distribution is
hampered by three significant obstacles: (1) accurate
determination of water saturations using conven-
tional wire-line logs is not possible because of deep
invasion by filtrate for typical drilling programs (Olson
et al., 1997; George et al., 2004); (2) wellhead shut-in
pressures are strongly influenced by high-permeability
interval properties and do not accurately represent all
interval pressures; and (3) the reservoir is layered,
differentially depleted, and pressure data for individ-
ual layers are minimal. To define the remaining gas
in place, we used an integrated approach using core,
core-derived lithofacies-specific petrophysical relation-
ships, and engineering data. Because petrophysical
property relationships (e.g., permeability-porosity,
capillary pressure) differ among lithofacies, the con-
struction of a geomodel with appropriate lithofacies
is critical. Because only a relatively small number of
wells were cored and the reservoir volume is im-
mense, a methodology was developed for predicting
lithofacies from wire-line-log response at wells not
cored and then between wells.
RESERVOIR GEOLOGY
Regional Geology
The Hugoton field lies on the west side of the Hugo-
ton embayment of the Anadarko basin and is bounded
to the northwest by the Las Animas arch and to the
northeast by the central Kansas uplift. The Anadarko
basin is an asymmetric foreland basin associated with
the early Pennsylvanian Ouchita–Marathon orogeny,
and the Hugoton embayment and the rest of the Kansas
shelf form the flatter side of the asymmetry (Figure 4).
The Anadarko basin was initiated and had its greatest
subsidence in Pennsylvanian-Morrowan time, with
subsidence rates decreasing through the Permian. The
basin was nearly filled by the end of the Wolfcampian
when the Anadarko basin was covered by shelf carbon-
ates (Kluth and Coney, 1981; Rascoe and Adler, 1983;
Kluth, 1986; Perry, 1989).
Marine carbonate reservoirs thin toward the up-
dip margin and many pinch out at, or just west of,
the field margin, particularly in the Council Grove.
FIGURE 3. Stratigraphic column, Hugoton field area, with
the names of gas fields in Kansas and Oklahoma adjacent
to the intervals from which they produce (compiled
from Zeller, 1968; Pippin, 1970; Baars, 1994; D. P. Merriam,
2006, personal communication). The combined Hugoton
and Panoma fields in Kansas and the Guymon-Hugoton
field in Oklahoma are lumped as Hugoton in this chapter.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 311
Red continental rocks, primarily very fine to coarse
siltstones, are thickest at the margin and thin basin-
ward across the shelf (Figure 5). The red beds have
been thought by many to be the lateral seal that,
when accompanied by a Leonardian-age evaporite
top seal, created a giant stratigraphic trap (Garlough
and Taylor, 1941; Mason, 1968; Pippin, 1970; Parh-
man and Campbell, 1993). However, laterally con-
tinuous, high-porosity and high-permeability, ma-
rine, and continental sandstones are common at the
updip margin in the northwest part of the field.
These rocks are gas productive inside the field bound-
aries and water saturated outside the field despite
being in a higher structural position and without
evidence of a physical barrier. These conditions argue
against these rocks being a lateral seal and suggest
that mechanisms other than lithofacies change alone
are responsible for trapping (Dubois and Goldstein,
2005).
Present-day structure of Wolfcampian-age rocks
was strongly influenced by a Laramide-age eastward
tilt (Figure 6), whereas the Wolfcampian isopach
(Figure 7) better reflects the shelf geometry. From the
west field margin, the Wolfcampian strata thicken
FIGURE 4. Distribution of major litho-
facies in the mid-continent during
the late Wolfcampian (modified from
Rascoe, 1968; Rascoe and Adler, 1983;
Sorenson, 2005; used with permission
from the Rocky Mountain Association
of Geologists and AAPG). Approximate
paleolatitude was 38N (Scotese, 2004).
basinward at a rate of approximately
1.2 ft/mi (0.24 m/km) to location on
the shelf where the rate of thicken-
ing increases by a factor of 10. The
axis of thickening is coincident with
an area of present-day steep dip and
may mark a shelf margin or the
axis of a steepened slope. It is also
nearly coincident with the edge of
a Virgillian-age starved basin and
the transition from marine carbon-
ate to marine shale (Rascoe, 1968;
Rascoe and Adler, 1983). Dubois and
Goldstein (2005) estimated the max-
imum relief across the Kansas part of
the shelf during Council Grove de-
position to have been 100 ft (30 m),
with a slope of approximately 1 ft/mi
(0.2 m/km). Notable is the absence
on the Hugoton shelf of dark, fissile shale, a common
deep-water lithofacies in the Wolfcampian in out-
crop in eastern Kansas and northeast Oklahoma (Maz-
zullo et al., 1995; Boardman and Nestell, 2000), sug-
gesting that water depths on the Hugoton shelf were
less than those at the present-day outcrop 300 mi
(480 km) to the east. The closest equivalent to the typ-
ical deep-water lithofacies in Hugoton core are dark,
marine siltstones found near the base of the marine
carbonate intervals in four cycles, the Grenola (C_LM),
Funston (A1_LM), Wreford, and Ft. Riley. For this chap-
ter, we will refer to most of the extremely gently slop-
ing area as shelf and the area of steeper dip and strati-
graphic thickening as the shelf margin.
Reservoir Lithofacies
The Hugoton in Kansas and Oklahoma produces gas
from 13 fourth-order marine-continental (carbonate-
siliciclastic) sedimentary cycles (Figure 8), six in the
Chase and seven in the Council Grove, reflecting rapid
glacioeustatic sea level fluctuations (Olson et al.,
1997; Heyer, 1999; Boardman et al., 2000; Olszewski
and Patzkowsky, 2003). The marine and continen-
tal lithologic units are laterally continuous and can
 312 / Dubois et al.

FIGURE 5. Stratigraphic cross section of the Chase and Council Grove groups with the top of the Council Grove as the datum. At the wells, lumped lithofacies
TM
are those predicted by neural network models (small well symbols) or from core (large symbols) and are interpolated in Geoplus Petra between wells. The
lumped lithofacies include continental sandstone (L0), continental siltstone (L1-2), mud-supported carbonate and marine siltstone (L3-5), grain-supported
carbonate and dolomite (L6-9), and marine sandstone (L10). The Council Grove is thinnest at a midfield (midshelf) position. Log curves are gamma ray (left)
and corrected porosity (right).
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 313

FIGURE 6. (A) The present-day structure of the top of the Wolfcampian reservoir (top of Chase) is mostly a function
of eastward tilt during the Laramide orogeny. Note the shelf margin or area of steepened slope at the southeast margin
of the Hugoton field outline. (B) Three-dimensional view of the same area. The present-day structure on the top of
the Chase and base of Council Grove.

be traced across the shallow shelf to the outcrop in
eastern Kansas. The main pay zones are 13 thin
(mean thickness 6– 70 ft [2 –21 m]) marine, mainly
carbonate intervals, deposited during sea level high-
stands. Pay zones are separated by continental, mainly
siltstone (red-bed) intervals (mean thickness 6–25 ft
[2 – 8 m]) deposited during sea level lowstands, when
most of the shelf was exposed. The siltstones gen-
erally have poor reservoir quality and vertically iso-
lated, or restricted communication among, the 13 pay
intervals (Seimers and Ahr, 1990; Oberst et al., 1994;
Ryan et al., 1994; Olson et al., 1997). The principal
factor in determining the reservoir storage and flow
capacity (hydrocarbon pore volume and permeabil-
ity) of Hugoton reservoir rock is primary deposi-
tional texture. Diagenesis, both early and after burial,
including leaching of grains and cements and early
and late dolomitization, are important factors in en-
hancing or reducing porosity (Seimers and Ahr, 1990;
Olson et al., 1997; Luczaj and Goldstein, 2000). How-
ever, the dominant reservoir rocks are marine car-
bonate with grain-supported textures and, to a lesser
extent, siliciclastic sandstone (Seimers and Ahr, 1990;
Caldwell, 1991; Olson et al., 1997; Heyer, 1999; Dubois
et al., 2003).
Depositional Model
Climate, shelf geometry, and glacially forced sea
level changes all influenced sediment supply, depo-
sitional patterns, accommodation, and stabilization
of both the marine and continental sediments. The
Hugoton shelf was near the paleoequator (Figure 4),
and monsoonal climate conditions are likely to have
prevailed at the annual scale (Parrish and Peterson,
1988). Generally, arid conditions accompanied gla-
cially induced sea level lowstands with more humid
conditions and high sea level present during inter-
glacial periods (Rankey, 1997; Soreghan, 2002). Pre-
vailing winds are thought to have been from the
present-day west during winter and east during sum-
mer (Parrish and Peterson, 1988). Extremely low re-
lief enabled rapid migration of the shoreline posi-
tion and changes in shelf hydrodynamics, with only
314 / Dubois et al.
FIGURE 7. Isopach of the Wolfcampian reservoir (top of
Chase Group to base of Grenola Limestone, Council Grove
Group). Wolfcampian rate of thickening increases by a
factor of 10 at the shelf margin.
minimal absolute changes in sea level (as little as
100 ft [30 m]). These conditions set the stage for
the vertical succession of lithofacies repeated from
one sedimentary cycle to the next, as well as the re-
markable lateral continuity of thin lithofacies units
in each cycle.
The cyclical nature of the Council Grove and
Chase is widely recognized (Seimers and Ahr, 1990;
Caldwell, 1991; Puckette et al., 1995; Olson et al.,
1997). Vertical succession of lithofacies in a shoaling-
upward pattern in both the Council Grove and
Chase (Figure 9) is a result of depositional environ-
ments changing across the shelf in response to
rapid sea level fluctuation. Differences in the style
(symmetry) and pattern (lithofacies) among the Chase
cycles recognized by Olson et al. (1997) are confirmed,
for the most part, in our study. Exceptions are that we
do see fine-grained sandstone of marginal-marine
origin at both the top and base of the Towanda and
Winfield near the updip margin of the field, although
it is more common to find the situation as they de-
picted elsewhere (sandstone at the base of the Towan-
da and top of the Winfield). Although similar in many
respects, the Council Grove cycles are typically more
asymmetric than the Chase cycles and tend to have
better developed, thin, packstone-grainstone litho-
facies at the base of the marine half-cycle. Figure 10
presents a composite of the vertical distribution of
lithofacies from model node wells that also help il-
lustrate the difference in symmetry between a Coun-
cil Grove and Chase cycle.
Intervals were defined within a simple cyclic frame-
work instead of a sequence-stratigraphic framework.
Existing formation or member tops that are half-
cycle boundaries between marine and continental
intervals represent a sequence boundary and flooding
surface. Because the transgressive systems tract (flood-
ing surface to maximum flooding) is relatively thin
and consistent in most of the cycles, little is gained by
correlation of an additional surface for sequence-
stratigraphic classification.
Idealized depositional models (Figure 11) for the
Council Grove Chase and Chase are generally sim-
ilar, but differences exist because of gradual changes
in climate, ambient sea level position, and sea level
fluctuation rate. Differences may be related to a shift
from more icehouse to more greenhouse conditions in
the Permian (Parrish, 1995; Olszewski and Patzkowsky,
2003). The entire Hugoton shelf was above sea level
during maximum lowstand for all studied Council
Grove cycles, and continental red-bed siliciclastics
accumulated and were stabilized by vegetation and
built relief preferentially near the field’s west updip
margin (Dubois and Goldstein, 2005). Accommoda-
tion for the carbonate sediments of the overlying
marine half-cycle was reduced, leading to nondepo-
sition, or pinch-outs, of several marine intervals in
the Council Grove at that position.
At the end of each lowstand, a relatively rapid sea
level rise resulted in deposition of a thin (1 –4-ft; 0.3–
1.2-m) transgressive carbonate-siliciclastic interval
at the base of each marine half-cycle. Only in the
Funston (A1-LM) and Neva (C-LM) cycles are well-
developed marine siliciclastics (shaly siltstone) de-
posited during maximum flooding. After maximum
flooding, shallowing, accompanied by conditions that
fostered increased carbonate production, resulted in
a shoaling-upward lithofacies stacking pattern. A fall
in absolute sea level caused progradation of broad
facies belts (e.g., carbonate sand shoals), resulting in
laterally extensive lithofacies bodies. With continued
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 315
sea level fall, continental sabkha, coastal-plain, and
savannah environments followed the retreating shore-
line and covered the carbonate surface. Evidence for
prolonged direct subaerial exposure and erosion of
the carbonate surface is absent in all seven Council
Grove cycles in the nine cores examined. Instead of
calcretes, microkarst, erosion, or other indicators of
prolonged exposure in the upper part of the marine
carbonate, there is a vertical succession of lithofa-
cies that suggests continuous sedimentation that
accompanied a sea level fall and withdrawal: sub-
tidal carbonate, tidal-flat carbonate, red siltstone
and muddy siltstone with anhydrite (sabkha), and
finally, red siltstone with paleosols (coastal plain or
savannah).
Although Chase deposition was similarly influ-
enced by absolute sea level, it differs from the Council
Grove in significant ways. Specifically, during Chase
lowstand, the lateral extent of subaerial exposure on
the shelf was generally more limited, and in some
continental intervals, tidal-flat siltstone and very
fine-grained sandstone are prevalent, particularly
in positions lower on the shelf. Fine-grained eolian
sandstone present in nearly all continental half-cycles
in the Council Grove is nearly absent in the Chase.
Marine transgressions in the Chase generally extend-
ed farther landward than they did during Council
Grove deposition, with marine sediments extending
beyond the updip margin of the field in all six cycles,
whereas they pinch out in four of the seven Council
Grove cycles studied. During the maximum high-
stand and the subsequent fall in sea level, carbonate
sand shoals of the Chase tend to be coarser grained,
and constituents include bioclasts and, occasionally,
ooids, instead of oncoids and peloids (in the Council
Grove), indicating more open-marine conditions.
Another significant difference between the Chase
and Council Grove is the presence of fine-grained
sandstone deposited in tidal-flat and marginal-marine
FIGURE 8. Formation- and member-level stratigraphy
correlated to the wire-line well log in the A-1 Flower well,
Stevens County, Kansas. Commonly used formation and
member letter-number combinations are shown for the
Council Grove. Twelve of the thirteen marine-continental
(carbonate-siliciclastic) sedimentary cycles that are gas
productive are shown (Grenola Limestone, C_LM is not
logged). Stratigraphic names that include ‘‘limestone’’
are marine half-cycles when combined with an adjacent
continental half-cycle, intervals with stratigraphic names
that include ‘‘shale’’ form a complete cycle. Color-coded
lithofacies are derived from core. Three were deposited
in a continental setting (L0 = sandstone; L1 = coarse
siltstone; and L2 = shaly siltstone) and eight in a marine
environment (L3 = siltstone; L4 = carbonate mudstone;
L5 = wackestone; L6 = very fine-crystalline dolomite; L7 =
packstone; L8 = grainstone and phylloid algal bafflestone;
L9 = fine-medium-crystalline moldic dolomite; and L10 =
sandstone). L0 is absent in this well. Wire-line-log abbre-
viations are caliper (CALI), gamma ray (GR), corrected
porosity (PHI_GM3), photoelectric effect (PEF), density po-
rosity (DPHI), neutron porosity (NPHI), core permeability
(K_MAX), and core porosity (CORE_POR). Logged interval
is 520 ft (160 m).
316 / Dubois et al.
settings at either the top, base, or top and base of all
cycles above the Fort Riley in the northwest part of
the field (Winters et al., 2005).
In all Chase or Council Grove intervals, the nature
of the lithofacies present in a succession are a func-
tion of the position on the shelf: the farther west
and updip, the greater the volume of siliciclastics,
whether in the marine or continental setting. Ma-
rine carbonate tends to be muddier to the west and
northwest, and grain-supported carbonate tends to
be finer toward the west, with the dominant grains
being hardened pellets (round, very fine grained,
micritic) and peloids (subrounded, fine grained, mi-
critic) instead of oncoids, bioclasts, or ooids (found
in upper Chase only). Marine environments become
more restricted in a westerly direction, and rocks
FIGURE 9. Idealized Chase and Council
Grove Groups cycles. Chase cycles are
from Olson et al. (1997), used with
permission from the AAPG, and our
Council Grove cycles are similarly for-
matted. One exception is that we extend
the cycle and approximate sea level
curve through the continental half-cycle
based on earlier work (Dubois and
Goldstein, 2005). Five cycle types dis-
tinguished the basis of lithofacies
stacking pattern and inferred relative
sea level curve.
with normal-marine assemblages
are absent in most of the Council
Grove cycles at or near the west
margin of the field. Both the Chase
and Council Grove cycles exhibit
gradual changes through time that
may be related to third-order cy-
clicity (Boardman et al., 2000) and
the overall shift from icehouse to
greenhouse conditions that began
in late Pennsylvanian and contin-
ued until the end of glacial condi-
tions in the Permian (Parrish, 1995).
Most likely, as a consequence of the
climate change trend, Chase ma-
rine carbonate intervals tend to be
three to five times thicker than their
Council Grove counterparts, at least
in the Crouse through Cottonwood
interval (B1-LM–B5-LM) on the Hu-
goton shelf.
Layered Reservoir and Differential Depletion
The Hugoton and Panoma fields appear to be
one large reservoir system that may have filled and
changed pressure in stages (Sorenson, 2005). How-
ever, the six pay zones in the Chase are being de-
pleted by production at different rates, as indicated
by different interval pressure tests and reservoir simu-
lation (Fetkovich et al., 1994; Oberst et al., 1994; Ryan
et al., 1994). Table 1 lists zonal pressures showing
differential depletion in both the Chase and Council
Grove. Composite tests covering multiple zones ex-
hibit pressures that primarily represent the produced
zone(s) having the lowest pressure and the highest
permeability as measured from core and drillstem test
(DST) analysis. The more recent of the wells tested
(2005) showed pressures as low as 19 psi (131 kPa) in
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 317
the main pay zones, the Herington and Krider half-
cycles. This explains why comingled wellhead shut-
in pressures (32 psi [220 kPa] field average in Kansas)
do not provide the necessary data for determination
of overall depletion.
STATIC MODEL WORKFLOW
Although many details exist, the workflow for
developing the Hugoton geomodel can be described
simply in four steps: (1) define interval tops from
logs, lithofacies in core, petrophysical properties for
lithofacies, and accurate log correction algorithms
to obtain a log database; (2) train a neural network
and predict lithofacies at node wells; (3) populate a
3-D cellular model with lithofacies and porosity;
FIGURE 10. Vertical histograms
showing the average relative
distribution of lithofacies in two
Wolfcampian marine half-cycles
from wells having predicted
lithofacies data (node wells).
Data for the Crouse (B1_LM),
Council Grove Group, are from
1146 wells, and for the Krider,
Chase Group, data are from
1069 wells. Histograms and
probabilities demonstrate the
difference in symmetry in ver-
tical lithofacies distribution be-
tween the Chase and Council
Grove. Probability distributions
were used to condition litho-
facies modeling by sequential
indicator simulation between
node wells. Layer annotations
refer to layering in the half-
cycle respective models (dis-
cussed later). Abbreviations are
fine-grained (Fg), fine crystal-
line (Fxln), and fine to medium
crystalline (F-mxln).
using stochastic methods
and (4) populate model with
lithofacies-specific petro-
physical properties and fluid
saturations. Figure 12 pro-
vides a more comprehensive
overview. The following is a
discussion of each of the steps
for integrating core-defined
lithofacies, core-derived pe-
trophysical properties, wire-
line-log response, and prediction of reservoir prop-
erties to characterize the reservoir system at the core
and well scale and at the field scale.
Lithofacies Classification
Because petrophysical properties vary among litho-
facies, fundamental to reservoir characterization and
the construction of a cellular reservoir model is the
population of cells with lithofacies. Determining the
number of lithofacies classes and the criteria for de-
fining classes involved four standards: (1) maximum
number of lithofacies recognizable by neural net-
works using petrophysical wire-line-log curves and
other variables; (2) minimum number of lithofacies
needed to accurately represent lithologic and petro-
physical heterogeneity; (3) maximum distinction
318 / Dubois et al.

FIGURE 11. Idealized depositional models for the Council Grove (A) and Chase (B) showing the distribution of dominant
lithofacies on the Hugoton shelf. Depicted are approximate depositional environments and associated lithofacies for
typical Chase and Council Grove cycles at maximum sea level lowstand and during the falling sea level stage of the
marine highstand.

Table 1. Pressures by zone for two relatively closely spaced wells.*

Group Zone 1994 Science Well 2005 Replacement Well

TM
DST-Sip psi (kPa) Composite psi (kPa) XPT -SIP* psi (kPa)

Chase Group Herington 120 (830) 104 (720) 19 (130)

Krider 88 (610) 104 (720) 21 (145)
104 (720) 30 (210)
Winfield SS 105 (720) 104 (720) 141 (970)
Winfield LS 121 (830) 104 (720) 217 (1500)
Towanda 230 (1590) 104 (720) 165 (1140)
U. Ft. Riley >400 (2750) 104 (720) 192 (1320)
Florence 398 (2740) 104 (720) 265 (1830)
Wreford 372 (2570) 104 (720) 219 (1510)
Council Grove Group A1_LM 400 (2760) 156 (1080) nt
B1_LM 350 (2410) 156 (1080) nt
B2_LM 131 (900) 156 (1080) nt
B3_LM 368 (2540) 156 (1080) 386 (2660)
B4_LM 215 (1480) 156 (1080) nt
B5_LM 160 (1100) 156 (1080) 348 (2400)

*The well drilled in 1994 was a research well (Flower A1, Figure 8), drilled with a foam fluid to limit filtrate invasion and formation damage.
Pressures are 24-h shut-in pressures from drillstem tests. The well drilled in 2005 is located 6 mi (10 km) north of the earlier well, and pressures
were recorded in the open hole by Schlumberger’s XPT TM tool, a repeat formation tester.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 319

FIGURE 12. Workflow for field-scale Hugoton model. Workflow can be divided into three broad tasks: (1) gather and
qualify data; (2) process data to provide basic geomodel input files (Develop/Define/Properties/Algorithms); and
(3) build the geomodel. The figure suggests the process is linear, whereas in reality, there are more feedback loops,
multiple iterations at subtask level, and testing and validation at smaller scales.
320 / Dubois et al.
of core petrophysical properties among classes; and
4) the relative contribution of a lithofacies class to
storage and flow. An optimal solution using these
criteria resulted in 11 lithofacies distinguished on
the basis of rock type (siliciclastic or carbonate),
texture (Folk, 1954, grain size for siliciclastics; Dun-
ham, 1962, classification for carbonates), and princi-
pal pore size (visual estimate). In classifying dolomite
rocks, we did not consider depositional texture, but
instead, the present texture and pore size that is pri-
marily a function of crystal size and the presence or
lack of molds of leached carbonate grains. Classes
based on differences in core petrophysical properties
coincided well with major lithofacies classes of rocks
and have fairly distinctive wire-line-log response to
petrophysical properties, the principal variables used
for neural network prediction of lithofacies. Although
defining more classes might have improved petro-
physical prediction accuracy, the inability of neural
networks to effectively recognize and distinguish finer
lithofacies classes discouraged finer class distinctions
(e.g., discriminating between fine-grained packstone
and coarse-grained packstone).
A quantitative, digital lithofacies description sys-
tem (Table 2) was used in describing core at 0.5-ft
(0.15-m) intervals. Three of the five factors illustrated
in the tables were sufficient to segregate lithofacies
classes, although other digits were considered initial-
ly in the process of determining class boundaries. For
each interval, as much as 12 variables were recorded.
Factors recorded in addition to those in Table 2
included the degree of consolidation and fracturing,
subsidiary pore size, cement mineralogy, bedding,
water depth, faunal assemblage, and color. Classify-
ing lithofacies in a digital form facilitated changes in
classification criteria and correlation of lithofacies
with core and log petrophysical properties involved
in the iterative process of determining optimal litho-
facies class boundaries. This digital system is designed
to provide a continuous numerical classification that
corresponds to the continuum in lithological and petro-
physical properties. In using this system, instead of a
mnemonic system, error in classifying a given sam-
ple is generally only one class up or down, and there-
fore, the predicted property values are within a class
step up or down from the true value. Once an object
is numerically classified, mapping to alternate clas-
sification schemes can be performed automatically.
Fourteen of approximately one hundred contin-
uous cores were selected for lithofacies analysis on
the basis of length (longest selected), geographic po-
sition (sampling distribution), and availability of core
analysis and wire-line-log data (Figure 13). In most
cases, selected cores included either the entire Chase
(five) or Council Grove (six) interval or covered both
intervals (three). Examples of the 11 major lithofacies
classes are shown in Figures 14 and 15. Common sub-
classes are represented, but the examples illustrated
do not show the range exhibited by a lithofacies. For
example, the marine carbonate packstone, packstone-
grainstone class (L7), includes rocks having a variety
of principal grain types and grain size but were de-
posited in a variety of environments (e.g., fine-grained
pellets = tidal flat and lagoon; peloid and oncoids =
restricted shelf and shoals; bioclasts = open shelf and
shoals).
Relative proportions of the 11 lithofacies in 4250 ft
(1295 m) of core described in this study are shown in
Figure 16. Continental lithofacies comprising fine-
grained sandstones (L0), coarse siltstones (L1), and
fine or shaly siltstones (L2) represent 42% of the rock
volume, whereas marine carbonates and marine sili-
ciclastics represents 45 and 13%, respectively. Litho-
facies with the greatest storage and flow capacity
in marine rocks include L6 to L10, consistent with
the principal reservoir lithofacies defined in previ-
ous studies (Siemers and Ahr, 1990; Olson et al.,
1996; Heyer, 1999). These lithofacies represent 31%
of the rock volume (Figure 16) and include very fine-
crystalline dolomite (L6), fine-medium-crystalline
dolomite (L9) with grain-moldic porosity, and litho-
facies with grain-supported texture, including pack-
stone (L7), grainstone (L8), and marine sandstone
(L10). Continental, fine-grained sandstones represent
only 6% of the rock volume, but are important to flow
and storage in the Council Grove. All 11 lithofacies
are present in the Chase Group, but continental sand-
stones and very fine-crystalline, sucrosic dolomites
are less common than in the Council Grove. Coarser
crystalline dolomite with grain-moldic porosity, typi-
cally dolomitized bioclastic or ooid grainstone or pack-
stone, is absent, and marine sandstone is uncommon
in the Council Grove. The very fine-crystalline do-
lomite is interpreted to have originally been mud-rich
carbonate for the most part.
Lithofacies Prediction
To predict lithofacies using neural network anal-
ysis, we used a standard, single-hidden-layer neural
network (Hastie et al., 2001) based on wire-line well
logs in 1364 node wells throughout the Hugoton and
Panoma fields. As illustrated in Figure 17, the input
vector to the neural network included two computed
Table 2. Digital lithofacies description system.*

Variables

1 2 3 4 5

Code Rock Type Dunham/Folk Classification Grain Size Principal Pore Size Argillaceous Content

9 Evaporite
8 Dolomite
7 Dolomite-limestone
6 Dolomite-siliciclastic
5 Limestone
4 Carbonate-siliciclastic
3 Siliciclastic-carbonate
2 Marine siliciclastic
1 Continental siliciclastic
0 Shale
Cobble conglomerate vcrs rudite and cobble congl (>64 mm) cavern vmf (>64 mm) Frac-fill 10 – 50%
Sucrosic/pebble conglomerate m-crs rudite/pebble congl (4 – 64 mm) med – lrg vmf (4 – 64 mm) Frac-fill 5 – 10%
Baffle-boundstone/vcrs sandstone fn rudite/vcrs sand (1 – 4 mm) sm vmf (1 – 4 mm) Shale > 90%
Grainstone/crs sandstone Arenite/crs sand (500 – 1000 Mm) crs (500 – 1000 Mm) Shale 75 – 90%
Packstone-grainstone/med Ss Arenite/med sand (250 – 500 Mm) med (250 – 500 Mm) Shale 50 – 75%
Packstone/fn sandstone Arenite/fn sand (125 – 250 Mm) fn (125 – 250 Mm) Shale 25 – 50%
Wackestone-packstone/vfn Ss Arenite/vfn sand (62 – 125 Mm) pin-vf (62 – 125 Mm) Shale 10 – 25%
Wackestone/crs siltstone crs lutite/crs silt (31 – 62 Mm) Pinpoint (31 – 62 Mm) Wispy 5 – 10%
Mudstone-wackestone/vf-m silt fn-med lutite/vf-m silt (4 – 31 Mm) Microporous (<31 Mm) Trace 1 – 5%
Mudstone/shale/clay Clay (<4 Mm) Nonporous Clean < 1%

Digital Description Lithofacies Lithofacies Class

1/>2 0 NM sandstone
1/2 1 NM siltstone
1/0-1 2 NM shaly siltstone
0.2/<3 3 Mar shale and siltstone
3-8/01 4 Mdst/Mdst-Wkst
3-8/2-3 5 Wkst/Wkst-Pkst
6-8/8/<3 6 Vfxln sucrosic (Dol)
3-8/4-5 7 Pkst/Pkst-Grnst
3-8/6-7 8 Grnst/PA Baff
7-8/8/>2 9 F-Mxln sucrosic moldic Dol
2/>2 10 Marine sandstone

*After Dubois et al. (2003). (A) Five-digit classification system used for core descriptions at 0.5-ft (0.15-m) intervals, gathered by visual observation with the aid of binocular microscope. A total of
seven other variables were recorded but were not used in determining lithofacies. (B) Digital code for 11 lithofacies. An example, 13323, is a continental siliciclastic, very fine-grained sandstone (203 –
410 ft; 62 – 125 m), with pinpoint porosity and wispy clay laminations (5 – 10% clay). Abbreviated are nonmarine (NM), marine (Mar), carbonate mudstone (Mdst), wackestone (Wkst), packstone,
(Pkst), grainstone (Grnst), phylloid algal bafflestone (PA Baff), dolomite (Dol), very fine crystalline (Vxln), and fine to medium crystalline (F-Mxln).
Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 321
322 / Dubois et al.

geologic variables, a depositional environment indica-

FIGURE 13. Distribution of Hugoton cores (continuous)
for which lithofacies were defined at 0.5-ft (0.15-m)
intervals.
tor (MnM) and a stratigraphic cycle relative position
(RelPos), in addition to the following wire-line-log pa-
rameters: gamma ray, logarithm of deep induction resis-
tivity, average of neutron and density porosity, neutron-
density porosity difference, and photoelectric factor
(PE, where available). No adjustment was made for thin-
bed or boundary effects. For each input vector, the net-
work computed a vector of output values representing
the corresponding lithofacies membership probabil-
ities, and the most probable lithofacies was assigned
for the logged interval. The network was trained based
on associations between core-defined lithofacies and
the log and geologic constraining variables. A com-
parison of core-defined lithofacies, lithofacies mem-
bership probabilities, and predicted discrete lithofa-
cies is shown in Figure 18.
The two geologic variables were derived from a
25-formation (or member) set of tops (Figure 8), which
are also the tops of marine or nonmarine (continen-
tal) half-cycles. Values for the depositional environ-
ment indicator were (1) nonmarine (continental) and
(2) marine and (3) intertidal, for the Herington and
Holmesville, where intertidal environments are domi-
nant. The MnM variable helps to distinguish between
lithofacies with similar petrophysical properties but
developed in different broad depositional environ-
ments. Values for the stratigraphic cycle relative po-
sition parameter (RelPos) range linearly with depth
from 0 at the bottom of each half-cycle interval to 1 at
the top, indicating position in each interval. Includ-
ing this curve allowed the network to encode infor-
mation regarding the fairly regular succession of litho-
facies commonly exhibited within each interval and,

FIGURE 14. Major lithofacies in Chase and Council Grove, lithofacies code 0-5. (A) Continental sandstone (L0): Example:
Blue Rapids (Council Grove, B1_SH), Cross H Cattle 1-6, 2652 ft (808 m). Coarse silt to very fine-grained sandstone,
mostly quartz, massive bedded, adhesive meniscate burrows (S. Hasiotis, 2005, personal communication). Low-relief
migrating eolian system. Digital classification: 13322. (B) Continental coarse siltstone (L1): Example: Stearns (Council
Grove, B4_SH), Newby 2-28R, 2963 ft (903 m). Coarse quartz silt, rhizolith (Rz) and root traces with reduction haloes
(Ho). Savannah, slow accumulation of silt by airfall, stabilized by vegetation and soil processes. Digital classification:
12213. (C) Continental shaly siltstone (L2): Example: Hooser (Council Grove, B3_SH), Newby 2-28R, 2944 ft (897 m).
Fine- to medium-grained quartz silt and clay, caliche (Ca), rhizolith (Rz), and root traces with reduction haloes (Ho).
Coastal plain, slow accumulation of silt by airfall, stabilized by vegetation and soil processes. Digital classification: 11114.
(D) Marine siltstone and shale (L3): Example: Funston (Council Grove, A1_LM), Newby 2-28R, 2872 ft (875 m). Very fine-
grained shaly siltstone. Siliciclastic-dominated shelf during maximum flooding. Plug $ = 4.6%; k = 0.0001 md. Digital
classification: 21104. (E) Mudstone and mudstone-wackestone (L4): Example: Crouse (Council Grove, B1_LM), Alexander
D-2, 2962 ft (903 m). Silty mudstone-wackestone, wispy laminations, and ministylolites (Ms), burrowed in part (Bh),
sparse normal-marine fauna, including fusulinids (Fs). Low-energy shelf at a time close to maximum flooding. Plug $ =
3.1%; k = 0.00239 md. Digital classification: 41113. (F) Wackestone and wackestone-packstone (L5): Example: Fort Riley
(Chase), Flower A-1, 2700 ft (823 m). Slightly dolomitized wackestone, normal-marine faunal assemblage includes
echinoids, brachiopods, bryozoan, and fusulinids. Intercrystalline micropores (blue in thin section) in dolomitized mud
matrix is dominant porosity (core slab and thin section stained with alizarin red). Low-energy normal-marine shelf.
Full-diameter $ = 15.2%; k = 0.413 md. Digital classification: 52111.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 323

thus, transfer some of that character to the sequence of
predicted lithofacies in each well. The two curves were
computed for the node wells using Visual Basic code
in an ExcelR spreadsheet run in a batch-processing
routine and exported as log curves in a Log Ascii Stan-
dard (LAS) file format. They were then combined with
the wire-line-log curves to complete the feature vector.
The neural network code has been added to
Kipling.xla, an ExcelR add-in for nonparametric
regression and classification (Bohling and Doveton,
2000). For this study, the network was trained to
match observed associations between logs and litho-
facies identified in core from a set of key wells shown
in Figure 13 (8 Chase wells with 3952 0.5-ft [0.15-m]
324 / Dubois et al.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 325

tuned to the training data and unable to generalize.

FIGURE 16. Relative proportions of 11 lithofacies in 4250-ft
(1300-m) core from Chase and Council Grove groups,
Hugoton field. Eleven lithofacies are identified by a code
(L0–L10): L0 = sandstone; L1 = coarse siltstone; L2 = shaly
siltstone; L3 = siltstone; L4 = carbonate mudstone; L5 =
wackestone; L6 = very fine-crystalline dolomite; L7 = pack-
stone; L8 = grainstone and phylloid algal bafflestone;
L9 = fine-medium-crystalline moldic dolomite; and L10 =
sandstone.
intervals in the training set, and 10 Council Grove
wells with 4593 0.5-ft [0.15-m] intervals). Fundamen-
tal parameters controlling the network behavior are
the number of nodes in the hidden layer (network
size) and a damping or decay parameter. Increasing
the network size allows the network to match the
training set more closely, but using too many hidden-
layer nodes leads to the network becoming precisely
Increasing the damping parameter counteracts pre-
cise tuning but can result in the network developing
smoother representations of the boundaries between
lithofacies.
We used cross-validation to estimate the optimum
values for network size and damping parameter. The
cross-validation was done in two ways: (1) splitting
the entire training data set, regardless of well, into
random subsets, with two thirds of the cases used for
training and the other one third for testing (compari-
son of predicted and actual lithofacies); and (2) tak-
ing out each well in turn, training on the remaining
wells, and testing on the removed well. Training and
testing were repeated several times for each param-
eter combination to account for the random varia-
tion between different realizations of the network.
This computationally intensive cross-validation was
performed using scripts in the R statistical computing
language (R Development Core Team, 2005) using
the nnet function developed by Venables and Ripley
(2002). The scripts computed several measures of cor-
respondence between actual and predicted lithofa-
cies, including the average lithofacies misallocation
cost for all intervals in the test well. This value is
computed from a misallocation cost matrix that as-
signs a cost for misallocation that is proportional to
the distance in lithofacies code units in the lithofacies
spectrum from actual. For example, calling a pack-
stone (L7) a marine siltstone (L3) carries a higher cost
than confusing it with a grainstone (L8) with similar

FIGURE 15. Major lithofacies in Chase and Council Grove, lithofacies code 6-11. (A) Very fine-crystalline sucrosic
dolomite (L6): Example: Cottonwood (Council Grove, B5_LM), Beatty E-2, 2800 ft (853 m). Finely crystalline, sucrosic
dolomitized mudstone, locally with anhydrite cement and replacement in nodules and along fracture (An). Porosity
(blue in thin section) is microporous (intercrystalline) and pinpoint (molds-Mo). Restricted, protected lagoon. Plug $ =
13.9%; k = 1.37 md. Digital classification: 88120. (B) Packstone and packstone-grainstone (L7): Example: Winfield (Chase),
Flower A-1, 2579 ft (768 m). Medium- to coarse-grained bioclastic-oncoid packstone, patchy anhydrite cement (An).
Most porosity (blue in thin section) is intergranular. Carbonate sand shoal on open shelf. Full-diameter $ = 16.4%;
k = 5.98 md. Digital classification: 54520. (C) Grainstone or phylloid algal bafflestone (L8): Lithofacies have similar core
and wire-line-log properties and were lumped because of their small populations. Example 1: Cottonwood (Council
Grove, B5_LM), Alexander D-2, 3024 ft (922 m). Medium-coarse grained oncoid-peloid grainstone. Well-connected
intergranular porosity is blue in thin section. Carbonate sand shoal on restricted shelf. Full-diameter $ = 18.8%; k =
39.0 md. Digital classification: 56540. Example 2: Cottonwood (Council Grove, B5_LM), Newby 2-28R, 2992 ft (912 m).
Phylloid algal bafflestone. Phylloid algal blade molds (Pm) partially filled with anhydrite cement (An). Matrix is largely
peloid-pellet packstone (Pp). Phylloid algal mound on slightly restricted shelf. Full-diameter $ = 20.6, k = 1141 md. Digital
classification: 57770. (D) Fine to medium crystalline moldic dolomite (L9)-Example: Krider (Chase), Flower A-1, 2516 ft
(767 m). Fine-medium crystalline moldic dolomite. Large molds (Mo), possibly ooids and bioclasts, dominate the well-
connected pore system in a dolomitized medium-coarse grained grainstone. Patchy anhydrite cement (An) occludes
some porosity. Carbonate sand shoal on an open shelf. Full-diameter $ = 22.3%; k = 275 md. Digital classification: 88550.
(E) Marine sandstone (L10): Example: Herington (Chase), Flower A-1, 2485 ft (757 m). Planar (Px) and ripple (Rx) cross-
bedding and vertical burrows (Bv). Tidal flat. Very coarse silt to very fine-grained sandstone, well-sorted, subarkose (83%
of detrital fraction is quartz, by x-ray diffraction), well-connected intergranular porosity (blue), patchy anhydrite cement
(An). Full-diameter $ = 20.8%; k = 48.2 md. Digital classification: 23321.
326 / Dubois et al.
petrophysical properties. Absolute accuracy in litho-
facies prediction, although desirable, is unrealistic,
and our goal was to limit error to the nearest litho-
facies class. Plots of median average misallocation cost
versus damping parameter and network size (panel
variable) for the well-by-well cross-validation, includ-
ing the photoelectric wire-line log (PE) in the Council
Grove (Figure 19), illustrate the median computed
over 40 average misallocation costs for each param-
eter combination: five trials per well for each of the six
Council Grove wells with PE logs. Cross-validation
plots for the Council Grove case without PE, and for
the Chase, with and without PE, were similar (not
illustrated). Although the network performed reason-
ably well throughout a range of parameter values, we
chose to use a network size of 20 hidden-layer nodes
and a damping parameter of 1.0. The damping pa-
rameter chosen (1.0) exhibited consistently low mis-
allocation values, and the number of hidden-layer
nodes (20) was chosen over configurations with fewer
nodes that tended to overgeneralize.
Initially, we trained four neural networks: Chase
with and without PE logs and Council Grove with
and without PE logs. Predictions using the model
including PE were used wherever possible. The Chase
models included all 11 lithofacies, but the Council
Grove models included only 9 because the fine- to
medium-crystalline dolomite and marine sandstone
do not occur in the Council Grove in sufficient vol-
ume to be considered as separate classes. Later, we
FIGURE 17. Schematic repre-
sentation of single hidden-
layer neural network used to
predict lithofacies from wire-line
logs and geologic constraining
variables. Inputs two geologic
constraining variables (MnM =
depositional environment indi-
cator; RelPos = relative position
in stratigraphic interval) and
includes an array of nuclear and
electrical wire-line-log curves:
gamma ray, (GR); logarithm
of deep induction log (LogILD);
average of neutron and den-
sity porosity (%N + %D)/2);
difference between neutron
and density porosity (%N  %D);
and photoelectric effect (PE).
Outputs are lithofacies occur-
rence probabilities.
added two additional models for the Chase below the
Towanda (with and without PE) to better represent
the distribution of marine sandstone in that interval
and region. These six neural network models were then
applied as appropriate to produce predicted lithofa-
cies versus depth logs at 0.5-ft (0.15-m) intervals in
the 1364 node wells distributed throughout the field
(Figure 20A). The batch prediction capability of the
Kipling program was used in this case, with logs being
read from LAS files and the predicted lithofacies curves
being written out to LAS files.
Following the construction of predicted lithofacies-
depth curves, the predicted lithofacies curves were read
TM
into Petrel (Schlumberger 3-D modeling software)
and upscaled to the resolution of the model layers
(roughly 2 ft [0.6 m] in marine intervals and 4 ft [1.3 m]
in nonmarine intervals) by majority vote: the litho-
facies for each model cell intersecting a well is taken
to be the most frequently occurring lithofacies in that
layer in the well. Voxel-based methods were chosen
over object-based methods for facies and property mod-
eling because of the relatively dense well control and
geometry of the lithofacies bodies being modeled
(thin and laterally extensive). Sequential indicator sim-
ulation (Deutsch and Journel, 1998) as implemented
in Petrel was used to generate lithofacies values in
all model cells conditioned on the upscaled litho-
facies values in the node well cells. The development
of variograms for the lithofacies by data analysis on a
zone-by-zone basis (24 zones and 11 lithofacies) is
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 327

FIGURE 18. Comparison of predicted lithofacies versus core-defined lithofacies. Illustrated vertical plots of lithofacies
membership probabilities, predicted discrete lithofacies (most probable), and lithofacies in core at the 0.5-ft (0.15-m)
scale for the Chase and Council Grove from two separate wells in the training set (Youngren and Stuart, respectively).
Neural networks were those used in estimating lithofacies in node wells (trained on all wells). Probabilities were not
used as an input for modeling, but they do illuminate some of the misallocations (actual lithofacies is commonly in
second place).

difficult and not justified considering the well control
and lithofacies geometry. Regions were therefore clas-
sified using a limited number of variograms for each
lithofacies and large horizontal ranges. Using large
horizontal ranges encouraged the stochastic simula-
tion to produce facies bodies that are as laterally ex-
tensive as possible, but still subject to conditioning
data at the node wells. This provided facies bodies
consistent with geologic depositional models but
constrained by node well data. Lithofacies bodies
are laterally extensive because of both the width of
the lithofacies spectrum used in determining class
(lumped lithofacies) and the geologic conditions
under which they were deposited. Node well spac-
ing is much closer than the width of the lithofacies
bodies. For the marine half-cycles, the major axis was
set at 30,000 ft (9144 m) and minor axis at 25,000 ft
(7620 m) and the azimuth equal to 118, the approxi-
mate depositional strike, whereas in the continental
and tidal-flat half-cycles, a 30,000  30,000-ft (9144 
9144-m) range was used. Vertical range was set at ap-
proximately two times the mean layer height in the
zone, and nuggets ranged from 0.1 to 0.22 on the
basis of limited data analysis. The very large number
of conditioning data helped reduce the sensitivity of
the simulated results to the variogram model param-
eters, making the predictions more deterministic.
Table 3 shows the proportional distribution of litho-
facies defined in core in the training set (keystone
wells) and predicted lithofacies at the node wells and
in the lithofacies model in the study area. Consid-
ering bias by the geographic distribution of sam-
pling and the scale change from 0.5 ft (0.15 m) to
much thicker layering, consistency in the distribution
at each scale can be characterized as good. Most dis-
crepancies are between the core-defined lithofacies
328 / Dubois et al.
(training wells) and the lithofacies predicted by
neural networks (node wells) and are caused by the
position of the training wells on the shelf. Training
wells are skewed to the west (because of core avail-
ability), whereas the node wells are more evenly
distributed. An exception is grainstone (L8), where
the neural networks as trained are not effectively
predicting this facies. Fortunately, this facies repre-
sents only a small part of the volume (2.3% in the
training wells) and is commonly mistaken as pack-
stone, a close neighbor, in the neural network train-
ing exercises.
Core Petrophysics
Previous petrophysical studies of the Hugoton and
Panoma fields have generally used average proper-
ties assigned to formations (e.g., Siemers and Ahr,
1990; Olson et al., 1996). Porosity and permeability
characteristics of reservoir-quality wackestone, pack-
stone, and grainstone lithofacies in the Council Grove,
Texas County, Oklahoma, were reported by Heyer
(1999). Byrnes et al. (2001) and Dubois et al. (2003)
presented lithofacies-specific petrophysical proper-
ties for the Council Grove in Panoma field and il-
lustrated the similarities between low-permeability
carbonates and sandstone. Fundamental to construc-
tion of the reservoir geomodel is the population of
cells with the basic lithofacies and their associated
petrophysical properties—porosity, permeability, and
FIGURE 19. Example results
for cross-validation analysis to
determine optimal values of
neural network size (number
of hidden-layer nodes) and
damping parameter. Results
shown are the median over
eight Council Grove wells
with core-defined lithofacies
and wire-line photoelectric
curve. The procedure was to
perform five trials per well;
leave out each well in turn,
train on the other seven wells,
and predict on the subject
well. Median average mis-
allocation cost versus damping
parameter and network size
for all wells was then plotted.
A network size of 20 hidden-
layer nodes and a damping
parameter of 1.0 were chosen.
fluid saturation. Petrophysical properties vary among
the 11 major lithofacies. Principal lithofacies-specific
petrophysical properties analyzed and discussed here
include routine helium and in-situ porosity, routine
air and in-situ Klinkenberg gas permeability, grain
density, capillary pressure, and gas-water relative
permeability.
Data for routine porosity, permeability, and grain den-
sity were compiled for more than 6000 full-diameter
core and plug samples from measurements performed
by commercial laboratories and the Kansas Geolog-
ical Survey. Lithofacies were determined for more
than 3500 samples. Core-plug sampling was designed
to represent the range in porosity, permeability, and
lithofacies in the study area. Routine air permeability
was generally measured under a confining pressure of
approximately 400 psi (2.8 MPa); in-situ Klinkenberg
gas permeability (high-pressure gas or liquid equiva-
lent) and in-situ porosity were measured under a hy-
drostatic confining pressure of 800 psi (5.5 MPa) and
a hydrostatic in-situ stress equal to the net effective
stress. A correlation between routine helium and in-
situ porosity was determined for 245 core samples
representing the range in porosity and lithofacies. In
addition, capillary-pressure curves were obtained for
252 samples using air-mercury capillary-pressure intru-
sion analysis. Data for 32 gas-water drainage-relative
permeabilities were compiled, and effective gas per-
meability at critical water saturation was measured
on more than 200 cores.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 329

FIGURE 20. (A) Map illustrating the location of 1364 node wells used for the static model construction. Fourteen of the
wells have core-defined lithofacies (circled) and the balance has lithofacies predicted by neural networks. The 1364 is a
mix of wells with Council Grove (1146) and Chase (1069) wells with lithofacies defined by neural net (854 of the 1364 have
both Council Grove and Chase lithofacies). Only wells with lithofacies defined at least to the top of the Fort Riley (Chase)
or the Florena Shale, B5_SH (Council Grove), were considered. (B) Map showing 8765 wells with formation-member level
tops was used for building the structural and stratigraphic framework for the Hugoton geologic model.

Porosity and Grain Density Permeability

Routine (unconfined) helium porosity ($He) values
range from 1 to 26% (Figure 21), with in-situ porosity
($i) values averaging 0.54 porosity units less than
routine porosity values (i.e., $i = $He  0.54) for all
lithofacies and porosity rocks. Comparing porosity
among lithofacies, average porosity decreases with
decreasing grain size (i.e., very fine sandstone to very
fine-medium siltstone) in the continental siliciclas-
tics and from grainstone to mudstone in the carbon-
ates. Continental rocks exhibit an average grain
density of 2.70 ± 0.08 g/cm3 (error expresses 2 stan-
dard deviations), limestones exhibit a grain density
of 2.72 ± 0.06 g/cm3, and dolomites exhibit a grain
density of 2.82 ± 0.08 g/cm3.
Core-measured in-situ Klinkenberg permeabilities
(kik) range from 0.00002 to 400 md (2  108 –0.4 mm2).
Permeability is a function of several variables, includ-
ing primarily pore-throat size, porosity, grain size and
packing (which controls pore body size and distribu-
tion), and bedding architecture. Lithofacies-specific
equations were developed to predict permeability using
porosity as the independent variable because porosity
data are the most economic and abundant and be-
cause porosity is well correlated with the other vari-
ables for most lithofacies.
More than approximately 75% of all rocks in the
Hugoton exhibit an in-situ Klinkenberg permeability
of less than 1 md (0.001 mm2). Obtaining accurate
330 / Dubois et al.

Table 3. Relative distribution of 11 lithofacies in core, node wells, and cellular model.*
Height Source 0.5 ft (0.15 m) 0.5 ft (0.15 m) Variable** Variable**
Actual (%) NNet Predicted (%) Upscaled (%) Modeled (SIS) (%)

Lithofacies Code Training Node Wells Node Wells All Cells

0 5.6 2.2 1.0 1.1

1 23.3 19.7 17.0 16.7
2 12.9 9.6 7.1 6.7
3 7.5 9.6 9.0 9.1
4 5.4 4.3 3.6 3.9
5 14.5 20.1 22.2 22.5
6 2.8 4.9 3.9 3.8
7 14.7 24.7 25.9 25.2
8 2.3 0.2 0.2 0.2
9 5.6 1.4 3.8 3.8
10 5.4 3.4 6.3 7.1
Count (N) 8545 993,146 183,949 107,765,147

*Core-defined lithofacies for 14 wells were used in neural network Training for lithofacies prediction in 1364 Node Wells. Lithofacies 0.5 ft (0.15 m)
in node wells were upscaled to model layer thickness (Variable Upscaled). Sequential indicator simulation (SIS) was used to populate the cellular
model (All Cells) between the node wells.
**Model layer h. Average of mean h = 3.3 ft (1 m). Range of mean h = 1.9 – 5.2 ft (0.57 – 1.58 m). Lithofacies 0 – 2 tend to be in thicker layers.

permeability values for rocks with low permeability
requires correction for the influence of confining
stress and care to avoid core data influenced by stress-
release microfractures. Figure 22 shows that both
continental and carbonate rocks in the study area
exhibit a significant decrease in permeability below
approximately 1 md (0.001 mm2) because of the in-
fluence of Klinkenberg gas slippage and confining
stress combined. This decrease is consistent with other
tight gas rocks (Byrnes et al., 2001).
Cores with identified macrofracturing exhibit in-
creasingly greater permeability with decreasing ma-
trix permeability (Figure 23). The permeability of
the fractured cores with matrix permeability below
0.5 md (0.0005 mm2) can be attributed to the core
permeability measurement reflecting the permeabil-
ity of fractures in the sample with the matrix contri-
bution being small or negligible. Full-diameter anal-
ysis, generally performed at confining pressures less
than 400 psi (2.8 MPa), commonly exhibit significant
difference from plug values for permeabilities below
0.5 md (0.0005 mm2), even for samples where frac-
tures were not identified but microfractures may have
been present. To obtain accurate matrix permeability-
porosity correlations, full-diameter permeability data
for permeabilities less than 0.5 md (0.0005 mm2) that
were more than 2 standard deviations outside the
plug-defined matrix permeability-porosity trend were
not used in final matrix correlations. The correlation
between matrix and well test permeability is discussed
in a following section.
As with many sedimentary rocks, the relationship
between permeability and porosity can be approxi-
mated by power-law functions, although the rela-
tionship changes slightly in some lithofacies at po-
rosities below approximately 6%. Each lithofacies
exhibits a relatively unique kik  $i correlation that
can be represented using a power-law equation of the
form (Figure 24)
kik ¼ AfBi
where kik is in millidarcys (md); porosity is in percent
(%); and values for A and B are shown in Table 4.
Subparallel trends are apparent for the continental
siliciclastics, the sucrosic dolomites, and the mudstone-
to-grainstone limestones. For these trends, the standard
error of prediction ranges from a factor of 3.3–9.1. At
$i > 6%, permeability in a grainstone-bafflestone can
be 35 times greater than mudstone and 150 times
greater than marine siltstone of similar porosity. These
differences illustrate the importance of identifying
lithofacies to correctly predict permeability from wire-
line-log porosity. Differences in permeabilities be-
tween continental very fine sandstone and coarse
siltstone and between coarse siltstone and very fine-
medium siltstone are approximately 2.5 times,
whereas the difference between marine sandstone
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 331
FIGURE 21. Histogram of routine helium porosity for
Chase and Council Grove nonmarine continental (NM)
sandstones and siltstones (A) and limestones (B). Litho-
facies codes and descriptions are provided in text. Porosity
generally increases with increasing grain size in siliciclastics
and with decreasing mud content from mudstone through
grainstone (Baff = bafflestone; Grst = grainstone; Pkst =
packstone; Wkst = wackestone; Mdst = mudstone).
and marine siltstone is approximately 10 times.
Although models for permeability prediction using
porosity were developed for utility, we recognized
that the dominant control on permeability is the
pore-throat size (Figure 25).
Water Saturation and Capillary Pressure
It is important to consider the presence of water
in the pore space of low-permeability reservoirs both
for accurate volumetric calculations and because
water occupies critical pore-throat space and can
greatly diminish gas permeability, even in rocks at
irreducible water saturation (Swi). In the Hugoton,
determination of formation water saturation from
electric wire-line-log response is problematic because
of deep mud filtrate invasion with conventional mud
programs because of the low reservoir pressure (Ol-
son et al., 1997; George et al., 2004). Because water
saturations could not be reliably determined for most
wells using logs, saturations were estimated based on
matrix capillary-pressure properties and determina-
tion of the free-water level (level at which gas-brine
capillary pressure is zero). Air-mercury capillary-pressure
data were compiled and measured for 252 samples
ranging in porosity, permeability, and lithofacies, and
relationships developed that allowed the prediction
of a capillary-pressure curve for any given lithofacies
and porosity.
To examine the lithofacies dependence of thresh-
old-entry pressure, gas-column height, and pore-
throat size (Figure 26), laboratory capillary-pressure
data were converted to reservoir gas-brine capillary-
pressure data using the standard equation (Purcell,
1949; Berg, 1975): Pcres = Pclab (scosures/scosulab); where
Pcres is the gas-brine capillary pressure (psia) at reser-
voir conditions, Pclab is the laboratory-measured cap-
illary pressure (psia), and scosures and scosulab is the
interfacial tension (s, dynes/cm) times the cosine of
the contact angle (u, degrees) at reservoir and labo-
ratory conditions, respectively. For the Hugoton and
Panoma fields, the interfacial tension is approximate-
ly 63–65 dynes/cm for the initial reservoir gas pres-
sures of about 400–450 psi (2.8–3.1 MPa) and tem-
peratures of 90–1008F (32–388C) (Hough et al., 1951;
Jennings and Newman, 1971). Conversion of capillary
pressure to height above free-water level to determine
the water saturation in any given rock type as a func-
tion of height above the free-water level requires the
conversion of capillary-pressure data to height above
free-water level. This conversion was performed using
the standard relation (Hubbert, 1953; Berg, 1975): H =
Pcres/(C(rbrine  rgas)), where H is the height (ft) above
free-water level, Pcres is the capillary pressure (psia) at
reservoir conditions, rbrine and rgas are the density of
brine and gas at reservoir conditions (rbrine = 1.16–
1.19 g/cm3 and rgas = 0.025–0.035 g/cm3, which are
reasonable intermediate values for these fields, and C
is a constant (0.433 [psia/ft]/[g/cm3]) for converting
density to pressure gradient in psia/ft. From the air-
mercury capillary-pressure data, pore-throat diameter
was calculated using the modified Washburn (1921)
relation: d = 4Cscosu/Pc, where Pc = capillary pres-
sure (psia); C = 0.145 psiacmmm/dyne; u = contact
angle (1408); s = interfacial tension (484 dynes/cm);
and d = pore-throat diameter (mm). This relation
332 / Dubois et al.
assumes that the nonwetting phase (i.e., gas) enters
the pores through circular pore throats.
Figure 26 illustrates selected capillary-pressure
curves for samples of different permeability. Differ-
ences among capillary-pressure curves for the various
lithofacies correspond to variations in threshold-entry
pressure, pore-throat diameter, and water saturation
for various gas-column heights above the free-water
level, including the thickness of the transition zone
from Sw = 100% to approximately Swi.
Capillary pressures and corresponding water satura-
tions (Sw) vary among lithofacies and with porosity-
permeability and gas-column height. Threshold entry
pressures and corresponding heights above free-water
level are well correlated with permeability (Figure 26).
This is consistent with the relationship between
pore-throat size and permeability. The figure shows
that for rocks with in-situ Klinkenberg gas permeabil-
ity below approximately 0.003 md (3  106 mm2),
threshold entry heights are greater than the gas-
column heights available in the Hugoton, and there-
fore, the samples have Sw = 100%.
FIGURE 22. Crossplot of routine air
permeability (kair) versus in-situ Klin-
kenberg gas permeability (kik) for Coun-
cil Grove rocks (gray circles) and Chase
rocks (black triangle). Influence of both
confining stress and Klinkenberg cor-
rection increase with decreasing perme-
ability. Values of kik can be predicted
approximately from kair using log10kik =
0.059 (log10kair)3  0.187 (log10kair)2 +
1.154 log10kair  0.159, where perme-
abilities are in millidarcys (md). Vari-
ance is caused by both differing routine
conditions and rock response.
Synthetic capillary-pressure curves
were constructed based on capillary
pressure-porosity-permeability-litho-
facies relationships exhibited by the
252 cores analyzed, representing the
range in lithofacies and permeability.
Capillary pressures in each lithofa-
cies can be represented to be a func-
tion of porosity. Modeled capillary-
pressure curves for two important
lithofacies (Figure 27) illustrate that
with decreasing porosity (and asso-
ciated permeability), threshold entry
heights and transition zone heights
increase. Example capillary-pressure
curves for different lithofacies at a
given 10% porosity (Figure 28) illustrate the signifi-
cant differences in Sw that can exist among lithofa-
cies at any given height above free-water level. Because
differences decrease with increasing height, satura-
tions for all lithofacies approximately approach a
similar irreducible saturation at gas-column heights
above approximately 300 ft (90 m), except for low-
porosity rocks where saturation differences are still
evident. Using the capillary-pressure model, it was
possible to predict water saturation for any given litho-
facies and porosity at any given height above free-
water saturation and, thus, populate every grid cell
in the 3-D geomodel with water saturation values.
Relative Permeability
Gas and water drainage relative permeability curves
reveal several characteristics similar to other low-
permeability rocks. Water permeability, even at 100% Sw,
is less than Klinkenberg gas permeability and decreases
with decreasing permeability. Gas relative perme-
ability is less than the absolute gas permeability at
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 333

FIGURE 23. Crossplot of in-situ Klinkenberg permeability versus in-situ porosity for whole core identified as fractured
(asterisk), whole core that were not identified as fractured but may contain microfractures (gray circle), and unfractured
core plugs (black triangle). Permeabilities shown are either measured in-situ values or routine values corrected to in-situ
conditions using the equation presented in Figure 22. Whole core (full-diameter) values diverge from matrix (plug)
values at porosities less than about 10% and matrix permeability of approximately 0.5 md, reflecting the influence of
microfracture(s) on permeability in whole core samples with porosity less than 10%. Above 10% porosity, influence of
microfracture(s) is small.

all water saturations greater than zero and gas relative
permeability decreases significantly as Sw increases
above 50%. Relative permeabilities can be reasonably
modeled using Corey-type equations (Figure 29), sim-
ilar to other low-permeability rocks (Byrnes, 2003).
Permeability at Different Scales
Fundamental to modeling the permeability distri-
bution in the Hugoton is the need to understand the
relative function of matrix and fracture flow and the
possible scale dependence of permeability. Figure 23
showed that for rocks below approximately 8% po-
rosity, or approximately 0.5 md (0.0005 mm2), micro-
fractures in core significantly increased permeability.
A fundamental question for these data is as follows:
are the microfractures present in the subsurface, or
are they a stress release or coring-induced phenom-
enon? This question can only be answered by com-
paring upscaled matrix permeabilities with unfrac-
tured full-diameter permeabilities and with DST- or
well test-calculated permeabilities. Comparing care-
fully examined unfractured full-diameter permeabil-
ity values with core-plug values measured on plugs
334 / Dubois et al.

Table 4. Power-law coefficients (A) and exponents

(B) by lithofacies for in-situ Klinkenberg (md)
versus in-situ porosity (%) trendlines illustrated
in Figure 24.
Lithofacies In-situ In-situ
Code Klinkenberg Klinkenberg
Permeability Permeability
Coefficient Exponent
kik = A$iB kik = A$iB
A B

0 1.000E – 09 7.90
1 3.715E – 10 7.90
2 1.585E – 10 7.90
3 1.995E – 11 8.31
4 2.088E – 10 7.98
5 2.967E – 08 6.26
6 1.967E – 09 7.10
7 1.527E – 07 6.17
8 3.631E – 09 8.24
9 2.553E – 07 6.30
10 1.995E – 10 8.31

taken from the full-diameter cores (Figure 30) indi-

FIGURE 24. Crossplot of in-situ Klinkenberg permeability
(kik) versus in-situ porosity ($i) for the continental silici-
clastics (A), limestones (B), and dolomites (C). Lithofacies
and codes are discussed in the text. Each lithofacies ex-
hibits a relatively unique kik  $i correlation that can
be represented using a power-law equation of the form
kik = A$ Bi ; values for A and B are shown in Table 4. For
some samples, routine permeability values were con-
verted to in-situ values using the equation in Figure 22,
and routine porosity was converted to in-situ porosity
using the equation in the text. Trend lines exhibit a stan-
dard error of prediction ranging from 3.3 to 9 times
depending on the lithofacies.
cates that for homogenous samples, matrix proper-
ties apply to the full-diameter core scale.
The ability to compare well-scale permeability with
matrix permeability is limited because so few wells
have DST or well test data for thin intervals for which
core data are available and which were tested prior
to hydraulic fracturing, which complicates artificial
fracture-enhanced permeability with reservoir perme-
ability. In four key research wells, permeability was
measured using DST for multiple intervals for which
core analysis was also performed. To compare with
core permeabilities, full-diameter and plug perme-
abilities were arithmetically averaged (representing
parallel-flow contribution from each depth interval)
to determine average interval permeabilities. Correla-
tion between DST and upscaled full-diameter and plug
core permeabilities (Figure 31) shows good correlation
for intervals with permeability greater than about
0.5 md (0.0005 mm2). For interval permeabilities below
0.5 md (0.0005 mm2), full-diameter permeabilities
exhibit nearly constant permeability between 0.5
and 3 md (0.0005 and 0.000033 mm2), characteris-
tic of microfracture-influenced permeability. Matrix-
scale plug permeabilities can be either higher or lower
than DST permeabilities.
Variance in the DST-matrix permeability correla-
tion is partially or predominantly related to the lim-
ited vertical sampling of the core plugs and difficulty
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 335

FIGURE 25. Crossplot of principal pore-throat diameter (PPTD, micrometers) versus in-situ Klinkenberg permeability
(kik) for lithofacies in Chase and Council Grove groups and sandstones from outside the Hugoton embayment for com-
parison. The good correlation through eight orders of magnitude shows the predominant influence that pore-throat size
exerts on permeability and explains permeability changes with grain size and Dunham classification at a given porosity.
Second Y-axis shows corresponding threshold entry heights necessary for gas column to enter sample for gas pressure
and temperature conditions in the Hugoton area and discussed in text. The correlation between kik and PPTD can be
expressed: PPTD = 2.2 kik0.42. Standard error of prediction for this correlation is a factor of 1.7.

FIGURE 26. Selected capillary-pressure

curves for rocks of different perme-
ability illustrating general curve char-
acteristics. Capillary pressure has been
converted to height above free-water
level (at which Pc = 0) using equations
in text. These curves illustrate how
threshold entry height and transition
zone height increase with decreasing
permeability.
336 / Dubois et al.
FIGURE 27. Model capillary-pressure curves converted
to height above free-water level for nonmarine coarse
siltstone (A) and packstone and packstone-grainstone
(B) for different porosities. Threshold entry heights and
transition zone heights increase with decreasing porosity
for all lithofacies. The siltstones exhibit both greater entry
heights and higher transition zones than corresponding
porosity packstones. Threshold entry heights for coarse
siltstones less than 6% porosity exceed existing closure
in the field, indicating these rocks are water saturated.
Curves for packstones less than 6% porosity are not
modeled.
in representing some pore properties that are larger
in scale than core plugs. The single, phylloid algal
bafflestone interval exhibits significantly lower ma-
trix permeability because core plugs did not sample
the larger scale vuggy nature of this lithofacies, which
exhibits high permeability. Because microfractures
do not contribute significantly to measured perme-
ability for rocks with permeability greater than 0.5 md
(0.0005 mm2), both full-diameter and plug data reflect
matrix properties, and the good correlation with DST
permeabilities indicates that the reservoir is not frac-
tured at the scale of investigation of the DST. The
better correlation of plug and DST permeabilities for
intervals with permeability below 0.5 md, and the
fact that upscaled permeabilities from plug data are
greater than or equal to DST permeabilities for three
of four intervals, can be interpreted to indicate that
these intervals are also unfractured. These data, and
less precise data from other wells, indicate that the
production characteristics of many wells in the Hu-
goton are consistent with matrix properties control of
flow, without significant natural fracture contribu-
tion. Data and statistics on the fraction of wells that
exhibit production greater than what would be pre-
dicted from matrix properties have not yet been com-
piled and calculated.
Log Petrophysics
The application and validation of statistical and
petrophysical concepts to porosity estimation is an
important factor in reserve calculations. As always,
accuracy is a major concern, but special attention
must be paid to potential bias introduced by gas ef-
fects and lithology variation. Cumulative effects of
the factors can result in significant under- and over-
estimation of total hydrocarbons in place. As dis-
cussed earlier, the Chase and Council Grove are in-
vaded extensively by mud filtrate during drilling with
conventional mud programs, so that water satura-
tions estimated from resistivity logs are adversely
overestimated. Pore volumes remain unchanged re-
gardless of invasion, so that density and neutron logs
can be used with confidence for porosity estimation
in the Hugoton. However, these logs must be eval-
uated carefully to eliminate, first, borehole environ-
mental effects, and then the factors associated with
variable mineralogy and variable gas effects. Because
we used only relatively modern log suites (post-1980)
and only neutron and density porosity for estimating
porosity, there was no need to calibrate, normalize, or
otherwise preprocess the logs other than checking
quality.
Because the density and neutron measurements
are from devices that make contact with the borehole
wall, the highly anomalous porosities caused by wash-
outs must be removed prior to detailed analysis. Po-
rosity cutoffs of 20–22.5% were applied by lithofacies
to eliminate washout effects. This was followed by a
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 337
remedial elimination of shoulders caused by the trun-
cation. Porosities substituted into these washout in-
tervals corresponded to average porosities of the litho-
facies assigned to the interval. This procedure was
found to be effective in removing washout effects and
was performed using an automated process with lim-
ited manual intervention to preserve intervals with
high, but real, porosities such as those that occur in
more coarsely crystalline dolomites (L9). Following en-
vironmental correction, the calculation of porosities
from logs could then be developed as a reliable pro-
cedure, particularly when validated through the use
of porosities from the extensive core database as the
calibration standard.
Porosity Log Calibration among Lithofacies
The subdivision of the Chase and Council Grove into
lithofacies allowed a statistical strategy of porosity-log
calibration that was sensitive to mineralogy effects
while simultaneously accommodating the effects of
gas saturation in residual or greater volumes. Initial
regression analysis that related core porosities to log
measurements in the Council Grove showed a distinc-
tive difference in calibration model between the sili-
ciclastic lithofacies and the carbonate mudstone litho-
facies (L0–L4) and the carbonate lithofacies (L5–L9).
The best correlation of log and core porosities in the
siliciclastic lithofacies was exhibited for the density
FIGURE 28. Model capillary-pressure curves
converted to height above free-water level
for lithofacies L0-L10 at a constant poros-
ity of 10%. Figure 27 illustrates how curves
change as a function of porosity for two
lithofacies. In this figure, for a given po-
rosity, comparison between curves at any
given height above free water shows that
water saturations are generally greater
for the siliciclastics than the carbonates.
The differences among curves illustrate
the importance of knowing lithofacies to
accurately predict water saturation.
log alone, because the neutron log
appears to be adversely and erratically
affected by clay minerals in the fine
fraction. Siliciclastic lithofacies were cal-
ibrated separately, and the relation be-
tween each one’s density porosity ($d)
to core porosity is shown in Figure 32. A
striking distinction is present between
the continental lithofacies (L1 and L2)
and the marine siliciclastic lithofacies
(L3), which exhibits relative homoge-
neity. It should be noted here that the carbonate mud-
stone lithofacies (L4) typically has high silt content,
and its population can be characterized as exhibiting
properties more like a fine-grained siliciclastic than a
pure carbonate mudstone. By contrast, in the carbon-
ate lithofacies L5 to L9, the average of the neutron ($n)
and density ($d) porosities, calibrated to a limestone
matrix, was found to be the best estimator of porosity,
compensating for changes in lithology between lime-
stone and dolomite, as well as minor saturations of
residual gas.
Compensation for Gas Effects
Although gas effects are minor in the Council Grove
and can be compensated for by the averaging of neu-
tron and density porosities, they can be significant in
the Chase (Figure 33) and must be accommodated in
an expanded equation set. The equation commonly
applied to estimate the porosity with compensation
for gas takes the form $ = (($n2 + $d2)/2)0.5. The equa-
tion closely approximates the formula derived by
Gaymard and Poupon (1968) from a petrophysical
model of a gas-filled reservoir. An alternative and em-
pirical equation that is also widely used is a simple
weighted average of the neutron and density porosi-
ties with a weighting of one third applied to the neu-
tron porosity and a two thirds weighting of the den-
sity porosity (Asquith and Krygowski, 2004). This
338 / Dubois et al.

rosities related to neutron and density fractional log

porosities resulted in an equation of the form $ =
0.37$n + 0.62$d for limestone (lithofacies L5, L7, and
L8) and $ = 0.04 + 0.31$n + 0.53$d for dolomite (L6
and L9). The extra term in the dolomite equation
accommodates the lithology correction required for
logs calibrated to a limestone matrix. Porosities in the

FIGURE 29. Relative gas permeability (A) and water rela-

tive permeability (B) curves for 32 samples of various
lithofacies. Black curves represent predicted values for
the Corey-type equation model: krg = (1  (Sw  Swc,g)/(1 
Sgc  Swc,g))p (1  ((Sw  Swc,g)/(1  Swc,g))2), krw = ((Sw  Swc)/
(1  Swc))q (kw/kik), where Swc is the critical water satu-
ration, Swc,g is the critical water saturation for gas flow,
and all saturation terms are in fractions. Black curves rep-
resent mean values of exponents p = 1.3 for gas curve and
q = 8.3 for water curve, whereas gray-bounding curves
represent outer limits of curves using exponents p = 1.3 ±
0.4 and q = 8.3 ± 3, which represent the range exhibited FIGURE 30. Crossplot of full-diameter core porosity versus
by the sample set, which had 0.1 < kik < 50 md. plug porosity (A) and permeability (B) for samples in which
the full-diameter cores did not exhibit any apparent micro-
fracturing. Good correlation indicates that matrix-scale
properties apply to full-diameter scale. Variance can be
empirical equation closely emulates the gas correc- attributed to full-diameter core sampling multiple litho-
tion shown on service company neutron-density cross- facies or a range in porosity not sampled by the corre-
plots (Figure 34). A regression analysis of core po- sponding core plug.
Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 339

FIGURE 31. Crossplot of calculated

interval drillstem test (DST) formation
permeability versus average interval per-
meability calculated from full-diameter
core for four wells and from core plugs
in well 1. Routine core data were cor-
rected for confining stress, Klinkenberg,
and relative permeability effects so as
to correspond to reservoir-condition
values. Good correlation down to about
0.5 md shows matrix-scale control of
flow in the region of DST investigation.
Below 0.5 md, microfractures in full-
diameter core result in permeabilities
higher than in the unfractured reservoir.
Higher DST than core-plug permeabil-
ities can be interpreted to indicate that
formation is not fractured in the range
of investigation, and that plug sam-
pling density was probably not ade-
quate to properly sample lower range
of permeability.

FIGURE 32. Porosity calibra-

tion of siliciclastic zones based
on regression of core mea-
surements on density-log po-
rosity in limestone-equivalent
units for continental facies
(L1 and L2) and marine facies
(L3 and L4).
340 / Dubois et al.

FIGURE 33. Example of strong

(and atypical) gas effect on the
neutron and density-porosity
logs in the Towanda Limestone
for a well in Stevens County,
Kansas.

marine sandstone (F10) were

calibrated to log porosities
with the equation $ = 0.05 +
0.15$n + 0.47$d. Although
there are comparatively few-
er core measurements in the
continental sandstone (F0),
the use of the averaged neu-
tron and density porosities
provided an excellent match
with core porosities.

Free-water Level and

Trapping Mechanism
Estimating the free-water
level (FWL) position is crit-
ical for calculating water
saturations using capillary
pressures and the height
above FWL. It has been rec-
ognized that the Hugoton
field has a sloped gas-water
contact, and we interpret a
sloped FWL that is several
hundreds of feet higher at
the west updip margin than
on the east downdip limits
(Garlough and Taylor, 1941;
Hubbert, 1953, 1967; Pippin,
1970; Sorenson, 2005). In
this study, we have defined
the gas-water contact as the
lowest position in the res-
ervoir that a well can pro-
duce gas economically, with-
out substantial water, and
the free-water level as the

FIGURE 34. Neutron-density

porosity log crossplot of gas
zones in the Towanda Lime-
stone from the example Stevens
County well (see Figure 33).
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 341
datum where gas-brine capillary pressure is zero. As
shown in the section on petrophysics, initial reser-
voir desaturation may not occur for some lithofacies
until several tens or hundreds of feet above the free-
water level (threshold entry height). For typical res-
ervoir rocks in the study area, packstone-grainstone
8–10% porosity, the FWL ranges from 50 to 70 ft (9 to
21 m) below the gas-water contact, a point at which the
water saturation is approximately 70% (see Figure 27).
The Hugoton gas reservoir is a dry-gas, pressure-
depletion reservoir with very little or no support from
the underlying aquifer. Vertical water flow is con-
strained by low vertical water permeabilities through
low-porosity siltstone layers (k < 106 md [109 mm2]
for $ < 4%) and by low water relative permeability in
carbonates with low water saturation. However, below
the transition zone, water can be produced freely,
and reservoir pressures (600 – 700 psi; 4.1 – 4.8 MPa)
approach regional hydrodynamic pressures for the
depth (Sorenson, 2005). The low reservoir gas pres-
sures (450 psi; 3.1 MPa) and subhydrostatic water
pressures below the transition zone were proposed by
Sorenson (2005) to be the result of water pressure
equilibrating with reservoir rocks exposed at outcrop
in eastern Kansas and gas cap expansion, and conse-
quent pressure decrease.
The Hugoton has long been considered a classic
example of a giant stratigraphic trap (Garlough and
Taylor, 1941; Parhman and Campbell, 1993) because
of updip changes in lithofacies and petrophysical
properties associated with these changes. However,
dips on the apparent gas-water contact and FWL that
cross stratigraphic boundaries cannot be fully ex-
plained by lateral heterogeneities. Hubbert (1953,
1967) proposed a conceptual model for the Hugoton
being a hydrodynamic trap with trapping resulting
from a hydraulic gradient coupled with permeabil-
ity changes at the updip margin of the field. Pippin
(1970) cited Hubbert’s hydrodynamics and updip
pinch-outs of reservoir rock as the trapping mecha-
nism. Olson et al. (1997) suggested that sealing faults,
at least in the western part of the field in Stanton
and Morton counties, Kansas, compartmentalize the
lower Chase reservoirs with the compartments hav-
ing dramatically different gas-water contacts that rise
to the west. Sorenson (2005) suggested that the down-
dip flow of gas during expansion of the Hugoton gas
bubble might be responsible for the gas-water contact
geometry.
Determining the mechanism for an uneven FWL
was not an objective of our investigations, but FWL
had to be established for the calculation of water
saturations using capillary pressure. Although others
have presented general descriptions of the gas-water
contact datum (e.g., Garlough and Taylor, 1941; Parh-
man and Campbell, 1993), it has not been rigorously
defined by earlier workers. Our estimation of the FWL
(Figure 35) was derived using a combination of four
indicators: (1) base of lowest perforations; (2) forma-
tion fluid resistivity estimated from wire-line logs;
(3) calculation of the FWL for an estimated original
gas in place (OGIP); and (4) pressure measurements
of deep-water productive intervals. Within the limits
of the Panoma field, we based the depth of FWL on
the average lowest reported productive perforations
in the Council Grove (FWL = base of perforations +
70 ft [20 m]), assuming that operators have been
efficient at identifying pay and avoiding water pro-
duction. Along the eastern and western margin of
the Hugoton in Kansas, where there is no underlying
Council Grove production, and the western margin
in Oklahoma, we used a formation fluid resistivity
method for estimating the FWL at the field bound-
ary. Here, FWL was estimated to be 30 ft (9 m) below
the structural datum of the point in the Chase pay
zones where the average apparent formation fluid
resistivity (mix of all fluids) estimated from logs was
at or near the resistivity of formation water. Limited
data in the Oklahoma Panhandle required that FWL
be estimated by backcalculating the FWL necessary
to calculate an OGIP that would allow total gas pro-
duction to equal 70% of OGIP. This method assumed
that the Panhandle reservoir exhibited similar pressure
depletion and gas production as reservoirs in Kansas.
The FWL subsea depth is approximately 30 ft
(9 m) at the east margin of the Hugoton and, mov-
ing west, remains relatively flat to the east margin
of the Panoma, where it begins to rise at a rate of
15 ft/mi (2.85 m/km) to a datum of +250 ft (+80 m),
where it then rises at 50 ft/mi (9.4 m/km) to a height
of +1000 ft (+300 m) at the western margin of the
Hugoton. The configuration closely parallels the gas-
water contact described by Parhman and Campbell
(1993), although our estimate places it 100 ft (30 m)
lower at the east margin. Our estimated gas-water
contact is +40 ft (12 m) at this position in the field
(FWL + 70 ft [21 m]).
Static Model Construction
The extremely large area (10,042 mi2; 26,008 km2),
small XY cells (660  660 ft; 201  201 m) and rela-
tively thin layers (169 layers, 3.3 ft [1 m] thick average;
Figure 36) resulted in a 108-million-cell model that
required subdividing the model into parts because of
342 / Dubois et al.

FIGURE 35. Free-water level (FWL) estimated for the Hugoton model area (A, B) is applicable for all production from
the Chase but only for Council Grove production that is inside the Panoma. The entire Wolfcamp (Chase and Council
Grove) is thought to have a common FWL across much of the area, but the Council Grove production outside the Panoma
(in Oklahoma) may be related to a different FWL. The FWL is at near sea level at the east Hugoton margin, climbs gradually
in a westerly direction to midfield, and then ascends more rapidly to a height of more than 1000 ft (300 m) above sea
level at the west field margin. The zero datum of the height above FWL surface for the Chase (C) and Council Grove (D)
correspond well to the edges of the Hugoton and Panoma field boundaries, respectively.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 343
computational limitations. Because the main pay
intervals are separated stratigraphically by intervals
with relatively poor reservoir properties and lithofa-
cies and property modeling is restricted by zones in
the modeling process, we chose to subdivide the
Chase and Council Grove model stratigraphically
instead of geographically, with three multizone, multi-
formation models in the Chase and three in the
Council Grove. Each of the six models were built
with the same starting architecture and layering to
facilitate cutting out selected parts of the six models
and assembly into smaller models. These would have
a complete vertical section of the reservoir system,
but limited aerial extent, and be more easily man-
aged for further analysis and reservoir simulation.
The structural framework for the six models was
based on a structural tops database for 8765 wells
(Figure 20B). Our data set of tops in Oklahoma is less
complete than it is in Kansas. Layering within the
models used the following hierarchy: (1) division
between formation and members, (2) further subdi-
vision between continental and marine intervals, and
(3) further subdivision into layers based on min-
imum vertical thickness of the key lithofacies in the
node wells (Figure 20A). Layers averaged 2 ft (0.6 m)
thick in the marine intervals and 4 ft (1.2 m) thick in
the continental intervals. Horizontal and vertical
FIGURE 36. Intersecting structural cross
sections (near midfield) in the Hugoton
model illustrate the 169 layers in the
24 zones from the combined six 3-D
grids. Marine zones are more finely
layered than the continental intervals
(vertical exaggeration is 200).
permeability were calculated at the
cell level using the petrophysical equa-
tions presented above. Water satu-
ration was calculated at the cell level
using capillary pressure curves devel-
oped for each cell based on the cell
lithofacies, porosity, and height above
free-water level. Finally, OGIP was
calculated for a bottom-hole pressure
equal to 450 psi (3.1 MPa) and com-
pressibility index (Z) equal to 0.92.
RESULTS AND DISCUSSION
The accuracy and utility of the
Hugoton geomodel can be measured
by several metrics, including pre-
diction accuracy of parameters like lithofacies, petro-
physical properties, and OGIP at the well to field
scale. The only direct measure for lithofacies is the
comparison of predicted and core-defined lithofa-
cies (Table 3). We can also qualitatively measure the
validity of the lithofacies model by (1) comparing
it with earlier work at smaller scales and (2) com-
paring the 3-D lithofacies patterns with deposition-
al models that have been proposed for the area and
for upper Paleozoic cyclic depositional systems in gen-
eral. Measures of accuracy for any parameter (e.g.,
permeability) at the lease and field-scale are con-
strained by lack of data directly measured at these
scales and the consequent need to compare other pa-
rameters, such as OGIP. OGIP requires integration of
many variables, which may be inaccurate due to er-
ror in a single input variable or improper upscaling
and integration of accurate, but different scale, vari-
ables. Ultimately, measures of accuracy and utility of
a geomodel at the lease and field scales are com-
monly defined by comparison of predicted and mea-
sured production and pressure history, where the
predicted pressure and production data are obtained
from input of the static geomodel (the focus of this
chapter) into a reservoir-flow simulator to obtain a
dynamic model. The workflow involved in calibration
of the geomodel with dynamic data is not discussed in
344 / Dubois et al.

this chapter or shown in Figure 12, but is an important
part of testing and refining a geomodel. Dynamic
modeling has been performed of 9- and 12-mi2 (23-
and 31-km2) areas, including histories for 28 and 37
wells, respectively, to test and refine the geomodel
discussed here. These simulations are ongoing and are
both generally validating the present geomodel and
showing how uncertainty in some properties (e.g.,
free-water level) must be reduced to provide a model
that is sufficiently constrained for use in accurately
predicting field performance.
Model Lithofacies and Properties
Lithofacies and property distribution in the 3-D
Hugoton cellular geomodel presented here on a larger
scale and with finer resolution are consistent with
earlier work on the Hugoton (Garlough and Taylor,
1941; Hubbert, 1953; Pippin, 1970; Seimers and Ahr,
1990; Parham and Campbell, 1993; Fetkovitch et al.,
1994; Oberst et al., 1994; Olson et al., 1997; Heyer; 1999;
Sorenson, 2005). The full-field geomodel presented
here reveals facies and property patterns that could
not be identified at smaller scales. Figures 37 – 39 are
a series of cross sections and map views of lithofa-
cies and properties for the six individual models.
General trends in thickness and lithofacies distri-
bution are evident in the 3-D volume: continental
rocks are thickest, marine carbonate intervals thin or
pinch out at Hugoton’s western updip margin, and
the relationship is nearly reciprocal basinward. The
important reservoir lithofacies (grain-supported car-
bonate, dolomite, and marine and continental sand-
stone) are laterally extensive, and the marine carbon-
ates, the primary pay zones, are separated by laterally
continuous continental siltstone with poor vertical
transmissibility.
Large-scale sedimentation patterns, interpreted
from lithofacies distribution, are striking when viewed
at the scale made possible by the full field-scale model.
In cross section at the intercycle scale, backstepping
of major lithofacies associated with changes in water
depth and energy are evident in the Chase marine
carbonate, particularly in the two dolomite litho-
facies. The position on the shelf where continental
siliciclastics thicken at a higher rate also backsteps in
a similar manner. Marine carbonates in the middle
three of the seven Council Grove cycles (Figure 37C)
pinch out at the western field margin. The gradual
shift in the paleoshoreline position, first to lower on
the shelf and then back to a higher position, is be-
lieved to be related to a change in sea level amplitude.
The symmetry demonstrated by sedimentation pat-
terns around the middle cycle of the seven fourth-
order Council Grove cycles suggests the possibility of
higher order cyclicity in the latter part of the middle
Paleozoic icehouse.
One of the more striking aspects of the model
from a reservoir perspective is the demonstration of
lateral continuity in the lithofacies illustrated in
Figure 38. County-scale connected volumes of the
more significant reservoir lithofacies are subparal-
lel to depositional strike and the field margin. Grain-
supported packstones in the Crouse (Figure 38A, B1_LM,
Council Grove) are found primarily in the eastern
half of the field, whereas muddier lithofacies are
dominant in the western, more sheltered shoreward
part of the shelf. Continental sandstone (eolian) is
limited to the northwestern updip margin in the
Council Grove (Figure 38B), and shallow-marine
and tidal-flat sandstone in the upper Chase is most
abundant in the northwestern half of the field area
(Figure 38C).
The core to model lithofacies workflow was suffi-
ciently robust to characterize smaller important litho-
facies bodies in the reservoir system (Figure 38D –F).
For example, dolomitized grainstone and packstone
of a relatively thick (30 ft, 10 m) carbonate sand shoal
system in the southern half of the field is known
to be an important contributor to storage and flow

FIGURE 37. Lithofacies in stratigraphic cross sections across the Hugoton shelf (A) for the Chase (B) and Council Grove (C).
Cross sections are 10 – 158 from being dip sections and are hung on the top of the Chase (B) and the Council Grove (C).
Some key observations can be made: (1) In both the Chase and Council Grove, continental half-cycles (yellow-orange
to red lithofacies) are thickest at the west field margin and thin basinward (southeasterly). The pattern for the marine
half-cycles is the opposite and somewhat reciprocal relationship with the continental half-cycles. (2) Backstepping pattern
in lithofacies distribution from one marine cycle to the next in the Chase. (3) Three Council Grove marine half-cycles
pinch out near the west field margin, marking a paleoshoreline that appears to then move northwesterly (landward)
upsection. (4) Trend in carbonate rock texture from mud dominated (landward) to grain dominated (basinward),
especially in the Council Grove. Large-scale sedimentation patterns and distribution of resultant lithofacies (at the cycle
scale) is largely a function of the position on the shelf and reflect the interaction of shelf geometry, sea level, and possibly,
the proximity to siliciclastic sources. Lithofacies distribution and cycle stacking patterns at larger scales may be a function
of lower order cyclicity and a shift from icehouse to greenhouse conditions (upward) during the Lower Permian.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 345

capacity in the Krider (R. Sorenson, 2004, personal Distribution of lithofacies-dependent properties
communication). Modeling of this important lithofa- (porosity, permeability, and water saturation) corre-
cies shows the lateral continuity of a 4  30-mi (6.5  lates well with lithofacies distribution (Figure 39).
50-km) sweet spot (porosity > 18%; Figure 38F). The best porosity and permeability coincide with
346 / Dubois et al.

FIGURE 38. Maps showing lithofacies distribution of selected lithofacies. Illustrated are map views of connected volumes
TM
generated in Schlumberger’s Petrel modeling application. Connected volumes are collections of touching cells in
the cellular model having common properties and, for this model, help to demonstrate the remarkable lateral continuity
of flow units in the Wolfcamp. (A) Thirty largest connected volumes in the Crouse limestone (B1_LM) packstone,
packstone-grainstone (L7, blue), and fine-crystalline dolomite (L6, pink) having porosity greater than 8%. (B) Fifteen
largest connected volumes in the Speiser shale (A1_SH) continental sandstone (L0) with porosity greater than 12%.
(C) Twenty largest connected volumes of marine sandstone (L10) having porosity greater than 15%. (D) Top 20 con-
nected volumes for Krider packstone, packstone-grainstone (L7, blue) and coarse-crystalline dolomite (L9, purple) having
porosity greater than 16%. (E) Enlarged area of (D). (F) Same as (E) except for porosity greater than 18%. Grant and
Stevens counties are outlined in green for reference in (D), (E), and (F).

the main reservoir lithofacies (Figure 37). Laterally
extensive low-permeability intervals separate the rela-
tively high-permeability pay zones of the layered res-
ervoir system. Water saturations are high in the con-
fining intervals, and the gas-water contact crosses
stratigraphic boundaries at the east downdip margin
as the pay intervals dip below the surface and on the
west where the free-water level is thought to rise
more quickly than the rate of dip.
Quantitative measures of lithofacies proportions
(Table 3) in core, neural network-predicted lithofa-
cies at node wells, upscaled at the node wells and in
the 108-million-cell model, show consistency at the
different scales, suggesting that the sample rate for
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 347

FIGURE 39. Property distribution in cellular Hugoton model in cross section. (A) Location of cross sections. (B) Chase
Group stratigraphic cross section (datum is top of Chase). Horizontal permeability is shown in west – east section, and
porosity (0 – 30%, yellow is 22%) is in the north – south section. (C) Chase Group through Easly Creek shale (B2_SH,
Council Grove) water saturation. Free-water level is the base of the cross section on the west and east side and the base
of the Easly Creek (B2_SH) in the middle, where the FWL is lower in the stratigraphic column (not able to display all
models simultaneously). Free-water level crosses stratigraphic boundaries in both updip and downdip positions. Highest
permeability (Kxy) and porosity (Phi), and lowest water saturation (Sw) is found in marine carbonates and sandstones.
Continental siltstones separating the marine carbonates are the intervals with Kxy and Phi and higher Sw.

training and lithofacies prediction in each of the
three steps was sufficient. Slight variations in mea-
sures are likely to be related to sample distribution
in core versus the node wells and the increase in
scales from 0.5 ft (0.15 m) to considerably thicker
cells (layers). Several factors related to the nature,
geometry, and distribution of the predicted litho-
facies and the variables chosen to predict lithofacies
are believed to have enabled the success of the neu-
ral networks to predict lithofacies at the node wells
and the modeling of lithofacies between nodes. Litho-
facies classes were chosen to maximize differences
in the signature of wire-line-log variables. Geologic
constraining variables (e.g., relative position curve
and marine-nonmarine curve) captured and leveraged
geologic information such as the predictable vertical
348 / Dubois et al.
succession of lithofacies in the sedimentary cycles and
the primary depositional environment. Extensive lat-
eral continuity and lithofacies-body sizes being much
larger than the lateral spacing of node wells helped
model the lithofacies between nodes. Significantly,
there was adequate core control to appropriately sam-
ple the reservoir system for the lithofacies training
set. Without core, the model could not have been built.
The most common misallocations of lithofacies
were prediction of grainstone (L8) as wackestone or
packstone (L5 and L7), predictions of carbonate mud-
stone (L4) as wackestone (L5), and predictions of the
continental sandstone (L0) as coarse siltstone (L1).
These incorrect lithofacies predictions reflect over-
laps of the log characteristics of the lithofacies. For-
tunately, the lithofacies involved are also sufficiently
similar in their petrophysical properties that the re-
sulting distributions of porosity, permeability, and
water saturation resulting from random misallo-
cations were not significantly different from correct
property distributions. For example, a grainstone (L8)
having 10% porosity and at a position 300 ft (90 m)
above FWL misclassified as a packstone (L7) would be
assigned a permeability of 0.35 md instead of 0.70 md
(Figure 24) and a water saturation of 24% instead of
17% (Figure 28).
The majority vote upscaling of lithofacies from
the 0.5-ft (0.15-m) sampling interval to the thick-
ness of the model layers intersecting each well did
not significantly alter the lithofacies populations.
Using this methodology, we would not expect to see
a significant difference in lithofacies populations
unless certain lithofacies typically occurred in thin
bodies separated by fairly large vertical intervals, re-
sulting in a systematic underrepresentation of those
lithofacies in the upscaled results. Finally, the litho-
facies proportions in the full 3-D model closely re-
flect those for the upscaled well cells. This is not
surprising because the stochastic indicator simula-
tion attempts to match the proportions observed in
the conditioning data. The distribution of our node
wells was sufficiently uniform such that the litho-
facies proportions in the conditioning set were very
similar to those in the reservoir system.
Controls on Petrophysical Properties
The petrophysical analysis indicates that accu-
rate reservoir properties prediction requires input of
lithofacies, use of properties that represent reservoir
conditions, and filtering of full-diameter data to avoid
microfractured core. The close correspondence of DST
permeabilities and upscaled plug-scale permeabili-
ties is interpreted to indicate that production from
many wells is controlled by matrix permeability and
not fractures.
Comparison of petrophysical properties among
lithofacies indicates that permeability increases and
threshold entry height and transition zone thickness
decrease with increasing energy in the depositional
environment and corresponding decrease in mud and
silt matrix. Within both the continental and marine
siliciclastics, permeability increases with increasing
mean grain size at any given porosity. The marine,
very fine-grained sandstone (L10) exhibits an order
of magnitude greater permeability than marine siltstone
(L3). However, continental, very fine-grained sand-
stone (L0) exhibits only approximately 2.5 greater
permeability at any given porosity than continental
coarse-grained siltstone (L1) and 5 greater perme-
ability than fine–medium-grained siltstone (L2). This
difference is likely because of poorer sorting of the
continental sandstone compared to the marine sand-
stone. The poor sorting of the continental siliciclas-
tics would be expected to result in a wide pore-throat
size distribution, which is reflected in the high capillary-
pressure transition zones of these rocks.
Limestone permeabilities exhibit subparallel perme-
ability-porosity trends among lithofacies (Figure 24)
and increasing mean porosity (Figure 21), upper po-
rosity range, and permeability from mudstone through
packstone-grainstone. These changes result in a corre-
sponding increase in threshold entry heights and wa-
ter saturations at any given height above free water.
The differences in the capillary-pressure properties
between lithofacies decrease as porosity and perme-
ability increase. Both threshold entry heights and
transition zone heights for the limestone are less
than for continental siliciclastics of similar porosity.
Original Gas in Place
Comparison of estimated OGIP by others, production
data, and OGIP from calculations in the 108-million-
cell model suggests that the geomodel successfully
models the Hugoton, particularly in the center of the
field where control on the free-water level is greatest.
The geomodel OGIP is calculated to be 21.8 tcf
(0.62 trillion m3) and have a hydrocarbon pore vol-
ume of 676 bcf (19.1 billion m3, 120 billion bbl) in Grant
and Stevens counties, Kansas, for the Chase (Hugo-
ton) and Council Grove (Panoma) intervals (Table 5;
Figure 40). Cumulative gas production for the area is
14.1 tcf (0.4 trillion m3) or 65% of calculated OGIP,
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 349

Table 5. Estimated original gas in place (OGIP).*

Group Zone OGIP (bcf) OGIP (109 m3) HCPV (bcf) HCPV (109 m3)

Chase Herington 1227 34.8 38.1 1.08

Krider 2795 79.1 86.7 2.46
Odell 295 8.4 9.2 0.26
Winfield 3215 91.0 99.8 2.83
Gage 807 22.9 25.0 0.71
Towanda 4686 132.7 145.5 4.12
Holmesville 663 18.8 20.6 0.58
Fort Riley 5212 147.6 161.8 4.58
Matfield 127 3.6 3.9 0.11
Wreford 1048 29.7 32.5 0.92
Council Grove A1_SH 331 9.4 10.3 0.29
A1_LM 656 18.6 20.4 0.58
B1_SH 76 2.2 2.4 0.07
B1_LM 143 4.1 4.4 0.13
B2_SH 9 0.3 0.3 0.01
B2_LM 167 4.7 5.2 0.15
B3_SH 56 1.6 1.7 0.05
B3_LM 34 1.0 1.0 0.03
B4_SH 67 1.9 2.1 0.06
B4_LM 22 0.6 0.7 0.02
B5_SH 3 0.1 0.1 0.00
B5_LM 113 3.2 3.5 0.10
C_SH 2 0.1 0.1 0.00
C_LM 20 0.6 0.6 0.02
Total 21,775 589.7 675.9 19.1

*By zone for part of the Hugoton model covered by Grant and Stevens counties, Kansas, for 450 psi (3103 kPa) initial bottom-hole pressure.

slightly low compared to earlier work by others. For
the Chase in Kansas, Oberst et al. (1994) estimated
OGIP volumetrically at 31.1 tcf (0.88 trillion m3),
whereas Olson et al. (1997) placed it at 34.5 – 37.8 tcf
(0.98 – 1.1 trillion m3). Because their estimates were
for different reservoir volumes than ours (Chase in
Kansas versus Chase and Council Grove in two coun-
ties in Kansas), we cannot compare directly, but assum-
ing similar reservoir performance, we can compare
the estimates on the basis of production efficiency.
The ratio of Chase cumulative production to date
(24.8 tcf; 0.7 trillion m3) to OGIP is 79.7% by Oberst
et al. (1994) and 65.6– 71.9% by Olson et al. (1997).
Our estimate for the entire Wolfcamp reservoir vol-
ume (65%) is closer to the Olson et al. (1997) esti-
mate. In Kansas, 89.2% of the Hugoton – Panoma
production is attributed to the Chase and 10.8% to
the Council Grove, although we believe the two be-
have as one reservoir system.
It is important to note that estimation of OGIP
using the matrix capillary-pressure method em-
ployed in this study is influenced by natural var-
iance in the capillary-pressure curves and the de-
termination of free-water level. Natural variance in
capillary-pressure curves can result in a one stan-
dard deviation confidence interval for predicted wa-
ter saturation of greater than 10% of the saturation
value (e.g., Sw = 10% results in 9% < Sw < 11% or,
for S w = 80%, 72% < S w < 88%), which results
in a one standard deviation confidence interval for
predicted OGIP of approximately 3%. Change in
the free-water level results in little change in water
saturation for intervals greater than 300 ft (91 m)
above FWL, but can have significant influence on
intervals that are within their transition zone and
have rocks that exhibit transition zones that are
only tens of feet high. For these intervals, shift in
the FWL by a few tens of feet can result in significant
350 / Dubois et al.
water saturation change and, consequently, OGIP
changes. This was important for the Council Grove
OGIP estimate and is still being refined.
CONCLUSIONS
The more than 100-million-cell, 10,000-mi 2
(26,000-km2), 3-D geologic and petrophysical prop-
erty geomodel of the Hugoton presented in this study
demonstrates the application of a detailed reservoir
FIGURE 40. Map views of original gas
in place (OGIP) for Grant and Stevens
counties, Kansas (A), are the two most
prolific counties in Hugoton. Hydro-
carbon height (HCH) at surface condi-
tions for Chase Group (B), Council Grove
Group (C), and combined, Wolfcamp
(D). Wolfcamp volume map varies little
from the Chase because the Council
Grove volume is small relative to the
Chase (note the smaller scale range
and contour interval for the Council
Grove). Areas of high Chase OGIP are
coincident with Krider ooid shoal
complex illustrated in Figure 38D – F.
characterization and modeling work-
flow for a giant field. Core-based cali-
bration of neural net prediction
of lithofacies using wire-line-log
signatures, coupled with geologic-
constraining variables, provided ac-
curate lithofacies models at well to
field scales. Differences in petrophy-
sical properties among lithofacies and
within a lithofacies among different
porosities illustrate the importance of
integrated lithologic-petrophysical
modeling and of the need for closely
defining these properties and their
relationships. Lithofacies models,
coupled with lithofacies-dependent
petrophysical properties, allowed
the construction of a 3-D geomodel
for the Hugoton that has been ef-
fective at the well, section (1 mi2;
2.6 km2) and multisection scales.
The model provided a tool to pre-
dict lithofacies and petrophysical
properties distribution, water satura-
tions, and OGIP. It will likely provide
a quantitative basis for evaluating
remaining gas in place, particularly
in low-permeability intervals, and help direct field
management and production practices that will po-
tentially enhance ultimate recovery. The reservoir char-
acterization and modeling from pore to field scale
discussed provides a comprehensive lithological and
petrophysical view of a mature giant Permian gas sys-
tem. Both the knowledge gained and the techniques
and workflow employed have implications for un-
derstanding and modeling similar reservoir systems
worldwide.
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 351
ACKNOWLEDGMENTS
We are thankful for the support from partners
in the Hugoton Asset Management Project, includ-
ing Anadarko Petroleum Corporation, BP America
Production Company, ConocoPhillips Company, Ci-
marex Energy Company, EOG Resources Inc., Exxon-
Mobil Production Company, Medicine Bow En-
ergy Corporation, Osborn Heirs Company, OXY
U.S.A., Inc., and Pioneer Natural Resources U.S.A.,
Inc., and from the Kansas Geological Survey. We
thank Nathan Winters for his assistance on core de-
scriptions, Raymond Sorenson for insightful discus-
sions, Shane Seals in work on earlier models, and
geoPlus Corporation and Schlumberger for provid-
ing the software. Reviewers Eugene Rankey, Sean
Guidry, and Krishnan Srinivasan provided help-
ful comments and suggestions that improved this
manuscript.
REFERENCES CITED
Asquith, G., and D. Krygowski, 2004, Basic well log analy-
sis, 2d ed.: AAPG Methods in Exploration, v. 16, 244 p.
Baars, D. L., ed., 1994, Revision of stratigraphic nomen-
clature in Kansas: Kansas Geological Survey Bulletin,
v. 230, 80 p.
Berg, R. R., 1975, Capillary pressures in stratigraphic traps:
AAPG Bulletin, v. 59, p. 939 – 956.
Boardman II, D. R., and M. K. Nestell, 2000, Outcrop-based
sequence stratigraphy of the Council Grove Group of
the midcontinent, in K. S. Johnson, ed., Platform car-
bonates in the southern midcontinent, 1996 Sympo-
sium: Oklahoma Geological Survey Circular 101, p. 275–
306.
Bohling, G. C., and J. H. Doveton, 2000, Kipling.xla: An
Excel add-in for nonparametric regression and classi-
fication, Kansas Geological Survey: http://www.kgs
.ku.edu/software/Kipling/Kipling1.html (accessed
December 31, 2005).
Byrnes, A. P., 2003, Aspects of permeability, capillary
pressure, and relative permeability properties and
distribution in low-permeability rocks important to
evaluation, damage, and stimulation: Proceedings of
the Rocky Mountain Association of Geologists Petro-
leum Systems and Reservoirs of Southwest Wyoming
Symposium, Denver, Colorado, September 19, 2003,
12 p.
Byrnes, A. P., M. K. Dubois, M. Magnuson, 2001, Western
tight gas carbonates: Comparison of Council Grove
Group, Panoma field, southwest Kansas and western low
permeability sandstones (abs.): AAPG Annual Meeting
Program, v. 10, p. A31.
Caldwell, C. D., 1991, Cyclic deposition of the Lower Perm-
ian, Wolfcampian, Chase Group, western Guymon–
Hugoton field, Texas County, Oklahoma, in W. L.
Watney, A. W. Walton, C. G. Caldwell, and M. K.
Dubois, organizers, Midcontinent Core Workshop on
Integrated Studies of Petroleum Reservoirs in the
Midcontinent: Midcontinent AAPG Section Meeting,
Wichita, Kansas, p. 57 – 75.
Deutsch, C. V., and A. G. Journel, 1998, Geostatistical soft-
ware library and user’s guide: Oxford, Oxford Univer-
sity Press, 369 p.
Dubois, M. K., and R. H. Goldstein, 2005, Accommodation
model for Wolfcamp (Permian) redbeds at the updip
margin of North America’s largest onshore gas field
(abs.): Proceedings AAPG 2005 Annual Convention,
June 19– 21, Calgary, Alberta, Canada, and Kansas Geo-
logical Survey Open-file Report 2005-25: http://www
.kgs.ku.edu/PRS/AAPG2005/2005-25/index.html
(accessed December 31, 2005).
Dubois, M. K., A. P. Byrnes, G. C. Bohling, S. C. Seals, and
J. H. Doveton, 2003, Statistically-based lithofacies pre-
dictions for 3-D reservoir modeling: Examples from
the Panoma (Council Grove) field, Hugoton embay-
ment, southwest Kansas (abs.): AAPG Annual Meet-
ing Program, v. 12, p. A44, and Kansas Geological Sur-
vey Open-file Report 2003-30, 3 panels: http://www
.kgs.ku.edu/PRS/publication/2003/ofr2003-30/index
.html (accessed October 10, 2005).
Dubois, M. K., A. P. Byrnes, and G. C. Bohling, 2005,
Geologic model for the giant Hugoton and Panoma
fields (abs): AAPG Midcontinent Section Meeting,
Oklahoma City, Oklahoma: http://www.kgs.ku.edu
/PRS/Poster/2005/MidcontAAPG/index.html (accessed
October 10, 2005).
Dunham, R. J., 1962, Classification of carbonate rocks accord-
ing to depositional texture, in W. E. Ham, ed., Classifi-
cation of carbonate rocks: AAPG Memoir 1, p. 108 –121.
Dutton, S. P., E. K. Kim, C. L. Broadhead, W. D. Raatz, S. C.
Ruppel, and C. Kerans, 2005, Play analysis and digital
portfolio of major oil reservoirs in the Permian Basin:
Bureau of Economic Geology Reports of Investigations,
RI0271, 302 p.
Fetkovitch, M. J., D. J. J. Ebbs Jr., and J. J. Voelker, 1994,
Multiwell, multilayer model to evaluate infill-drilling
potential in the Oklahoma Hugoton field: 65th So-
ciety of Petroleum Engineers Annual Technical Con-
ference and Exhibition, New Orleans, Louisiana, SPE
Paper 20778, p. 162 – 168.
Folk, R. L., 1954, The distinction between grain size and
mineral composition in sedimentary rock nomencla-
ture: Journal of Geology, v. 62, p. 344 – 359.
Garlough, J. L., and G. L. Taylor, 1941, Hugoton gas field,
Grant, Haskell, Morton, Stevens, and Seward counties,
Kansas, and Texas county, Oklahoma, in A. I. Levorsen,
ed., Monograph Stratigraphic type oil fields: AAPG,
p. 78 – 104.
Gaymard, R., and A. Poupon, 1968, Response of neutron
and formation density logs in hydrocarbon bearing
formations: The Log Analyst, v. 9, no. 5, p. 3 – 12.
George, B. K., C. Torres-Verdin, M. Delshad, R. Sigal, F.
352 / Dubois et al.
Zouioueche, and B. Anderson, 2004, Assessment of in-
situ hydrocarbon saturation in the presence of deep
invasion and highly saline connate water: Petrophysics,
v. 45, no. 2, p. 141– 156.
Grammer, G. M., G. P. Eberli, F. S. P. van Buchem, G. M.
Stevenson, and P. Homewood, 1996, Application of
high-resolution sequence stratigraphy to evaluate lat-
eral variability in outcrop and subsurface; Desert Creek
and Ismay intervals, Paradox basin, in M. W. Longman
and M. D. Sonnenfeld, eds., Paleozoic systems of the
Rocky Mountain region: Rocky Mountain Section So-
ciety for Sedimentary Geology, p. 235 – 266.
Hastie, T., R. Tibshirani, and J. Friedman, 2001, The ele-
ments of statistical learning: Data mining, inference,
and prediction: New York, Springer, 533 p.
Hecker, M. T., M. E. Houston, and J. D. Dumas, 1995,
Improved completion designs in the Hugoton field
utilizing multiple gamma emitting tracers: Society of
Petroleum Engineers Annual Technical Conference and
Exhibition, Dallas, Texas, SPE Paper 30651, p. 223 –235.
Hemsell, C. C., 1939, Geology of Hugoton gas field of south-
western Kansas: AAPG Bulletin, v. 23, no. 7, p. 1054–
1067.
Heyer, J. F., 1999, Reservoir characterization of the Council
Grove Group, Texas County, Oklahoma, in D. F. Mer-
riam, ed., AAPG Midcontinent Section Meeting Trans-
actions, Geosciences for the 21st Century, p. 71 – 82.
Hough, E. W., M. J. Rzasa, and B. B. Wood, 1951, Interfacial
tensions at reservoir pressures and temperatures;
apparatus and the water-methane system: American
Institute of Mechanical Engineering Petroleum Trans-
actions, Technical Publication 3019, v. 192, p. 57 – 60.
Hubbert, M. K., 1953, Entrapment of petroleum under hy-
drodynamic conditions: AAPG Bulletin, v. 37, p. 1954 –
2026.
Hubbert, M. K., 1967, Application of hydrodynamics to
oil exploration: 7th World Petroleum Congress Pro-
ceedings, Mexico City: Amsterdam, Elsevier Publishing
Co. Ltd, v. 1B, p. 59 – 75.
Jennings Jr., H. Y., and G. H. Newman, 1971, The effect of
temperature and pressure on the interfacial tension of
water against methane-normal decane mixtures: Trans-
actions of the American Institute of Mining Metallur-
gical and Petroleum Engineers, v. 251, p. 171 – 175.
Kluth, C. F., 1986, Plate tectonics of the ancestral Rocky
Mountains, in J. A. Peterson, ed., Paleotectonics and
sedimentation of the Rocky Mountains, United States:
AAPG Memoir 41, p. 353 – 369.
Kluth, C. F., and P. J. Coney, 1981, Plate tectonics of the
ancestral Rocky Mountains: Geology, v. 9, no. 1, p. 10 –
15.
Konnert, G., A. M. Afifi, S. A. Al-Hajri, K. de Groot, A. A. Al
Naim, and H. J. Droste, 2001, Paleozoic stratigraphy
and hydrocarbon habitat of the Arabian plate, in
M. W. Downey, J. C. Threet, and W. A. Morgan, eds.,
Petroleum provinces of the twenty-first century:
AAPG Memoir 74, p. 483 – 515.
Luczaj, J. A., and R. H. Goldstein, 2000, Diagenesis of the
Lower Permian Krider Member, southwest Kansas,
U.S.A.: Fluid-inclusion, U-Pb, and fission-track evi-
dences for reflux dolomitization during latest Permian
time: Journal of Sedimentary Research, v. 70, no. 3,
p. 762 – 773.
Mason, J. W., 1968, Hugoton and Panhandle field, Kansas,
Oklahoma and Texas, in W. B. Beebe and B. F. Curtis,
eds., Natural gases of North America, v. 2: AAPG
Memoir 9, p. 1539 – 1547.
Mazzullo, S. J., C. S. Teal, and C. A. Burtnett, 1995, Facies
and stratigraphic analysis of cyclothemic strata in the
Chase Group (Permian Wolfcampian), south-central
Kansas, in N. J. Hyne, ed., Sequence stratigraphy of
the mid-continent: Tulsa Geological Society Special
Publication 4, p. 217 – 248.
McGillivray, J. G., and M. I. Husseini, 1992, The Paleozoic
petroleum geology of central Arabia: AAPG Bulletin,
v. 76, p. 1491 – 1506.
Oberst, R. J., P. P. Bansal, and M. F. Cohen, 1994, 3-D
reservoir simulation results of a 25-square mile study
area in Kansas Hugoton gas field: Society of Petroleum
Engineers Mid-Continent Gas Symposium, SPE Paper
27931, p. 137 – 147.
Olson, T. M., J. A. Babcock, and P. D. Wagner, 1996,
Geologic controls on reservoir complexity, Hugoton
giant gas field, Kansas, in D. L. Swindler and C. P.
Williams, compilers, AAPG Mid-continent Section Meet-
ing Transactions: Mid-continent Section, p. 189–198.
Olson, T. M., J. A. Babcock, K. V. K. Prasad, S. D. Boughton,
P. D. Wagner, M. K. Franklin, and K. A. Thompson,
1997, Reservoir characterization of the giant Hugo-
ton gas field, Kansas: AAPG Bulletin, v. 81, p. 1785 –
1803.
Olszewski, T. D., and M. E. Patzkowsky, 2003, From cy-
clothems to sequences: The record of eustacy and cli-
mate on an icehouse eperic platform (Pennsylvanian –
Permian), North American Midcontinent: Journal of
Sedimentary Research, v. 73, no. 1, p. 15 – 30.
Parham, K. D., and J. A. Campbell, 1993, PM-8. Wolf-
campian shallow shelf carbonate — Hugoton embay-
ment, Kansas and Oklahoma, in D. G. Bebout, ed., Atlas
of major midcontinent gas reservoirs: Gas Research
Institute, p. 9 – 12.
Parrish, J. T., 1995, Geologic evidence of Permian climate,
in P. A. Scholle, T. M. Peryt, and D. S. Ulmer-Scholle,
eds., The Permian of northern Pangea: v. I. Paleoge-
ography, paleoclimate, stratigraphy: Berlin, Germany,
Springer-Verlag, p. 53 – 61.
Parrish, J. T., and E. Peterson, 1988, Wind directions pre-
dicted from global circulation models and wind di-
rections determined from eolian sandstones of the
western United States — A comparison: Sedimentary
Geology, v. 56, p. 261 – 282.
Perry, W. J., 1989, Tectonic evolution of the Anadarko
basin region, Oklahoma: U.S. Geological Survey Bulle-
tin, v. 1866-A, p. A1 – 16.
Pippin, L., 1970, Panhandle–Hugoton field, Texas–Okla-
homa–Kansas— The first fifty years, in M. T. Halbouty,
 Geologic and Petrophysical Modeling of the Hugoton Gas Field, Kansas and Oklahoma / 353
ed., Geology of giant petroleum fields: AAPG Memoir 14,
p. 204–222.
Puckette, G. R., D. R. Boardman II, and Z. Al-Shaieb, 1995,
Evidence for sea-level fluctuation and stratigraphic se-
quences in the Council Grove Group (Lower Permian)
Hugoton embayment, southern mid-continent, in N. J.
Hyne, ed., Sequence stratigraphy of the mid-continent:
Tulsa Geological Society Special Publication 4, p. 269–
290.
Purcell, W. R., 1949, Capillary pressure — Their measure-
ments using mercury and the calculation of perme-
ability therefrom: American Institute of Mechanical
Engineers Petroleum Transactions, v. 186, p. 39 – 48.
R Development Core Team, 2005, R: A language and en-
vironment for statistical computing: R Foundation for
Statistical Computing, Vienna, Austria: http://www
.R-project.org (accessed December 31, 2005).
Rankey, E. C., 1997, Relations between relative changes in
sea level and climate shifts; Pennsylvanian – Permian
mixed carbonate-siliciclastic strata, western United
States: Geological Society of America Bulletin, v. 109,
no. 9, p. 1089 – 1100.
Rascoe Jr., B., 1968, Permian system in western midconti-
nent: Mountain Geologist, v. 5, p. 127 – 138.
Rascoe Jr., B., and F. J. Adler, 1983, Permo-Carboniferous
hydrocarbon accumulations, midcontinent, U.S.A.:
AAPG Bulletin, v. 67, p. 979 – 1001.
Ryan, T. C., M. J. Sweeney, and W. H. Jamieson Jr., 1994,
Individual layer transient tests in low-pressure, multi-
layered reservoirs: Society of Petroleum Engineers
Mid-Continent Gas Symposium, Amarillo, Texas, SPE
Paper 27928, p. 99 – 113.
Scotese, C. R., 2004, A continental drift flipbook: Journal
of Geology, v. 112, p. 729 – 741.
Siemers, W. T., and W. M. Ahr, 1990, Reservoir facies, pore
characteristics, and flow units: Lower Permian Chase
Group, Guymon – Hugoton field, Oklahoma, Society
of Petroleum Engineers Proceedings 65th Annual Tech-
nical Conference and Exhibition, New Orleans, Louisi-
ana, September 23 – 26, 1990, SPE Paper 20757, p. 417 –
428.
Soreghan, G. S., 2002, Sedimentologic-magnetic record of
western Pangean climate in upper Paleozoic loessites
(lower Cutler beds, Utah): Geological Society of Ameri-
ca Bulletin, v. 114, no. 8, p. 1019 – 1035.
Sorenson, R. P., 2005, A dynamic model for the Permian
Panhandle and Hugoton fields, western Anadarko
basin: AAPG Bulletin, v. 89, no. 7, p. 921 – 938.
Venables, W. N., and B. D. Ripley, 2002, Modern applied
statistics with S, 4th ed.: New York, Springer, 512 p.
Washburn, E. W., 1921, A method of determining the dis-
tribution of pore sizes in a porous material: Proceed-
ings of the National Academy of Sciences, v. 7, p. 115 –
116.
Weber, L. J., F. M. Wright, J. F. Sarg, E. Shaw, L. P. Harman,
J. B. Vanderhill, and D. A. Best, 1994, Reservoir de-
lineation and performance; application of sequence
stratigraphy and integration of petrophysics and en-
gineering data, Aneth field, southeast Utah, U.S.A., in
E. L. Stout and P. M. Harris, eds., Hydrocarbon reser-
voir characterization; geologic framework and flow
unit modeling: Society of Sedimentary Geology, p. 1 –
29.
Winters, N. D., M. K. Dubois, and T. R. Carr, 2005, Depo-
sitional model and distribution of marginal marine
sands in the Chase Group, Hugoton gas field, southwest
Kansas and Oklahoma Panhandle (abs.): AAPG Mid-
continent Section Meeting, Oklahoma City, Oklahoma:
http://www.kgs.ku.edu/PRS/Poster/2005/MidcontAAPG
/index.html (accessed October 10, 2005).
Zeller, D. E., ed., 1968, The stratigraphic succession in Kan-
sas: Kansas Geological Survey Bulletin, v. 169, 81 p.
 10
Ruppel, S. C., and R. H. Jones, 2006, Key role of outcrops and cores
in carbonate reservoir characterization and modeling, Lower Permian
Fullerton field, Permian basin, United States, in P. M. Harris and L. J.
Weber, eds., Giant hydrocarbon reservoirs of the world: From rocks
to reservoir characterization and modeling: AAPG Memoir 88/SEPM
Special Publication, p. 355 – 394.

Key Role of Outcrops and

Cores in Carbonate Reservoir
Characterization and Modeling,
Lower Permian Fullerton Field,
Permian Basin, United States
Stephen C. Ruppel
Bureau of Economic Geology, Jackson School of Geosciences, University of Texas at Austin,
Austin, Texas, U.S.A.

Rebecca H. Jones
Bureau of Economic Geology, Jackson School of Geosciences, University of Texas at Austin,
Austin, Texas, U.S.A.

ABSTRACT

T
he analysis of reservoir sequence and cycle stratigraphy, of depositional
and diagenetic facies, and of the interrelationships between these
attributes and reservoir properties is key to the construction of an accurate
reservoir framework needed for reservoir modeling and improved imaging of
remaining hydrocarbons. Fundamental steps in such a rock-based process of
model construction applied at Fullerton Clear Fork field, a shallow-water car-
bonate platform reservoir of middle Permian age, included (1) creating and ap-
plying an analogous outcrop depositional model; (2) describing and interpreting
subsurface core and log data in terms of this initial model; (3) defining the
sequence-stratigraphic architecture of the reservoir section; (4) developing a cycle-
based reservoir framework; and (5) defining controls, interrelationships, and dis-
tribution of porosity and permeability. Data used in this analysis included cores,
thin sections, three- and two-dimensional seismic data, borehole image logs, and
outcrop models.
Key geological elements addressed and incorporated into the models include
stratal architecture, differential dolomitization, karst fill, mineralogical variations,
and rock-fabric distribution. These components were used to constrain inter-
pretation and definition of flow units, permeability distribution, and saturation.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215882M88698

355
356 / Ruppel and Jones
In addition to resulting in improved and more geologically realistic reservoir
models, the rock-based methods used in this study provide key insights into the
formation, characterization, and interpretation of carbonate platform reservoirs;
these insights have widespread application worldwide.
INTRODUCTION
The Clear Fork Group (Figure 1) in the Permian
basin comprises a thick (as much as 800 m [2500 ft])
succession of dominantly shallow-water platform
carbonates that were deposited across west Texas and
New Mexico during the Early Permian (Leonardian).
Reservoirs developed in these carbonates (Figure 2)
have accounted for more than 3.2 billion bbl of oil
production (Dutton et al., 2005), more than 10% of
the total recovered from the Permian basin to date.
Despite this substantial production, estimates of origi-
nal oil in place (OOIP) indicate that, overall, Leonar-
dian reservoirs contained more than 14.5 billion bbl
of oil at discovery. Recovery efficiency is thus only
about 22%, considerably below the 32% average for
carbonate reservoirs in the Permian basin (Tyler and
Banta, 1989; Holtz and Garrett, 1990). Recovery from
the shallow-water platform reservoirs of the Clear Fork
Group has been even less efficient. Holtz et al. (1992)
estimated a recovery efficiency of only 18% of OOIP
for these reservoirs.
To recover the remaining oil in Clear Fork res-
ervoirs, operators must turn to increasingly sophisti-
cated recovery technologies, e.g., water flooding, gas
injection, horizontal wells, etc. To effectively deploy
these technologies, however, it is critical that an accu-
rate reservoir framework first be constructed to form
the basis for modeling and interpreting past, present,
and future recovery operations. This report details
approaches used to develop such a framework in the
Fullerton Clear Fork field in west Texas (Figures 2, 3).
By any standard, Fullerton is a giant. As of 2002,
cumulative production stood at about 310 million
bbl, about 18% of the approximately 1.69 billion bbl
of OOIP (Wang and Lucia, 2004). Like many mature
Clear Fork reservoirs in the Permian basin, Fullerton
field has undergone a major decline in oil production
rate and a major increase in water production rates
for several years. An understanding of the deposi-
tional and diagenetic facies, the cycle and sequence
stratigraphy, and the architecture of these stratigraph-
ic elements is the crucial first step to determining the
probable oil distribution at discovery and defining the
best strategies for recovering the sizeable remaining
oil volume.
METHODS
Data and interpretations presented in this report
are based on investigation of 29 cores totaling 4384 m
(14,383 ft), the examination of more than 1700 rock
thin sections, and the correlation of approximately
45,000 stratigraphic tops in more than 800 wells.
The basic procedure followed in this study to develop
a full-field sequence-stratigraphic model and reser-
voir framework is as follows:
(1) identify depositional facies
and vertical facies-stacking
patterns and cycles in cores;
(2) calibrate facies and cycle
patterns to wire-line logs;
(3) construct two-dimensional
(2-D) cross sections of core/
log data sections; (4) define
FIGURE 1. Leonardian strati-
graphic section in the Permian
basin including the Clear Fork
Group and analogous units in
New Mexico and in outcrop.
Productive interval at Fullerton
field is shown in color.
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 357
FIGURE 2. Regional map of the Permian basin showing the location of the Fullerton field and major reservoirs producing
from Leonardian rocks.
2-D cycle and sequence stratigraphy from core and
log sections; and (5) extrapolate 2-D cycle and se-
quence correlations into three-dimensional (3-D) space
by correlating well logs. Stratigraphic sequences were
tied to available 3-D seismic data to check their ac-
curacy and geometry. Conventional core analysis data
were available for all cores; these data were used to
determine relationships between facies, cyclicity, and
porosity and permeability development. Figure 3
shows the distribution of cores and 3-D seismic data
in Fullerton field.
PREVIOUS WORK
Mazzullo (1982) and Mazzullo and Reid (1989)
presented overviews of lower Leonardian stratigra-
phy and depositional systems in the Midland basin.
Presley and McGillis (1982; see also Presley, 1987)
documented the highly cyclic, predominantly evap-
oritic facies of the upper Leonardian Glorieta and
Upper Clear Fork units in the Texas Panhandle. Ruppel
(1992, 2002) described the facies, cyclicity, and dia-
genesis in the Glorieta and Upper Clear Fork at Mona-
hans Clear Fork field on the Central Basin platform
and postulated that reservoir development was caused
by cyclic deposition and diagenesis driven by episodic
sea level rise and fall. Atchley et al. (1999) described
Clear Fork facies at Robertson field at the north end of
the Central Basin platform and proposed a similar
model for structural control over facies deposition
and reservoir development. Ruppel et al. (2000) de-
scribed outcrop equivalents of the producing subsur-
face Clear Fork reservoirs from outcrops in the Sierra
Diablo Mountains of west Texas. Kerans et al. (2000)
358 / Ruppel and Jones
FIGURE 3. Structure of the Fullerton Clear Fork field.
Datum is a prominent subtidal flooding event near the
base of the Tubb Formation (see Figures 4, 5).
documented the depositional setting, facies, and archi-
tecture of the Abo from outcrops in the Sierra Diablo
(Hudspeth and Culberson counties, Texas) and showed
that karsting has had a major effect on both the Abo
and the overlying Lower Clear Fork succession.
GEOLOGIC SETTING
The Fullerton Clear Fork field is the largest of a
large number of fields developed in the Leonardian
series on the Central Basin platform of the Permian
basin (Figure 2). The field covers an area of about
14,000 ha (35,000 ac); the reservoir column averages
about 210 m (700 ft) in thickness at a depth of about
2040 m (6700 ft) (Figures 2, 3). In some parts of the
Permian basin, oil production is obtained from the
entire Leonardian section (Figure 1). The productive
reservoir section at Fullerton, however, is essentially
restricted to the Lower Clear Fork, Wichita, and Abo
stratigraphic units (Figure 4), although very minor
production has been reported locally from the Upper
Clear Fork section. By far, the bulk of the oil produc-
tion has come from the lower two-thirds of the Lower
Clear Fork and the Wichita sections of the reservoir.
Even where oil saturated, the Abo has proven difficult
to exploit in part because of its active water drive (as
opposed to the pressure depletion drive that charac-
terizes the overlying parts of the reservoir) and the
fact that it is at or near the oil-water contact through-
out most of the field.
Regionally, the Leonardian is dominated by shallow-
water platform carbonates. Each of the component
stratigraphic units contains updip peritidal tidal-flat
carbonates and downdip subtidal carbonates. Min-
eralogy in each is dominated by dolomite and an-
hydrite. Limestone is relatively uncommon; however,
where present, it is most common in the lower part of
the Leonardian section and in distal, downdip sec-
tions (Ruppel, 2002). The reservoir section at Fuller-
ton is generally consistent with this regional pattern,
but it does contain a higher volume of limestone than
most other Leonardian platform reservoir successions
in the Permian basin.
Structurally, Fullerton field is developed over a
large, compound structural high (Figure 3) that re-
flects deep-seated faulting and differential uplift of
the area that began in the Pennsylvanian (Jones and
Ruppel, 2004). Deeper oil production comes from
block-faulted Silurian (Wristen Group) carbonates in
the southern and central parts of the fields. The Clear
Fork reservoir seal is provided by evaporite-rich car-
bonates of the Upper Clear Fork and Glorieta.
The reservoir is currently drilled to well spacings of
16 to 4 ha (40 to 10 ac) and is under active water
flood. The greatest well density is in the northern
part of the field (mostly 8- and 4-ha (20- and 10-ac)
well spacings); poorer well control exists in the
southern half of the field. In general, these closely
spaced wells provide good control for definition of
stratigraphic horizons and reservoir attributes. In
many parts of the field, however, especially along the
western edge of the field and in the southern half of
the field, wells are represented only by poor-quality
(old gamma-ray and neutron) logs and cannot be
correlated or interpreted with any precision.
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 359
OUTCROP ANALOGS
Studies of analogous reservoir outcrops in the
Sierra Diablo in Hudspeth and Culberson counties,
Texas (Figure 2), provide important insights into the
geological controls on reservoir development in the
Fullerton reservoir and on the reservoir architecture.
The Leonardian of the Sierra Diablo contains direct
analogs of all major reservoir intervals at Fullerton
field, including the shallow-water platform carbon-
ates of the Lower Clear Fork Group (Fitchen et al.,
1995; Ruppel et al., 2000) and Wichita units and the
FIGURE 4. Type log of producing res-
ervoir section at Fullerton Clear Fork
field showing cyclicity and general
facies distribution. Shaded areas of the
PE log indicate the local presence of
limestone in the dolomite-dominated
section. Zones of elevated corrected
gamma-ray (CGR) log response indicate
the presence of silt and clay associated
with tidal-flat facies. SGR = spectral
gamma ray; PE = photoelectric; Neu =
neutron; Den = density.
karsted, platform-margin subtidal
carbonates of the Abo (Fitchen et al.,
1995; Kerans et al., 2000). Styles of
depositional architecture and facies
development are very representative
of subsurface succession and, thus,
form excellent, if not essential, mod-
els for interpreting sparser subsur-
face data sets. Key observations es-
tablished from these outcrops are
described below.
Abo
Integrated studies of outcrops and
subsurface data sets indicate that the
Abo represents the basal depositional
sequence (sequence L1) of the Leonar-
dian (Fitchen et al., 1995). Studies of
Abo outcrops in the Sierra Diablo
(Fitchen et al., 1995; Kerans et al.,
2000) have demonstrated three im-
portant aspects of this succession
in the Permian basin: (1) it consists
of dominantly open-marine, outer-
platform facies; (2) it displays clino-
formal architecture; and (3) it is over-
printed by karst features (sinkholes,
caves, cave fill, and collapse features).
The dominance of clinoformal, outer-ramp, fusulinid-
crinoid packstones and wackestones and less common
ramp-crest, ooid-peloid, grain-rich packstones and
grainstones in the Abo contrasts with the flat-lying,
alternating tidal-flat and shallow subtidal wackestone-
packstone successions of the Wichita and Lower Clear
Fork. The top-lapping clinoforms of the Abo document
rapid basinward progradation, a forced regression prob-
ably caused by a rapid fall in sea level. Karsting in the
Abo, although initiated at the exposed top of the Abo
during sea level fall, is manifested downsection in
the Abo by caves and sinkholes and upsection in the
360 / Ruppel and Jones
overlying Lower Clear Fork as collapse features. Where
karsting and associated thickness variations are de-
veloped, largely in platform-marginal settings, the
contact between the Abo and Lower Clear Fork is a
relatively sharp and undulating, unconformable sur-
face. Updip, karsting is less apparent, and the contact
is less pronounced.
Lower Clear Fork
The outcropping Lower Clear Fork succession in
the Sierra Diablo represents a single depositional se-
quence (L2). Key elements of this succession are
(1) basal backstepping tidal-flat deposits that locally
fill relief on the underlying, karsted Abo surface;
(2) an updip succession of amalgamated tidal-flat
facies; (3) a cyclic, downdip succession of alternating
tidal-flat and midramp, subtidal facies; and (4) an over-
all backstepping (upward-deepening) trend (Fitchen
et al., 1995; Kerans et al., 2000; Ruppel et al., 2000).
Lower Clear Fork deposits are characterized by
alternating peritidal, tidal-flat deposits and subtidal,
skeletal wackestones and packstones that document
cyclic rise and fall of sea level at the high-frequency
sequence and cycle scale. These cycles, which average
about 6 m (20 ft) in thickness, display consistent
patterns of facies stacking (cycle-base skeletal wacke-
stones and overlying peloidal grain-rich packstones)
and appear to be widely continuous (Ruppel et al.,
2000). These midramp, shallow subtidal platform
rocks pass downdip into clinoformal fusulinid-crinoid
wackestones and packstones of the outer ramp-slope
in less than 3 km (2 mi) basinward. Updip (landward),
the Lower Clear Fork is characterized by amalgamated,
inner-platform, peritidal, tidal-flat deposits. These
updip tidal flats are analogous to Wichita facies in
the subsurface. The absence of shallow-water high-
stand deposits at the top of the Lower Clear Fork
suggests a rapid, perhaps forced, regression, followed
by exposure and possible erosion at the top of the
L2 sequence.
FULLERTON RESERVOIR FACIES
AND STRATIGRAPHY
The productive reservoir section at Fullerton field
includes parts of the Abo, Wichita, and Lower Clear
Fork formations (Figures 4, 5). Like many stratigraph-
ic names used in the subsurface, each of these units is
commonly considered to display both regional time
equivalency and facies constancy across the region.
In other words, each is considered to be both a rock-
stratigraphic and a time-stratigraphic unit. In fact,
these names are best considered rock-stratigraphic
terms (essentially facies) at the formation level of
nomenclature. To place these units in their proper
perspective, however, it is best to consider both facies
and time interrelationships. In this report, we treat
these named stratigraphic units (i.e., formations) as
facies or rock-stratigraphic units but place them in a
sequence-stratigraphic (i.e., time-stratigraphic) frame-
work that has been developed from previous studies
(e.g., Fitchen et al.,1995; Kerans et al., 2000; Ruppel
et al., 2000). On the basis of these studies and data
from Fullerton field, we interpret the Abo and the lower
part of the Wichita Formation to be time-equivalent
facies (or systems tracts) of the earliest composite
(third-order) depositional sequence of the Leonardian
(sequence L1). The overlying sequence (L2) comprises
updip peritidal deposits of the upper Wichita Forma-
tion and downdip, dominantly subtidal deposits of the
Lower Clear Fork Formation (Figure 5). Sequence L3
contains basal transgressive systems tract (TST) silic-
iclastics of the Tubb Formation and carbonates of part
of the overlying Upper Clear Fork (Figure 4).
Correlations of usable wire-line-log suites (about
850) show that across most of the area of the Fuller-
ton Clear Fork unit (Figure 3), stratigraphic units are
relatively isopachous. This is consistent with the de-
positional setting of the field area on the Central
Basin platform, a broad, flat carbonate platform not
unlike the modern-day Bahamas platform.
The generally isopachous nature of the Wichita and
Lower Clear Fork is confirmed by 3-D and 2-D seismic
data. These data, however, show that the Wichita and
Abo vary considerably in thickness across the field.
Wichita deposits thin, and Abo deposits thicken to
the east and southeast (Figure 6). Synthesis of core,
wire-line, and seismic data suggests that this recipro-
cal thickness relationship is the result of facies change;
i.e., the Wichita represents the updip, shallower water
(inner-platform) equivalent of the distal, deeper water
(outer platform to slope) Abo. Support for this con-
clusion is presented in subsequent sections.
Facies and Depositional Setting
Reservoir facies encountered at Fullerton field are
typical of those observed throughout most carbonate
platform successions of Leonardian and early Gua-
dalupian age in the Permian basin. The character-
istics of these and similar facies have been document-
ed by many authors, including Bebout et al. (1987),
Ruppel and Cander (1988a, b), Garber and Harris
(1990), Longacre (1990), Kerans et al. (1994; see also
Kerans and Fitchen, 1995; Kerans and Kempter, 2002),
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 361

FIGURE 5. Northwest – southeast cross section (AA0 ) across the Fullerton field area showing sequence architecture and
general facies development based on cored well control. Line of section shown in Figure 3. Depth in feet.

FIGURE 6. Three-dimensional
seismic section (in time) across
the southern part of Fullerton
field showing the seismic defi-
nition of the Clear Fork reservoir
section. Yellow lines are time-
line boundaries; dotted line de-
fines the top of the Abo Forma-
tion. Note that whereas the
upper Wichita and Lower Clear
Fork intervals are essentially iso-
pachous and continuous across
the field, the lower Wichita and
Abo display reciprocal thick-
ness relationships. Dashed lines
are top-lapping Abo clinoforms.
UCF = Upper Clear Fork; LCF =
Lower Clear Fork.
362 / Ruppel and Jones
and Ruppel and Bebout (2001). Leonardian facies
successions have been documented both in outcrop
(Ruppel et al., 2000) and in the subsurface (Atchley
et al., 1999; Ruppel, 2002).
Facies groupings defined in this study and pre-
sented below were defined with three goals in mind:
(1) to document the characteristics of distinct depo-
sitional settings; (2) to document widespread and po-
tentially correlatable sediment packages; and (3) to
create a sound geological basis for assessing relation-
ships between rock textures and fabrics and petro-
physics. In many ways, it is the last of these goals that
is most important because it provides the foundation
for reservoir modeling. However, as a result of this,
facies characteristics used for subdivision are domi-
nantly small-scale matrix properties, e.g., grain size
and shape, grain type. Larger scale features, such as
fractures, anhydrite nodules, and evidence of burrow-
ing, although noted, are not used for facies subdivision.
Twelve facies can be at least locally identified in
the Fullerton reservoir succession. Most of these are
intergradational with one or more others. In many
cases, their distinction is somewhat subjective, based
FIGURE 7. Core and thin-section photo-
graphs of typical Wichita tidal-flat facies.
(A) Slab photo of peritidal mudstone-
wackestone showing weak laminations
and local burrowing. Core is 10 cm
(4 in.) wide. Exxon FCU 5927, depth:
2140 m (7021 ft). (B) Photomicrograph
of peritidal mudstone-wackestone fa-
cies showing abundant intercrystalline
porosity. Exxon FCU 6122, depth: 2162 m
(7092 ft). Porosity: 13.8%; permeabil-
ity: 1.6 md. (C) Slab photo of clay-rich
carbonate mudstone facies in the Wichita.
Core is 10 cm (4 in.) wide. Pan American
FM-1, depth: 2234 m (7329 ft). Note
underlying fenestral mudstone.
on differences in grain preservation,
diagenesis, grain size, etc. Most can
be found in any part of the reservoir
succession.
Peritidal Mudstone-wackestone
These rocks, which are most abun-
dant in the Wichita but also locally
common in the Lower Clear Fork, are
massive to parallel-laminated mud-
rich rocks that only sparsely con-
tain grains other than a few peloids
(Figure 7A, B). They are most typi-
cally associated with the exposed tidal-flat facies
and the clay-rich carbonate mudstone facies and are
interpreted to represent peritidal deposition on a very
low-energy tidal flat. They are dominantly dolomi-
tized, and where dolomite, they may display very high
porosities (as much as 15%). Because this porosity is
caused by intercrystalline pores in fine-crystalline
dolomite, however, the permeability is generally low.
In limestone, the facies invariably exhibits very low
porosity (typically less than 2%) and permeability.
Gamma-ray response is typically medium to high and
highly variable in these rocks because of the local
presence of clay minerals.
Clay-rich Carbonate Mudstone
These rocks are found intimately associated with
rocks of the peritidal mudstone-wackestone facies in
the Wichita. They are typically dark gray to nearly
black and are found in beds 1– 8 cm (0.4 –3.1 in.) in
thickness (Figure 7C). Where cyclic upward-shallow-
ing successions can be identified, these deposits are
found at or near cycle tops in both the Lower Clear Fork
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 363
and the Wichita. Although they are probably con-
tributors to the high gamma-ray response observed in
the Wichita and in Lower Clear Fork peritidal intervals,
because of their thinness, they cannot be discretely
defined by logs and, therefore, are not correlatable.
Their color and clay content and lack of apparent lat-
eral continuity suggest that they were formed as local
organic-rich ponds or stagnant pools on the tidal flat.
They may act as local baffles to reservoir fluid flow but
are probably very discontinuous laterally.
Exposed Tidal Flat
These deposits show obvious evidence of exposure
such as fenestral pores, pisolites, mud cracks, insect
burrows, sheet cracks, tepee structures, and cyano-
bacterial or microbial laminations (Figure 8). They
are locally common in both the Lower Clear Fork and
the Wichita. In the Lower Clear Fork, they define cycle
tops (Figure 8D); in the Wichita, their eustatic sig-
nificance is less apparent. Their sedimentary structures
FIGURE 8. Core, thin-section, and
borehole image log photographs of
typical tidal-flat facies in the Wichita
and Lower Clear Fork. (A) Slab photo of
exposed tidal-flat facies showing typi-
cal parallel laminations and fenestral
pores. Core is 10 cm (4 in.) wide. Lower
Clear Fork Exxon FCU 7322, depth: 2075 m
(6808 ft). (B) Image log of laminated
tidal-flat facies. Lower Clear Fork, HFS
L2.1. Exxon FCU 2564, depth: 2103–2104 m
(6900–6904 ft). Width of image is 63.5 cm
(25 in.). (C) Photomicrograph of exposed
tidal-flat facies showing fenestral vuggy
porosity. Lower Clear Fork Exxon FCU
6122, depth: 2104 m (6903 ft). (D) Slab
photo of cycle top showing fenestral
exposed tidal-flat facies overlain by
subtidal, burrowed peloid wackestone.
Note small lithoclasts above cycle top.
Lower Clear Fork, top HFS L2.2. Core is
10 cm (4 in.) wide. Exxon FCU 7322,
depth: 2074 m (6804 ft).
demonstrate, however, that they were
formed during at least local sea level
fall and exposure. Porosity is locally
very high (at least 30%) and is as-
sociated with a combination of fe-
nestral pores and intercrystalline
and interparticle pores. Permeability
is lower than in grain-dominated
subtidal rocks but can be significant
where high porosities are present.
However, this study and others (e.g.,
Ruppel, 2002) show that effective permeabilities
(0.1 md) are associated only with porosities above
approximately 12%. Like the peritidal mudstone-
wackestone facies, these rocks typically display some-
what elevated and variable gamma-ray response. This
can sometimes be used to distinguish these rocks
from overlying and underlying subtidal facies.
Peloid Wackestone
This grain-poor facies differs from the rocks of the
peritidal mudstone-wackestone facies in its associa-
tion and sedimentary structures. The rocks are in-
variably associated with other, demonstrably subtidal,
facies; they are burrowed; and they locally contain
skeletal debris (Figure 9). These features suggest that
they were deposited in low-energy subtidal settings.
Dominant grains are 80 –150-mm-diameter peloids
that are probably fecal pellets created by infaunal bur-
rowers. The absence of skeletal allochems supports a
low-energy, perhaps restricted, setting.
364 / Ruppel and Jones
Anhydrite nodules are locally common in these
rocks, probably reflecting the postdepositional entry
of sulfate-bearing diagenetic fluids into permeability
pathways created by burrowers. In many cases, nod-
ules occupy solution-widened vertical burrow path-
ways that are surrounded by alteration halos (Figure 9A).
These halos (which have been reported from many
Permian platform carbonate successions) commonly
display a more grain-rich texture, higher porosity and
permeability, and a depleted oxygen isotope signature
(Major et al., 1990; Ruppel and Bebout, 2001). At Ful-
lerton, they are most commonly developed in high-
frequency sequence (HFS) 2.2. This stratigraphic po-
sition (in the late highstand of the L2 composite
sequence) is analogous in terms of accommodation to
similar features documented from the younger Gray-
FIGURE 9. Core, thin-section, and image
log photo images of nodular peloid
wackestone typical of the Lower Clear
Fork HFS L2.2. (A) Core photo showing
anhydrite nodules surrounded by light-
colored halos of higher porosity. Exxon
FCU 6739, depth: 2130 m (6989 ft). Width
of core is 10 cm (4 in.). (B) Image log
depicting anhydrite burrows and rim-
ming porosity as white (high-resistivity)
patches surrounded by black (low-
resistivity) rims. Exxon FCU 2564, depth:
2097 – 2098 m (6879 – 6883 ft). Width
of image is 63.5 cm (25 in.). (C) Thin-
section photomicrograph of low-porosity,
burrowed, peloid wackestone. Exxon
FCU 6429, depth: 2078 m (6817 ft). (D)
Thin-section photomicrograph of peloid
wackestone with moldic pores. Exxon
FCU 6229, depth: 2085 m (6841 ft).
burg Formation (Ruppel and Bebout,
2001), suggesting that these features
are an expected and perhaps predict-
able feature of late highstand sedi-
mentation on Permian shallow-water
carbonate platforms. Burrowed, nod-
ular fabrics are readily defined on
borehole imaging logs (Figure 9B).
Porosity, except where enhanced
around burrows, is generally moldic
and low (Figure 9C, D).
Peloid Packstone
The distinction between these
rocks and those assigned to the pe-
loid wackestone facies (Figure 10) is
commonly based on apparent pe-
loid abundance and can be subjec-
tive. As with the peloid wackestone facies, peloids are
mostly fecal pellets created by burrowing infauna,
although skeletal debris (chiefly mollusk fragments)
are locally common. It should be noted that the pres-
ervation of these pellets is largely a function of diagen-
esis, both early and late. Nearly all of the low-energy
mud-dominated facies in the Permian contain obvious
pellets, indicating that burrowers were ubiquitous in
these deposits. A complete gradation in texture exists
between pelleted mudstones and peloid packstones.
Differences in texture are probably most commonly
caused by differences in early diagenesis. Pellets that
were early lithified and stabilized are more likely
to be preserved. Thus, peloid (i.e., pellet) packstones
are sediments that underwent significant amounts
of early diagenesis, whereas peloid wackestones and
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 365
pelleted mudstones underwent relatively little. The
textural significance of these pelleted facies is thus
more diagenetic than depositional. Accordingly, ap-
parent variations in peloid abundance in these mud-
dominated facies packstones, wackestones, and mud-
stones are not necessarily an indication of differing
depositional environment or fluctuations in wave
energy. Instead, they may be predominantly caused
by local changes in rates and effects of early diagen-
esis. Reflecting these subtleties, borehole image logs
cannot differentiate between peloid wackestones
and packstones, imaging them all as granular, rela-
tively homogeneous sediment (Figure 10B). Porosity
in these rocks is commonly associated with intercrys-
talline pores and uncommon skeletal moldic pores
(Figure 10C, D).
Peloid Grain-dominated Packstone
Unlike mud-rich peloid facies discussed above,
grain-dominated (or grain-rich) peloid packstones
commonly display evidence of possible wave-related
FIGURE 10. Core, thin-section, and
image log photographs of mud-rich
peloid packstone-wackestone. (A) Slab
photo of burrowed Lower Clear Fork
peloid packstone (dolostone). Exxon
FCU 6229, depth: 2133 m (6997 ft). Core
is 10 cm (4 in.) wide. (B) Image log of
peloid wackestone-packstone facies
typical of Lower Clear Fork HFS L2.2.
Exxon FCU 2564, depth: 2105 – 2106 m
(6905–6909 ft). Width of image is 63.5 cm
(25 in.). (C) Photomicrograph of Lower
Clear Fork peloid packstone (limestone)
with moldic porosity. Exxon FCU 5927,
depth: 2104 m (6903 ft). (D) Photomicro-
graph of Lower Clear Fork peloid pack-
stone (limestone) with moldic porosity.
Exxon FCU 6229, depth: 2078 m (6817 ft).
transport. These rocks, which con-
tain pellets (90–120 mm), less com-
mon ooids (150–250 mm), and un-
identifiable spherical grains, are
typically well sorted and contain in-
terparticle pores that are either open
or filled with cements (Figure 11).
The interparticle pores indicate
that these peloids acted as true grains
instead of pelleted mud, in contrast
to the intercrystalline and moldic
pores that typify mud-dominated
facies. As discussed above, in some
instances, the grain-dominated tex-
ture is associated with vertical burrows and probably
owes its origin to burrow-related diagenesis. For the
most part, this facies is restricted to the subtidal legs of
Lower Clear Fork depositional sequences. These rocks
are among the highest quality reservoir facies in the
field. Highest porosity (as much as 20%) and per-
meability are encountered in peloid grain-dominated
packstone limestones (Figure 11C), although dolo-
stones also exhibit good porosity and permeability
(Figure 11B, D). Where burrowed, porosity and per-
meability can be greatly enhanced in and around
burrows (Figure 12A). These high-porosity burrow fills
commonly account for apparent moldic fabric on
image logs (Figure 12B).
Ooid-peloid Grain-dominated
Packstone-grainstone
These rocks differ from the peloid grain-dominated
packstone facies, with which they are commonly close-
ly associated, in having recognizable ooids (as much
as 250–300 mm in diameter) in addition to pervasive
366 / Ruppel and Jones
pellets. Skeletal grains in the form of fusulinids, mol-
lusks, and crinoids are also common. In most cases,
these deposits, which are restricted to subtidal sections
of the Lower Clear Fork, are also well sorted, probably
because of wave action. Grainstones, although relatively
uncommon, display excellent size sorting and, in some
cases, possess inclined or cross-laminations (Figure 13B).
These rocks, which occur as both dolostone and lime-
stone, represent the highest energy facies in the reservoir
succession and, in many cases, display the best porosity
and permeability. Pores are dominantly interparticle,
although moldic pores are also very abundant, especial-
ly in limestone-dominated intervals (Figure 13D).
FIGURE 11. Core and thin-section
photomicrographs of peloid, grain-rich
packstones. (A) Slab photo of Abo
peloid grain-dominated packstone.
University Consolidated IV-25, depth:
2210 m (7252 ft). Core is 10 cm (4 in.)
wide. (B) Photomicrograph of Lower
Clear Fork peloid dolostone grain-
dominated packstone showing inter-
particle porosity. White areas are poi-
kilotopic anhydrite. Exxon FCU 6122,
depth: 2126 m (6976 ft). (C) Photomi-
crograph of Lower Clear Fork peloid
(ooid?) dolostone grain-dominated
packstone-grainstone with interparticle
porosity. Exxon FCU 6946, depth: 2083 m
(6835 ft). (D) Photomicrograph of
Lower Clear Fork peloid grain-dominated
packstone dolostone showing inter-
particle porosity. Exxon FCU 6229, depth:
2078 m (6816 ft). (E) Slab photo of
Lower Clear Fork (L 2.2) peloid grain-
dominated packstone (dolostone). Exxon
FCU 6229, depth: 2079 m (6822 ft).
Width of photo is 10 cm (4 in.).
Fusulinid Wackestone-packstone
Fusulinid-bearing rocks (Figure 14)
are found in all three formations
(Abo, Wichita, and Lower Clear Fork),
although they are very limited in the
Wichita. In all occurrences, they are
most commonly dolomitized. Fusu-
linids, which constitute as much as
40% of the rock are preserved either
as open (Figure 14C) or anhydrite-
filled (Figure 14A) molds or as well-
preserved fossil tests (Figure 14D).
In the Lower Clear Fork and Wichita,
they are associated with abundant
peloids (probably pellets). In the Abo,
they commonly co-occur with cri-
noids and less commonly with brachiopods. Fusu-
linids are thought to have occupied water depths of
30 m (100 ft) or more and here represent the deepest
water facies observed at Fullerton field. Accordingly,
their presence is an indicator of platform flooding
and relative sea level rise, making them key indicator
facies of cycle and sequence boundaries.
Skeletal Wackestone-packstone
Typically, these rocks contain small volumes of skel-
etal debris (most commonly mollusk fragments, but
also including crinoids and less common ostracods)
and the ubiquitous peloids. They are gradational into
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 367
peloid wackestones and packstones. Evidence of bur-
rowing is common. The dominance of mollusks and
the essential absence of more normal-marine organ-
isms in these rocks suggest that they were deposited
in an inner-platform setting. Typically, they exhibit
low porosity; pore space is created by skeletal molds
and intercrystalline pores.
Oncoid Wackestone-packstone
Oncoids (or oncolites) are large microbially coated
grains (Figure 15) formed under conditions of fre-
quent wave agitation in shallow water. They are abun-
dant at the base of the Lower Clear Fork throughout
the entire field area. Invariably, they are associated
with fusulinids and other faunas indicative of open-
marine deposition. This association and their strati-
graphic position immediately above the top of the tidal-
flat–dominated Wichita indicate that they represent
marine flooding of the platform during sea level rise
(transgression). In some downdip wells (e.g., Pan
American FM-1; Figure 3), they are common at Lower
Clear Fork cycle bases. Their distribution suggests that
they document relatively high-energy conditions
developed during platform flooding. Like the fusuli-
nid facies, the oncoid wackestone-packstone facies is
FIGURE 12. Core photo (A) and image
log (B) images of burrowed peloid
packstone. Although low-resistivity
(black) features are commonly inter-
preted to be open vugs on image logs,
core photo reveals that vugs are ac-
tually burrow fills (light colored) of
higher porosity. Lower Clear Fork. Core:
Exxon FCU 6739, depth: 2123 m (6966 ft).
Image log: Exxon FCU 2564, depth:
2095 – 2096 m (6873 – 6877 ft). Core is
10 cm (4 in.) wide. Image log is 63.5 cm
(25 in.) wide.
a marker facies indicative of trans-
gression or platform deepening (i.e.,
sea level rise).
Siltstone-sandstone
Quartz silt- and sand-bearing rocks
are common only at the top of the
Lower Clear Fork and in the overly-
ing Tubb Formation. Because the
quartz is associated with potassium-
and thorium-rich clay minerals, its
presence in the Lower Clear Fork and
Tubb is well defined by high cor-
rected gamma-ray (CGR) wire-line-
log response (Figures 4, 5). Their maturity, size (fine
sand to coarse silt), sorting, and shape (generally sub-
angular) suggest that these sediments were origi-
nally wind blown (Fischer and Sarntheirn, 1988).
This indicates that most of the quartz in the section
was delivered to the area during sea level lowstands
when the platform was emergent. Quartz silt-sand
occurs in two scenarios: (1) in peritidal tidal-flat fa-
cies and (2) in reworked subtidal facies. The former,
which are probably formed as small volumes of silt,
sand, and clay, are blown onto intermittently exposed,
slowly accumulating tidal flats and admixed with peri-
tidal carbonate sediment. The presence of potassium
and thorium in these clastics produces the increased
gamma-ray response that is associated with many
tidal-flat deposits in the Clear Fork and facilitates rec-
ognition of these cycle-capping deposits in these sedi-
ments (Ruppel, 2002). Most of the occurrences of silt,
sand, and clay in the Lower Clear Fork at Fullerton
are of this type.
Subtidal silt, sand, and clay deposits are most com-
mon in the Tubb. These rocks were formed first by
large-volume eolian deposition during extended sea
level lowstand and carbonate nondeposition, then re-
worked during the ensuing sea level rise and marine
368 / Ruppel and Jones
flooding of the platform. These deposits are generally
richer in clastic content, are intermixed with carbonate
mud, and show evidence of subtidal conditions (e.g.,
burrows, stratification). Although these rocks may
locally exhibit some minor porosity, they do not ap-
pear to display any significant reservoir permeability.
Lithoclast Wackestone
Thin intervals (commonly less than 0.3 m [1 ft])
containing scattered lithoclasts are locally encoun-
tered at the contacts between transgressive marine
facies and underlying tidal-flat deposits or other facies
showing evidence of subaerial exposure (Figure 8D).
Clasts are variable in composition but most com-
monly consist of fragments or intraclasts derived from
the underlying bed. Clasts typically are 1 cm (0.4 in.)
in width or less. Although volumetrically minor, these
rocks are an important indicator facies of lithification
because of exposure or nondeposition and subsequent
sea level rise and, thus, of cycle boundaries.
FIGURE 13. Core, thin-section, and
image log photographs of Lower Clear
Fork and Abo grainstones. (A) Slab
photo of Lower Clear Fork ooid grain-
stone showing cross-laminations. Exxon
FCU 6122, depth: 2109 m (6920 ft).
Core is 10 cm (4 in.) wide. (B) Image
log of cross-bedded grainstone. Lower
Clear Fork. Exxon FCU 2564, depth:
2134–2135 m (7002–7005 ft). Width of
image is 63.5 cm (25 in.). (C) Photo-
micrograph of Lower Clear Fork ooid
limestone-grainstone showing oomoldic
pores. Exxon FCU 5927, depth: 2107 m
(6913 ft). (D) Photomicrograph of
Lower Clear Fork ooid dolostone grain-
stone wackestone showing interparti-
cle and minor moldic pores. Exxon
FCU 4828, depth: 2175 m (7137 ft).
Depositional Model
The relative distribution of the
facies described above can best be
understood when considered in light
of a conceptual geological model.
Figure 16 portrays idealized 3-D re-
lationships among major facies types
and depositional environments typ-
ical of most middle Permian (Leo-
nardian and Guadalupian) platform
carbonate successions in the Perm-
ian basin. It should be understood
that this model reflects only the
relative interrelationships among facies tracts. The
actual position of individual facies tracts at a given
point in time is a function of many factors, including
accommodation, rates and magnitude of sea level
rise, rising versus falling sea level trends, climate, tec-
tonics, etc. However, during normal relative rises in
sea level, facies tracts normally step landward,
whereas during falls, they step basinward. Exceptions
to this general pattern are common. For example,
ramp-crest facies may be much better developed dur-
ing sea level highstand and fall than during trans-
gression and rise. The basal Leonardian (L1) deposi-
tional sequence at Fullerton, for example, appears
to be dominated by an updip, inner-ramp succes-
sion (Wichita) and a downdip outer-platform succes-
sion (Abo) with a very poorly developed middle-ramp
to ramp-crest facies tract. Nevertheless, the vertical
successions (facies-stacking patterns) formed by sea
level-driven migrations of these facies tracts are the
key to identifying depositional cycles in cycles from
cores and logs.
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 369
Among the identified facies at Fullerton, the peri-
tidal mudstone-wackestone facies, the clay-rich carbon-
ate mudstone facies, and the exposed tidal-flat facies,
for the most part, represent deposition on the inner
ramp in a sabkha-tidal-flat setting as low-exposure
index (cf. Hardie and Garrett, 1977) peritidal deposits,
tidal-flat ponds, and high-exposure tidal-flat deposits,
respectively (Figure 16). It should be noted, however,
that exposure fabrics, like those developed in the inner
platform, can also form on the ramp crest following
complete aggradation and exposure at any time.
The peloid- and pellet-rich facies at Fullerton (peloid
wackestone facies, peloid packstone facies, peloid grain-
dominated packstone facies, and skeletal wackestone-
packstone facies) are also dominantly associated with
the inner ramp but occupy a somewhat more distal,
lagoonal to restricted subtidal setting, where sedi-
ment formation is dominated by infaunal and epi-
faunal burrowing activities (Figure 16).
FIGURE 14. Core, thin-section, and
image log photographs of the fusulinid
wackestone-packstone facies. (A) Slab
photo of fusulinid wackestone with open
and anhydrite-filled fusumoldic pores.
Exxon FCU 7322, depth: 2132 m (6996 ft).
Core is 10 cm (4 in.) wide. (B) Image
log of fusulinid wackestone-packstone
facies. Exxon FCU 2564, depth: 2131 –
2132 m (6990–6994 ft). Width of image
is 63.5 cm (25 in.). (C) Photomicrograph
of fusulinid wackestone showing open
fusumolds. Exxon FCU 7630, depth:
2066 m (6778 ft). (D) Photomicrograph
of fusulinid wackestone showing well-
preserved fusulinids but little porosity.
Exxon FCU 4828, depth: 2139 m (7018 ft).
The ooid-peloid grain-dominated
packstone-grainstone facies in the
Lower Clear Fork at Fullerton is typi-
cal of facies deposited in a platform
ramp crest at the platform margin
where relatively high wave energies
are developed (Figure 16). However,
none of the cored wells in the field
reveal the kind of vertical facies stack-
ing and amalgamation that typically
defines a well-developed ramp crest
(see, for example, Kerans et al., 1994;
Ruppel et al., 2000). This may reflect
insufficient core control or the ab-
sence of a well-developed ramp crest
on the Lower Clear Fork platform. As
previously stated, the Abo also appears to lack a ramp-
crest facies succession.
As in most middle Permian carbonate-platform
successions, the fusulinid wackestone-packstone fa-
cies documents outer-platform deposition. Such rocks
are well developed in both the Lower Clear Fork and
the Abo. In the Abo, these rocks display classic clino-
formal bedding and the general lack of cyclicity that
typifies a distal outer-platform setting (Figure 16). By
contrast, fusulinid wackestone-packstone facies rocks
in the Lower Clear Fork are horizontally bedded and
interbedded with middle- and inner-platform facies.
This suggests that these fusulinid-rich rocks were de-
posited in a much more proximal outer-ramp position.
The oncoid wackestone-packstone facies is closely
associated with the fusulinid wackestone-packstone
facies in the Lower Clear Fork at Fullerton, indicating
that the former was deposited during platform deep-
ening. This facies has not previously been reported
370 / Ruppel and Jones
from middle Permian rocks in the Permian basin. Its
occurrence at Fullerton field may indicate the de-
velopment of relatively higher energy conditions on
the outer platform in the Fullerton area than is gener-
ally developed in other regions in the Permian basin.
FIGURE 16. Depositional model for Permian shallow-water carbonate platforms in the Permian basin. This model is
applicable to most Leonardian and Guadalupian carbonate platform successions, including the reservoir succession at
Fullerton field. Modified from Kerans and Ruppel (1994).
FIGURE 15. Core photographs of
oncoid, wackestone-packstone facies.
(A) Small, algally coated fusulinids. Low-
er Clear Fork, Exxon FCU 5927, depth:
2128 m (6980 ft). (B) Large oncoids.
Lower Clear Fork, Pan American FM-1,
depth: 2182 m (7160 ft). Cores are 10 cm
(4 in.) wide.
Sequence Stratigraphy
Figure 17 illustrates diagrammat-
ically the sequence architecture,
basic facies tracts interrelationships,
and rock and time terminology of
the lower Leonardian at Fullerton
field. In this section, we describe the
sequence-stratigraphic characteris-
tics of each of the major Leonardian
units at Fullerton.
Abo Formation
As already discussed, the Abo rep-
resents the distal facies or systems
tract of the earliest (oldest) Leonar-
dian sequence in the Permian basin: the L1 sequence
(Figures 4, 5, 17). Regionally, the Abo consists of outer
ramp to slope, skeletal crinoid-fusulinid-dominated
subtidal facies and, less commonly, ramp-crest ooid-
peloidal grainstones (Kerans et al., 2000). In both
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 371

FIGURE 17. Generalized sequence-stratigraphic model of the lower Leonardian succession at Fullerton field showing
sequence boundaries, primary facies tracts, and stratigraphic nomenclature.

outcrop and the subsurface, the Abo is dominantly
characterized by clinoformal, top-lapping geometries
(Kerans et al., 2000; Zeng and Kerans, 2003). This
typical Abo architecture is very apparent in 3-D and
2-D seismic data from Fullerton field, where the Abo
is readily defined by its clinoformal reflectors, and its
top is defined by a prominent toplap surface that is
apparent throughout much of the field (Figure 6).
Although the Abo at Fullerton field is penetrated
by relatively few wells and has been cored in even
fewer wells, existing cores display facies and bedding
characteristics typical of sediments deposited in an
outer ramp-slope setting (such as open-marine fauna,
inclined bedding, graded beds, slump features, con-
glomerates, and lack of shallow-water sediments).
Facies consists of alternating beds of fusulinid-crinoid
packstones and wackestones and peloidal pack-
stones. No cyclicity is apparent in the Abo, although
alternations between skeletal-rich and skeletal-poor
intervals are evident. Inclined beds are locally com-
mon. Additionally, correlations in the Abo are not
apparent either from core sections or from wire-line
logs. In part, this is the result of limited core and well-
log control through the Abo section. But poor cor-
relatability is typical of outer-platform depositional
successions that display clinoformal architectures
because of the associated dipping bedding surfaces,
discontinuous facies packages, and poor vertical facies
contrasts.
Abo rocks are dominantly dolostones at Fullerton
field. Porosity is locally very high, reaching values of
as much as 25%. Pore types vary among fusumolds,
intercrystalline pores in coarse-crystalline dolostones,
and interparticle pores in skeletal-peloidal packstones.
The thickness of the Abo is indeterminate largely
for two reasons. First, few wells and no cores pen-
etrate the complete Abo to Wolfcamp section. Second,
outcrop studies reveal that the Abo – Wolfcamp con-
tact may be lithologically indistinct in many areas:
in outcrop, both the uppermost Wolfcamp and the
Abo are composed of clinoformal, outer-platform
fusulinid-crinoid wackestones (Kerans et al., 2000).
372 / Ruppel and Jones
A minimum thickness of 90 m (300 ft) is established
by the Pan American FM-1 cored well in the south-
eastern corner of the field (Figures 3, 5).
The contact between the Abo and the overlying
Wichita varies from sharp to gradational. In some
cores, the contact is marked by an interval (as much as
several feet thick) of intermixed clasts and fragments
of Wichita tidal-flat deposits and Abo fusulinid-
bearing subtidal deposits. The brecciated nature of
this contact interval suggests that it is the result of
karst-related dissolution and collapse. These breccias
and their distribution are discussed more fully in a
later section of this chapter.
Wichita Formation
The Wichita at Fullerton field consists of a diverse
assemblage of peritidal to supratidal tidal-flat depos-
its. Integration of outcrop and subsurface data sug-
gests that the Wichita Formation is the updip, proxi-
mal facies equivalent of both the Abo and the Lower
Clear Fork subtidal successions. These data also imply
that the Wichita actually comprises parts of two depo-
sitional sequences: the highstand leg of sequence L1
and the transgressive leg of sequence L2 (Figure 17).
The lower Wichita represents the updip, tidal-flat fa-
cies tract equivalent of the downdip, outer-platform
facies tract of the Abo in Leonardian sequence L1,
whereas the upper Wichita represents the updip tidal-
flat facies equivalent of the basal Lower Clear Fork
subtidal facies in sequence L2 (HFS 2.0; Figures 5, 17).
The Wichita is dominantly composed of dolostone;
however, intervals of limestone are common in the
upper Wichita in the northern part of the field. The
limestone intervals are relatively persistent laterally
and are generally subparallel to stratigraphic markers
in the field (Figure 18a).
The thickness of the Wichita is relatively consis-
tent throughout the northern two thirds of the field,
ranging from about 75 to 105 m (280 to 350 ft), but
thickness decreases markedly to as little as 34 m (110 ft)
in the southeastern part of the field (Figure 19). This
decrease in thickness, which occurs relatively abruptly
along a generally northeast-trending belt, marks the
platform-margin boundary between updip tidal-flat
Wichita facies and downdip outer-ramp Abo fusulinid
facies (Figure 5).
The position of the L1-L2 boundary is difficult to
precisely place within the thick succession of Wichita
peritidal deposits. The contact shown in Figure 5 is
extrapolated from downdip core and well control,
where the L1-L2 contact is well defined by a sharp,
karsted contact between the outer-ramp subtidal
deposits of the Abo and peritidal tidal-flat rocks of
the overlying Wichita.
Lower Clear Fork Formation
Rocks assigned to the Lower Clear Fork Formation
represent the subtidal systems tract of the Leonardian
2 (L2) sequence. Core studies show that the Lower
Clear Fork can be subdivided into three HFSs through-
out most of the Fullerton field area: L2.1, L2.2, L2.3
(Figures 4, 5). Each of these HFSs can be identified as
having a transgressive, dominantly subtidal lower leg
and an overlying, highstand upper leg.
High-frequency Sequence L2.0
High-frequency sequence L2.0 documents the
initial flooding of the platform following sea level
fall and exposure at the end of L1 deposition. In most
of the field area, HFS 2.0 consists of amalgamated
tidal-flat deposits of the upper Wichita. Subtidal Lower
Clear Fork deposits of L2.0 are only present at the
margins of the field (Figures 4, 5).
High-frequency Sequence L2.1
In most of the Fullerton field area, HFS L2.1 forms
the base of the Lower Clear Fork. As defined for this
study, L2.1 consists of a basal section of transgressive
to early highstand subtidal platform facies and an
upper section of highstand tidal-flat facies (Figures 4,
5, 17). The basal transgressive subtidal facies of the
Lower Clear Fork represent the first marine flooding
of the platform and a sharp change in depositional
style from the tidal-flat deposition of the Wichita to
subtidal deposition of the Lower Clear Fork.
The basal transgressive leg of L2.1 consists domi-
nantly of fusulinid wackestone-packstone and oncoid
wackestone-packstone facies that document landward
backstepping of outer platform facies across the plat-
form. These rocks are generally overlain by a succes-
sion of peloid packstones and grain-dominated pack-
stones that represent late transgression and early
highstand. Locally, within this succession, there are
exposed tidal-flat facies indicating periodic exposure.
L2.1 is capped throughout most of the field by a suc-
cession of one to three tidal-flat-capped cycles that rep-
resent exposure during late highstand (Figures 5, 17).
Tidal-flat facies in some cases exhibit an elevated
gamma-ray-log response that aids in their recogni-
tion (Figures 4, 5).
Like most of the Lower Clear Fork in the Permian
basin, HFS 2.1 is dominantly dolostone. However,
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 373

FIGURE 18. Distribution of limestone and dolostone in Wichita and Lower Clear Fork based on cores and wire-line logs.
(a) Cross section (BB0 ) across the northern part of the field showing cycle stratigraphy and mineralogy. (b) Maps
showing limestone in Wichita HFS L2.0, Lower Clear Fork HFS L2.1, and Lower Clear Fork HFS L2.2.

limestone is locally common at Fullerton field, es-
pecially in the transgressive leg of 2.1. Areally, lime-
stone is most abundant in an arcuate belt through
the field (Figure 18b).
The thickness of L2.1 is relatively constant across
the Fullerton field area, ranging from about 43 to 46 m
(140 to 150 ft) across most of the area. Good evidence
shows that local thickness changes are a function of
differential subsidence along deep-seated faults (Jones
and Ruppel, 2004). This is especially apparent along
the north-trending fault in the northeastern part of
the field, where the thickness changes from less than
43 m (140 ft) on the western, upthrown side of the
fault to more than 49 m (160 ft) on the eastern, down-
thrown side of the fault. Examination of modern struc-
ture (Figure 3) shows that considerable movement
374 / Ruppel and Jones
FIGURE 19. Thickness of the Wichita Formation in the
Fullerton field area. Marked thinning in southeast part
of the field corresponds to facies change from Wichita
to Abo.
on this fault occurred even after Lower Clear Fork
deposition.
Porosity is most abundant in the subtidal (trans-
gressive and early highstand) parts of the sequence.
Pore space is dominated by intercrystalline and moldic
pores, the latter being especially abundant in lime-
stone sections. Overall, these rocks are most porous
in areas where limestone is present. Tidal-flat rocks
are also locally porous but generally contain low
permeabilities. The subtidal rocks of L2.1 probably
constitute the most productive reservoir interval in
the field.
High-frequency Sequence L2.2
High-frequency sequence L2.2 is similar to HFS
L2.1 in consisting of a basal transgressive leg com-
posed of backstepping tidal-flat facies, a middle (late
transgressive to early highstand) leg composed domi-
nantly of subtidal facies, and an uppermost (late high-
stand) leg composed of tidal-flat facies (Figures 5, 17).
As is the case with L2.1, there is good evidence that the
marine flooding of the platform following the post-
L2.1 lowstand was progressive. Basal L2.2 tidal-flat
deposits are thickest in the center of the field area but
generally absent along the margins, reflecting greater
accommodation and early flooding of downdip areas.
The lower abundance of outer-ramp fusulinid-rich
facies in the TST of L2.2 relative to L2.1 suggests that
the overall accommodation during the L2.2 sea level
rise was somewhat less than during L2.1. Lower ac-
commodation and attendant lower wave energies are
also suggested by the near absence of the oncoid
wackestone-packstone facies in L2.2. These rocks are
found only in the most downdip core in the field. The
fusulinid wackestone-packstone facies is also largely
restricted to the eastern and southeastern parts of the
area (Figures 5, 17). Throughout most of the field
area, L2.2 is dominated by peloid wackestones and
packstones typical of middle-platform deposition.
The sequence is capped by a thin succession of tidal-
flat cycles (Figures 5, 17).
High-frequency sequence L2.2 is composed dom-
inantly of dolostone, with two important exceptions.
Like L2.1, limestone is dominant in the southern part
of the field (Figure 18b). Dolomite in this area is
largely restricted to mud-rich fusulinid wackestones;
virtually all grain-rich facies in this area are lime-
stone. Limestone is also abundant in a small area in
the north-central part of the field.
The thickness of L2.2 is generally significantly less
than that of L2.1. L2.2 ranges from about 26 to 27 m
(85 to 90 ft) across most of the central part of the area
to about 100 ft (30 m) along the northern and
southern field margins. Porosity development in L2.2
is similar to that in L2.1. Porosity is greatest in the
subtidal (transgressive and early highstand) parts of
the sequence and dominated by intercrystalline and
moldic pores. In addition, like L2.1, porosity is great-
est in sections containing significant limestone.
High-frequency Sequence L2.3
The uppermost Lower Clear Fork HFS (L2.3) is com-
posed of tidal-flat-capped restricted subtidal cycles
throughout most of the field area (Figures 5, 17).
These tidal-flat caps typically contain fenestral and
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 375
fine intercrystalline pore space that is definable on
porosity logs. Log correlations suggest that these cycle-
capping tidal-flat facies are relatively continuous across
significant areas of the field. The underlying cycle-
base subtidal rocks are dominantly mud-rich pack-
stones and wackestones. Porosity, as indicated, is
mostly restricted to tidal-flat caps. Because these rocks
are dominated by fine intercrystalline and moldic
pores, they contain little, if any, reservoir permeabil-
ity. Accordingly, locally they contain oil stain and
rarely contribute to oil production.
Tubb Formation
The Tubb is characterized by fine-grained silic-
iclastics (coarse siltstone and fine sandstone), which
are relatively easily defined by high-gamma-ray-log
response (Figures 4, 5). These clastics are interpreted
to represent eolian deposits that were deposited during
the post-L2 lowstand and then reworked during the
ensuing L3 sea level rise (Kerans et al., 2000; Ruppel
et al., 2000; Ruppel, 2002). Most intervals are varying
mixtures of siltstone-sandstone and carbonate (typ-
ically mud-rich, shallow-water facies). The Tubb
locally contributes to oil production in the Permian
Basin but is not part of the reservoir at Fullerton field.
Cycle-scale Stratigraphy
The fundamental goal of cycle stratigraphy is to
develop a correlation framework based on time-
equivalent surfaces. The basic underlying premise
for this approach is the assumption that widely
correlative depositional cycles are formed by punc-
tuated, allocyclic processes (e.g., sea level rise and
fall) that affected sedimentation over broad areas.
The methodology used to develop a cycle-scale strati-
graphic framework in the Leonardian section at Ful-
lerton field consists of the following: (1) characteriza-
tion of facies-stacking patterns and cycle development
in analogous outcrops; (2) description, interpretation,
and logging of facies, stacking patterns, and possible
cycle tops in cores; (3) integrated log- and core-based
correlation of tentative cycle tops; and (4) definition
of cycle architecture. At Fullerton field, nearly 4570 m
(15,000 ft) of core (from 29 cored wells) was described
in detail to provide the basic data for cycle definition.
Wichita Cyclicity
Correlations in the thick Wichita succession of
relatively similar tidal-flat deposits are difficult be-
cause of (1) the discontinuous nature of tidal-flat facies,
(2) the lack of systematic vertical stacking relation-
ships among facies to define cycles, (3) the nonuni-
form response of wire-line logs to these facies, and
(4) the overprinting effects of diagenesis. Internal,
cycle-scale correlations are possible, however, in the
upper Wichita in the northern part of the field, where
limestone beds are relatively correlative. Sedimentary
structures suggest that these limestones represent un-
dolomitized bases of cycles whose tops have been do-
lomitized by early diagenesis. The Wichita cycles de-
fined by these limestone-dolostone couplets average
6 m (20 ft) in thickness.
Lower Clear Fork Cyclicity: HFS L2.1
Facies-stacking patterns in HFS 2.1 define numer-
ous apparent cycle tops at thicknesses ranging from
0.6 to 4.6 m (2 to 15 ft). Cycles are typically char-
acterized by mud-rich facies at their bases and grain-
rich facies at their tops. Basal TST cycles in HFS L2.1
contain abundant oncoids at cycle bases, along with
accompanying fusulinids (Figure 20). These cycles
are capped by better sorted, peloid-rich facies. High-
frequency cycles (typically 1.5–3 m [5–10 ft] in thick-
ness) stack into cycle sets that average 7.5 – 12 m (25 –
40 ft) in thickness (Figure 20). Cycle sets display sim-
ilar facies-stacking patterns to cycles consisting of
fusulinid-rich bases and peloid-rich caps. Locally, tidal-
flat facies cap these cycle sets. Porosity is generally
FIGURE 20. Facies stacking and cycle development in
the fusulinid- and oncoid-rich, transgressive systems
tract of Lower Clear Fork HFS L2.1. Note that porosity
is typically developed at or near cycle tops. Note also
the lack of any systematic relationship between gamma-
ray log and facies or cyclicity.
376 / Ruppel and Jones
FIGURE 21. Facies stacking and cycle development in
the grain-rich, highstand systems tract of Lower Clear
Fork HFS L2.1. Note that porosity is typically developed
at or near cycle tops. As with other Lower Clear Fork
successions, gamma-ray logs do not show a systematic
response to facies or cyclicity.
highest at both cycle tops and cycle-set tops relative
to bases. However, because cycles in the lower parts of
cycle sets are generally more mud and fusulinid rich,
overall porosity in these basal TST cycles is commonly
relatively low.
Cores and outcrop studies suggest that Clear Fork
platform cycles and cycle sets are correlative across sig-
nificant distances. However, where cores are not avail-
able, these correlations can be difficult to establish.
Gamma-ray logs display virtually no systematic re-
sponse to subtidal facies and cycles (Figures 20, 21)
and, thus, cannot be reliably used for cycle-scale cor-
relation throughout most of the Lower Clear Fork. In
the absence of gamma-ray logs, the most effective
way of establishing cycle-scale correlations is through
the use of porosity logs. This approach is based on the
observation from outcrops and cores that cycle tops
consistently contain the most grain-rich facies, and
that these rocks are most likely to contain high po-
rosity. Accordingly, porosity logs can be used to cor-
relate both the high-porosity facies in the upper part
of the cycle and the overlying cycle top.
L2.1 highstand cycles are dominated by mud-rich
peloidal facies at their bases and grain-rich peloid-
or ooid-bearing facies at their tops (Figure 21).
Fusulinids are commonly sparse in these highstand
cycles, reflecting the basinward shift in facies tracts.
In general, these cycles are dominated by peloid
packstones and grain-rich packstones. Porosity is
commonly highest at both cycle tops and cycle-set
tops because of the abundance of these grain-rich
facies. Cycles typically average 1.5 – 3 m (5 – 10 ft) in
thickness; cycle sets are commonly 6 – 9 m (20 – 30 ft)
thick (Figure 21).
Lower Clear Fork Cyclicity: HFS L 2.2
Facies stacking and cycle development in HFS
L2.2 are very similar to those in L2.1. Cycles average
1.5– 3 m (5 – 10 ft) in thickness, and porosity is best
developed in cycle tops (Figure 22). L2.1 cycles dif-
fer, however, in the lack of oncoid facies and the
relative scarcity of fusulinid facies. This presumably
reflects decreasing overall accommodation in L2.2
because of platform aggradation and slowing rates
of long-term sea level rise. Additionally, cycle sets
are not generally definable in L2.2. This is largely
caused by the general absence of fusulinid facies that,
in L2.1, define these intermediate-scale sea level-rise
events.
Lower Clear Fork Cyclicity: HFS L2.3
High-frequency cyclicity in the Leonardian at Ful-
lerton field is most readily definable in L2.3. These
rocks, which are characterized by tidal-flat-capped
shallow subtidal cycles (Figure 23), appear to be much
more widely correlative than cycles in L2.1 or L2.2.
Although neither facies nor cyclicity is defined by
gamma-ray logs, both are distinguishable on porosity
logs because of the typically well-developed porosity
associated with cycle-capping tidal-flat deposits. The
porosity in these rocks is generally caused by the
presence of fenestral pores and can be relatively high,
especially relative to the generally low porosity ex-
hibited by cycle-base mud-rich wackestones and
FIGURE 22. Facies stacking and cycle development in
the grain-rich, late transgressive systems tract and early
highstand systems tract of Lower Clear Fork HFS L2.2.
Note that porosity is typically developed at or near cycle
tops. Note here again that gamma-ray logs do not display
any relationship to facies or cycle development.
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 377
FIGURE 23. Facies stacking and cycle development in
the nonreservoir Lower Clear Fork HFS L2.3. Porosity is
typically developed in cycle-capping tidal-flat facies, but
little or no permeability is associated with the fenestral
pores that dominate these caps.
packstones. However, because of the separate-vug
fenestral pores, permeability is generally low, and
L2.3 sparsely contributes to hydrocarbon production
in the field. Porosity-based wire-line correlations sug-
gest that cycle and facies continuity is high. This is
somewhat unexpected, considering outcrop observa-
tions that suggest that tidal-flat facies are highly dis-
continuous. The laterally continuous porosity develop-
ment at these cycle tops probably reflects the effects
of early diagenesis associated with sea level fall and
subsequent rise at these surfaces.
MINERALOGY AND DIAGENESIS
The reservoir section at Fullerton field is similar to
most other platform Leonardian successions in the
Permian basin in showing evidence of significant
postdepositional diagenesis. Principal products of this
diagenesis are matrix-replacive and pore-filling dolo-
mite and anhydrite. However, limestone is locally pres-
ent in the Leonardian section at Fullerton, including
the Abo, the Wichita, and the Lower Clear Fork.
Dolomite and Limestone Distribution
Dolomite is, by far, the dominant mineral in the
reservoir section. The Abo consists entirely of dolo-
mite except in the most downdip wells. In the FM-1
well, perhaps the most downdip well in the field, the
Abo contains alternating zones of limestone and
dolostone. Dolostone is more commonly associated
with mud-rich facies (peloid wackestone and fusuli-
nid wackestone) in the Abo clinoformal succession,
whereas limestone intervals are more commonly
grain-rich, skeletal facies.
Limestone is locally common in the Wichita, espe-
cially in the upper half of the formation in the north-
ern part of the field (Figure 18). In all cases, calcite-
rich rocks in the Wichita are peritidal mudstones or
wackestones (the peritidal mudstone-wackestone
facies). These calcite-dominated facies characteristi-
cally exhibit very low porosity (<2%) and essentially
no permeability. Log correlations suggest that lime-
stone intervals are locally correlative (Figure 18a).
Facies-stacking patterns suggest a systematic relation-
ship between mineralogy and cyclicity, and it seems
probable that these limestones are the result of cycle-
punctuated diagenesis: limestones representing un-
dolomitized cycle bases. Limestone is virtually absent
from the lower Wichita (L1 highstand deposits) but very
common in the upper Wichita (L2 TST) (Figure 18a).
Limestone abundance in the upper (L2) Wichita dis-
plays a systematic trend across the field. The highest
abundance of limestone is in the northwestern part
of the field; essentially, no limestone is present in the
southern part of the field (Figure 18b). Because of the
low porosity associated with limestones in the Wichita,
porosity logs can be reliably used to define and cor-
relate limestone even where cores, dual porosity, or
photoelectric effect (PE) logs are unavailable.
Lower Clear Fork rocks at Fullerton field also con-
tain locally abundant limestone (Figure 18a). In HFS
L2.1, limestone is especially common in a sinuous
belt that generally follows the margin of the outer
platform (Figure 18b). Essentially, all grain-rich facies
in this belt are dominantly calcitic.
Limestone also dominates L2.2 in the southern
part of the field. However, outside of this area, lime-
stone is uncommon, except in a small area in the
northwestern part of the field (Figure 18b). Porosity
is generally high in limestone-rich intervals in the
subtidal Lower Clear Fork. However, porosity is also
well developed in dolostones, so except where cores,
dual-porosity log suites, or PE logs are available, it is
not possible to differentiate limestone from dolo-
stone in the Lower Clear Fork.
Stable Isotope Chemistry
Stable isotopes can provide insights into the
timing and spatial distribution of diagenetic events.
Because diagenesis is a key factor in porosity preser-
vation and loss, such data can commonly provide in-
sights into causes and distribution of porosity trends.
Stable isotope data collected during this study and
those previously gathered by J. Kaufman (1991, per-
sonal communication) reveal similarities to other
Leonardian successions in the Permian basin but also
appear to define spatially and temporally distinct
patterns of diagenesis (Figure 24).
378 / Ruppel and Jones
FIGURE 24. Stable isotope data for the Tubb, Wichita,
Lower Clear Fork, and Abo at Fullerton field. Solid red
circles and blue triangles are dolomite and calcite samples,
respectively, from J. Kaufman (1991, personal communi-
cation). Open green circles and green triangles are do-
lomite and calcite samples, respectively, collected during
this study.
Dolostone samples from the Tubb, Lower Clear
Fork, Wichita, and Abo at Fullerton field exhibit d13C
values ranging from 1.41 to 5.54% Peedee belemnite
(PDB) (n = 50). These values are generally lighter than
data reported from the Guadalupian of the Permian
basin (Ruppel and Cander, 1988a, b; Leary and Vogt,
1990; Saller and Henderson, 1998) but similar to
most Leonardian values (Saller and Henderson, 1998;
Ruppel, 2002).
The d18O data obtained from Lower Clear Fork dolo-
stones range from 2.96 to +2.97% PDB (Figure 24).
With the exception of the lightest data, most of these
values are similar to Leonardian data reported by
Ye and Mazzullo (1993), Saller and Henderson (1998),
and Ruppel (2002), but very different from data re-
ported for Guadalupian rocks. Typical d18O values for
Guadalupian (San Andres and Grayburg formations)
platform dolomites are 3 – 6% (Vogt, 1986; Ruppel
and Cander, 1988a, b; Saller and Henderson, 1998;
Ruppel and Bebout, 2001; Ruppel, 2002). The rela-
tively enriched isotopic signatures of Guadalupian
dolomites have invariably been interpreted to have
been produced during dolomitization by evapora-
tively concentrated seawater brines (Bein and Land,
1982; Bebout et al., 1987; Ruppel and Cander, 1988a, b).
Taken collectively, the relatively depleted values for
the Leonardian dolostones at Fullerton and in other
Leonardian fields suggest that either (1) dolomitiza-
tion was caused by brines that were less evaporatively
concentrated than those that caused Guadalupian
dolomitization; (2) dolomitization was caused by fluids
of a mixed water origin; or (3) a combination of cases 1
and 2 above. Systematic spa-
tial variations in d18O and
d13C values of dolomites in
the Wichita and Lower Clear
Fork suggest an even more
complex history of dolomi-
tization and diagenesis.
Stable isotope values from
upper Wichita peritidal do-
lomites, for example, define
two spatially distinct trends
(Figure 25a). d18O and d13C
values from high-porosity do-
lomites along the other mar-
gins of the Wichita tidal flat
display relatively light values
(average: 0.0% d18O, +2.6%
FIGURE 25. Maps showing
trends of facies, porosity, and
stable isotope data in the upper
Wichita L1-L2 tidal-flat succes-
sion (a) and the Lower Clear
Fork L2 subtidal succession (b).
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 379
d13C). By contrast, values from low-porosity Wichita
peritidal dolomites elsewhere in the field are distinct-
ly heavier, especially with respect to d18O (average:
+2.8% d18O, +3.0% d13C). These marked differences
suggest that two mechanisms of dolomitization were
involved in the diagenesis of Wichita dolomites. Iso-
topically light d18O values along the arcuate belt of
higher porosity tidal-flat facies are consistent with
early, seawater-dominated dolomitization. The heavier
d18O values exhibited by updip dolomites are typical
of dolomitization by refluxing brines. Similar heavy
d18O values are encountered in overlying Lower Clear
Fork dolomites (average: +2.5% d18O). This isotopic
similarity suggests that dolomitization of nonporous
Wichita tidal-flat facies may have been caused by
evaporatively concentrated brines generated during
Lower Clear Fork time.
Limestone isotope data from Fullerton come en-
tirely from the Lower Clear Fork. Values of d13C range
from +1.61 to +5.46% PDB (n = 30), whereas d18O
data range from 3.89 to 1.55% PDB (Figure 24).
These data are very different from previously reported
values for other calcite-bearing samples in the Leo-
nardian and Guadalupian. Most reported d18O values
for Guadalupian calcite have ranged from about 7.6
to 10.4 (Leary, 1985; Vogt, 1986; S. Ruppel, 1996,
unpublished data from the Grayburg Formation at
South Cowden field). These values, however, were re-
corded from replacement calcite interpreted to have
precipitated from meteoric water possibly sourced
from deep basin fluids. The d13C values from these
replacement calcites are also very depleted (19 to
30% PDB), suggesting precipitation from bacteri-
ally mediated sulfate reduction (Leary, 1985; Vogt,
1986; Ruppel and Cander, 1988b). The more normal
d13C values for the Lower Clear Fork limestones at
Fullerton indicate a very different origin for these
calcites. Lower Clear Fork d18O values are very similar
to the current best estimate for marine calcite pre-
cipitates from seawater during the middle Permian
(2.8% PDB; Given and Lohmann, 1985; Lohman
and Walker, 1989). This suggests that Lower Clear
Fork limestones contain a preserved record of original
seawater chemistry and by extension that these rocks
have undergone relatively little chemical alteration.
Karst Development
Karst Fabrics
Karst diagenesis can be a major factor in reser-
voir heterogeneity and compartmentalization in car-
bonate successions (Loucks, 1999). Evidence of karst-
related diagenesis and dissolution is common in the
lower part of the reservoir section at Fullerton field;
karst features occur within the Wichita and in the top
of the Abo. These rocks are variable in fabric but
include four basic types: (1) polymict conglomerates,
(2) monomict breccias, (3) fractured and tilted beds,
and (4) void-filling cement.
The predominant style of polymict conglomerate
typically consists of rounded clasts of multiple peri-
tidal lithologies (Figure 26). Clasts range in size from
a few millimeters to several centimeters in maximum
dimension and are commonly subequant and rounded.
Clasts are commonly enclosed in mudstone or abut
one another at stylolitic contacts. Polymict fabrics are
most common in the middle of the Wichita (Figure 26).
Intervals of polymict conglomerate of at least 7.5 to as
much as 18 m (25 to as much as 60 ft) thick are present
in the Exxon FCU (Fullerton Clear Fork Unit) 6122
core. The multiple facies character of these clasts and
their rounded character indicate that they were formed
by sediment transport. Their discontinuous nature
and their association with other features indicative of
karst processes suggest that they originated as cave-
fill deposits.
Although probably not true polymict conglom-
erates, superficially similar deposits that also indi-
cate karst processes are present at the contact of the
Wichita and the Abo at the downdip edge of the field
area. These rocks are commonly characterized by a
mixture of two or more facies types, including Wichita
tidal-flat facies, green silty carbonate, dark-gray silty
carbonate, and Abo subtidal facies. Instead of being
interbedded with one another, the first three sur-
round the Abo facies in many cores. This suggests
that these polymict conglomerates actually represent
dissolution and/or erosion of the top of the Abo and
subsequent infilling of the irregular surface or differ-
ential compaction of the Abo (L1) and overlying trans-
gressive Wichita sediments (L2) at the Abo–Wichita
contact. In rare instances, there are also features sug-
gestive of collapse brecciation. All of these features
have been observed in analogous outcrops of the
Abo –Wichita contact in Apache Canyon in the Sierra
Diablo (Kerans et al., 2000). The outcrop succession
reveals that karst features (sinkholes and caves) were
formed during post-L1 sea level fall and then filled
with transgressive L2 tidal-flat deposits (i.e., Wichita
facies). These tidal-flat facies, in some instances, were
brecciated and intermixed with the underlying Abo
in sinkholes and collapsed caves.
Monomict breccias or conglomerates consist of
broken and rotated clasts of constant lithology and
380 / Ruppel and Jones
facies (Figure 27). These rocks are restricted to the
Wichita, commonly the middle of the section. Com-
monly associated with these deposits are fractures,
sediment infill (cracks and fissures), void-filling ce-
ment (chiefly anhydrite), and other evidences of dis-
solution. Most of these features can also be formed by
nonkarst processes in tidal-flat intervals not exposed
to true karst (e.g., tepee formation is commonly ac-
companied by broken and rotated blocks, cracks and
fissures, and cement and sediment infill). Thus, it is
possible that some of these deposits may not be the
result of true karst processes. However, their thickness,
abundance, and association with other features of karst
formation suggest that many are also karst related.
FIGURE 26. Core slab and image log
photographs of polymict conglomer-
ates in the Wichita Formation of prob-
able karst origin. (A) Exxon FCU 6122,
depth: 2195 m (7201 ft). (B) Amoco
University Consolidated IV-25, depth:
2199 m (7213 ft). Cores 10 cm (4 in.)
wide. (C) Image log of conglomerate in
the Wichita. Exxon FCU 2564, depth:
2186– 2187 m (7171– 7174 ft). Image is
63.5 cm (25 in.) wide.
Fractured and tilted beds are also
observed in some cores, especially in
the middle of the Wichita section.
Tilted beds have clearly been formed
by postdepositional collapse. In the
Exxon FCU 5927 core, they form part
of a succession that very much re-
sembles the classic cave fill-cave roof
succession described by Loucks (1999).
The succession in the Exxon FCU
5927 well consists of 6 m (20 ft) of
polymict cave-fill conglomerate com-
posed of mixed tidal-flat facies over-
lain by 20 ft (6 m) of fractured and
locally tilted but apparently gener-
ally in-situ beds of tidal-flat and
subtidal facies probably representa-
tive of a cave roof.
Large zones of void-filling anhy-
drite cement are also strong indica-
tors of karst-related dissolution. Zones
of massive anhydrite as much as
0.5 m (1.5 ft) thick are present in
cores in the Wichita at Fullerton field
(e.g., Exxon FCU 6122, 5927). Smaller
anhydrite-filled voids and fractures
are ubiquitous in the Wichita and at
the Wichita–Abo contact and also point to late ce-
mentation of dissolution voids by diagenetic fluids as-
sociated with reflux dolomitization of the succession.
Causes and Timing of Karst
Outcrop studies demonstrate that major karsting
of the Leonardian sequence occurred at the sea level
fall and rise event that is defined by the L1-L2 se-
quence boundary (Kerans et al., 2000). In downdip
areas of Fullerton field, cores demonstrate this same
relationship. Here, the top surface of the Abo outer
platform facies succession (L1) is karsted and is in-
filled and overlain by brecciated Wichita tidal-flat
facies (L2). In updip areas, however, establishing a
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 381
spatial and temporal relationship between karst for-
mation and the L1-L2 sequence boundary is more
problematic. Most karst features in the Fullerton field
area are found in an interval of about 45 m (150 ft) in
the middle of the Wichita section. In general, this
interval correlates approximately to the interpreted
L1-L2 boundary (Figure 5). However, karst features
are developed both below and above this horizon,
indicating that karst-related diagenesis was not lim-
ited to the L1 sequence. Instead, it appears that two
types of karst-related processes occurred as outlined
below.
Primary karsting and dissolution probably occurred
during the post-L1 lowstand. At this time, the exposed
FIGURE 27. Core slab photos of mo-
nomict breccias from the Wichita.
(A) Rotated clasts of tidal-flat facies. Exxon
FCU 5927, depth: 2203 m (7230 ft).
(B) Tilted blocks of laminated peritidal
facies. Exxon FCU 5927, depth: 2199 m
(7215 ft). (C) Brecciated clasts of peloidal
wackestone. Amoco University Consolidat-
ed V 15, depth: 2103 m (6899 ft). (D) Clasts
of peritidal mudstone-wackestone. Exxon
FCU 5927, depth: 2204 m (7233 ft).
Cores are 10 cm (4 in.) wide.
top L1 surface (consisting of Abo sub-
tidal sediments downdip and lower
Wichita tidal-flat sediments updip)
developed local caves, sinkholes, and
an irregular topography. During the
subsequent L2 sea level rise, trans-
gressive peritidal deposits of the Wi-
chita filled the irregular karsted sur-
face, including sinkholes and caves.
With continued sedimentation and
compaction, parts of the overlying
L2 (upper Wichita tidal flat) suc-
cession underwent local breccia-
tion and collapse probably because
of stress differences set up over un-
derlying karst features. This scenario
is consistent with outcrops (Kerans
et al., 2000) and with the distribution
of karst features seen at Fullerton
field.
Impact of Karsting on
Reservoir Quality
Some karst-related deposits at Ful-
lerton most certainly exhibit at least
local differences in petrophysical
properties (i.e., porosity, permeabil-
ity, and saturation) from surrounding undisturbed
and unaltered deposits. Polymict conglomerates and
anhydrite voids are two obvious examples. However,
two factors make quantification of the importance of
these differences difficult. First, most karst deposits
do not record significantly different porosity or perme-
ability from that of surrounding nonkarsted deposits
on both wire-line and core data. This is probably be-
cause of the fact that most karst fills are composed of
the same facies as nonkarsted intervals. Anhydrite
void fills are an obvious exception to this because
they contain no porosity or permeability; but they are
generally very small and probably of little impact on
reservoir flow.
382 / Ruppel and Jones
Second, with the exception of areas in which there
are cores, the distribution of karst features in the
reservoir section is not definable. Efforts to identify
karst fill using logs and 3-D seismic data are thwarted
by the similar lithological and petrophysical proper-
ties these features share with surrounding rocks. It is
tempting to conclude from these observations that
karst features have no impact of reservoir heteroge-
neity or fluid flow. However, there are anomalies in
water production and flow rates in the Wichita that
cannot be readily explained by matrix petrophysical
properties (T. Anthony, 2003, personal communica-
tion). These phenomena may be the result of karst
development.
RESERVOIR IMAGING
Accurate definition of reservoir architecture and
the distribution of rock fabrics in this architecture is
the key to defining improved methods for recovery
of hydrocarbons remaining in these systems. We
used several methods to better image the reservoir at
Fullerton field. Especially important in defining the
geologic architecture of the reservoir are (1) the cali-
bration and use of borehole image logs to aid in the
identification and mapping of facies, cyclicity, and
rock fabrics; and (2) the use of 3-D seismic data to
constrain the geologic framework.
Identifying Facies and Cyclicity from
Borehole Image Logs
Image logs are highly underused in the character-
ization of carbonate reservoirs. To most, the principal
use of such logs is in the identification of fractures.
However, image logs also have the potential to accu-
rately image many matrix properties that are key to
the proper characterization, modeling, and exploita-
tion of carbonate reservoirs. Traditionally, cores have
been obtained to provide the key data needed for con-
straining the distribution of reservoir facies, cyclicity,
and rock fabrics. However, because of cost, the num-
ber of cores available is generally far smaller than
needed to accurately constrain these elements. If
properly calibrated with core observations, borehole
image logs can provide most of the required data to
construct an accurate reservoir model at a fraction of
the cost.
Examination, correlation, and comparison of im-
age log character with cores show that seven facies
can reliably be identified from the image log in the
Abo–Wichita–Clear Fork succession at Fullerton. This
is fewer than the 12 facies defined from core studies
but is sufficient to provide necessary information to
define major facies successions and cyclicity and to
provide a strong basis for accurate correlation and
interpretation of facies and cyclicity to nearby wells.
Facies recognizable on image logs include (1) tidal-flat
facies (Figure 8B), (2) peloid wackestone-packstone
(Figure 10B), (3) nodular wackestone-packstone
(Figure 9B), (4) cross-bedded grainstone (Figure 13B),
(5) fusulinid wackestone-packstone (Figure 14B),
(6) karst breccia (Figure 26C), and (7) clay-rich mud-
stone. Additionally, siltstone-sandstone facies can be
identified from many wire-line logs.
The identification of tidal-flat facies is especially
crucial for defining both reservoir architecture and
reservoir quality. Because they generally occupy cy-
cle tops, their definition makes it possible to define
cycle boundaries and thereby facilitates cycle-scale
correlation. Equally important is the identification of
tidal-flat facies for their petrophysical significance.
As discussed previously, tidal-flat facies commonly
display relatively high porosities but low permeabil-
ity. Because of this, it is critical for accurate reservoir-
quality mapping to distinguish tidal-flat rocks from
subtidal rocks that are typically higher in permeabil-
ity. Definition of the fusulinid wackestone facies is
also of great importance because these rocks gener-
ally represent the deepest water facies in the Leo-
nardian succession and typically are found at cycle
bases. They are thus important indicators of sea level
rise and guides to cycle definition and correlation.
Using calibrated image log responses, it is possible
to create a very detailed record of facies in wells
having good-quality image logs. Vertical resolution is
potentially even better than that attainable from
cores (because of uncertainties of core and well depth
ties). Cycles defined from image logs range in thick-
ness from 1 to 6 m (3 to 20 ft). Highest facies and cycle
resolution was obtained in intervals characterized by
alternating subtidal and tidal-flat facies in the Lower
Clear Fork because of the marked contrast in image
log character of these two facies types.
Imaging Stratigraphic Architecture and Reservoir
Development from 3-D Seismic Data
Like image logs, 3-D seismic data are remarkably
underused in the characterization of carbonate res-
ervoirs. A critical need in developing approaches that
lead to improving recovery from reservoirs contain-
ing significant volumes of remaining hydrocarbons
is a better understanding of the 3-D distribution of
reservoir attributes. The approaches outlined in this
report for assembling and interpreting well data are
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 383
the most critical part of this effort. However, because
they are limited to well control, they leave important
gaps in our understanding. Three-dimensional seis-
mic data offer valuable data on interwell and extra-
well areas (meaning areas of the field where usable
well data are absent) and when properly interpreted
and applied can greatly improve and, thus, constrain
reservoir attribute models. Here, we provide some
brief insights into how 3-D data can be used to better
image both reservoir framework and porosity distri-
bution. A far more detailed and rigorous application
of these 3-D data to the construction of the geological
model at Fullerton is described by Zeng (2004).
Constraining Reservoir Architecture
Two- and three-dimensional seismic data at Ful-
lerton field provide important guides to the stratal
architecture of the reservoir succession. Seismic am-
plitude sections through the entire reservoir interval
reveal that much of the section is characterized by
generally parallel seismic reflectors (Figure 28). This
is not unexpected, considering the shallow-water-
platform depositional setting indicated by the cores
and the apparently subhorizontal correlations sug-
gested by wire-line logs. However, 3-D and some 2-D
data suggest a very different architecture for the basal
Leonardian, Abo Formation. Dip-oriented seismic lines
reveal sets of clinoformal reflections in the Abo that
dip generally basinward (toward the east). This appar-
ent clinoformal architecture of the Abo is consistent
with observations of clinoformal fusulinid wacke-
stones and packstones in outcropping Abo-equivalent
sections in the Sierra Diablo. These clinoforms dem-
onstrate that conventional horizontal correlations of
FIGURE 28. Fullerton field 3-D
seismic section (in time) show-
ing general continuity and iso-
pachous nature of Lower Clear
Fork and Wichita reservoir in-
tervals and contrasting clino-
formal nature of the Abo. Yellow
lines are time lines; dotted
line defines the top of Abo
Formation.
wire-line-log data are inap-
propriate for the Abo. In most
cases, it is not possible to re-
solve cycle-scale correlations
of either facies or time sur-
faces in such settings.
Defining Reservoir Quality
Seismic data can also be robust indicators of po-
rosity distribution in carbonate reservoirs. Zeng
(2004) demonstrated the strong agreement between
3-D seismic impedance data and reservoir porosity at
Fullerton field. Because of this robust relationship,
even simple amplitude extractions reveal field-scale
and fieldwide changes in porosity development that
are important for understanding controls on porosity
development and for defining interwell and extra-
well distribution of porosity.
Figure 29 is an amplitude extraction map for HFS
2.1 in a small area of the field created to examine
possible infill drilling locations. Note that whereas
the northern half of the map displays high negative
amplitudes indicative of high porosity, the southern
half of the area is characterized by markedly lower
amplitude. The porosity distribution revealed by this
map indicates that only those wells in the northern
half of the target area are likely to encounter good
porosity in this zone. It should be pointed out that
this information on porosity distribution is only ob-
tainable from the 3-D data volume; the quality of the
well logs in the area is too poor to determine porosity.
Areas of poor well control like these (caused by
either the lack of wells or poor-quality logs) are com-
mon through the field, and they greatly compromise
efforts to develop a fieldwide strategy to identify and
target oil resources. This is readily apparent from a
comparison of a porosity map based on well logs and
a seismic amplitude map for HFS L2.0 (Figure 30). The
value of the seismic data volume is especially ap-
parent in the southern end of the field where the
limited well control does not accurately image the
384 / Ruppel and Jones
FIGURE 29. Map of negative amplitude data extracted
from Fullerton 3-D data showing porosity distribution
in the L2.1 sequence. Data show that proposed wells in
the northern half of the area will encounter reservoir
porosity, whereas those in the southern half will not.
east – west-trending area of high porosity shown by
the amplitude map (compare Figure 30a with b). It is
apparent that basic amplitude data like these provide
a powerful supplement to well control in defining
and predicting porosity distribution in the field.
RESERVOIR FRAMEWORK
A critical component of a robust reservoir model is
a geologically constrained reservoir framework. Geo-
logically accurate models must be based on correla-
tions of time-stratigraphic units, the most readily
correlative of which are cycle and sequence bound-
aries. At Fullerton field, we used correlations of cycle
tops to construct the reservoir model for the Clear
Fork. Although these correlations are ultimately based
on correlations of wire-line logs, the underlying basis
for the interpretation and correlation of these logs is
a knowledge of 1-D and 2-D facies and cycle-stacking
relationships developed from integrated studies of
cores and outcrops.
Cycle-based Flow-unit Definition
Reservoir flow units are most appropriately based
on the definition and mapping of depositional cycle
boundaries. This is because when properly defined
and correlated, depositional cycles represent the best
available indicator of original depositional surfaces
or time lines. The procedure for identifying and
using cycle boundaries for flow-unit definition has
been well described by Ruppel and Ariza (2002) and
Lucia and Jennings (2002) for the South Wasson
Clear Fork reservoir.
Ideally, the correlation of depositional cycle tops
should be based on a log that can be directly tied to
facies and one that is independent of diagenesis or
porosity development. The gamma-ray log in certain
ideal settings serves this function. However, through-
out most of the Leonardian carbonate succession (in
fact, throughout most of the Permian carbonate sec-
tion in the Permian basin), the gamma-ray log is not
accurate for detailed correlation because of variable
volumes of uranium, potassium, and thorium. The
spectral gamma-ray log can help to distinguish varia-
tions in these elements and, thus, is useful in sepa-
rating clastic-rich sections (that contain high levels
of potassium and thorium). Because small volumes of
clastics are commonly associated with tidal-flat facies,
the spectral gamma ray is locally helpful in distin-
guishing tidal-flat facies from subtidal facies (Ruppel,
1992, 2002). However, variations in uranium content
are not always systematic, and because of this, gamma-
ray response commonly varies independently of facies.
Resistivity logs can also locally be used to separate
tidal-flat facies from subtidal facies on the basis of
differences in saturation (Ruppel, 2002). However,
even under ideal circumstances, neither gamma-ray
logs nor resistivity logs can accurately depict facies in
Leonardian rocks.
Accordingly, we used porosity logs to define and
correlate facies and cycle tops in the Lower Clear Fork
section at Fullerton field. The basis for this approach
comes from integrated studies of cores and wire-line
logs at Fullerton that demonstrate two important at-
tributes of Leonardian cycles. First, cycle-capping fa-
cies are either grain-rich subtidal or tidal-flat facies.
Second, porosity is most commonly associated with
these facies and, therefore, is most typically developed
at cycle tops. These relationships are key to the use of
wire-line logs for correlation of cycles and of flow units.
It is important to understand that there are limi-
tations to the accuracy of the porosity-log correlation
method. For example, because of diagenesis, all carbon-
ate facies display some variations in porosity. Cycle-
top facies in the Leonardian, whether subtidal grain-
stone or tidal flat, do exhibit lateral changes in porosity.
So locally, cycle tops defined from porosity may be
slightly mispositioned (e.g., because of a local decrease
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 385

FIGURE 30. Porosity development in HFS 2.0 (upper Wichita) across entire field based on wire-line logs (a) and 3-D
seismic data (b). Note that 3-D amplitude patterns suggest that well-based porosity mapping is inaccurate in some areas.

in porosity in the cycle-top facies or a local increase in
porosity in the overlying cycle-base facies). From a
strict chronostratigraphic point of view, this means
that some cycle tops are incorrect, and that cycle cor-
relations locally cross time lines. From a reservoir mod-
eling point of view, however, this result is actually
probably preferable. That is because these Leonardian
rocks uncommonly contain cycle-base flow barriers,
and flow probably does locally cross cycle boundaries.
In any case, the porosity-log correlation method, if
properly constrained by core and outcrop calibration of
facies, porosity, and cyclicity, is the most geologically
sound basis for constructing cycle correlations and
establishing the basis for true flow-unit correlations.
It is also important, however, to correlate geologi-
cally defined cycle tops at the highest resolution pos-
sible. For example, outcrop studies demonstrate that
cycles less than 3 m (10 ft) thick can be correlatable
across large areas of the platform (Ruppel et al., 2000).
Where wire-line data permit, an effort should be
made to correlate these thin cycles through the res-
ervoir as well, even if later upscaling is planned for
reservoir modeling. This is important for two reasons.
First, coarser scale correlations are much more likely
to be in error. At Fullerton, we found that early cor-
relations made at the sequence or cycle set scale were
later proven to be off by a cycle or two after we recor-
related the succession at the cycle scale. From a res-
ervoir modeling point of view, this means that flow
units defined by coarse-scale correlations are more
likely to cross-connect flow layers than those defined
by finer scale correlations. Second, upscaling (group-
ing of cycles into thicker flow-unit packages for
modeling) is more likely to retain the original geo-
logical architecture if based on fine-scale (i.e., cycle-
scale) correlations. Accordingly, we attempted to cor-
relate the reservoir succession at the highest possible
level of detail supported by outcrop and core obser-
vations and wire-line resolution.
Lower Clear Fork Reservoir Architecture
The robustness of the use of porosity logs for
defining facies and cyclicity is apparent from core
and log relationships in HFS L2.2. A comparison of
386 / Ruppel and Jones
core data and porosity logs in the Exxon FCU 6122
well (Figure 31) shows that porosity is nearly entirely
associated with cycle-top, grain-rich subtidal facies.
Thus, porosity logs can be used to define both facies
and cyclicity. Other cored wells in the field exhibit
the same core and log relationship.
Using this relationship, we defined and correlated
15 cycles in the Lower Clear Fork. The average thick-
ness of these cycles is about 5 m (17 ft) for L2.1 and
L2.2 throughout the entire interval. However, cycle
thickness varies systematically by sequence in this in-
terval. High-frequency sequence L2.2 cycles average
about 3.3 m (11 ft) in thickness, whereas L2.1 cycles,
with the exception of low-accommodation tidal-flat
cycles at the base and top of the sequence, are nearly
twice as thick. This is probably caused by two factors.
First, L2.1 deposits record the maximum flooding of
the platform and probably the development of maxi-
mum accommodation. The overall upward thinning
of cycles from L2.1 to L2.2 is consistent with an
overall upward decrease in accommodation. Second,
FIGURE 31. Comparison of core-
defined facies and cyclicity with po-
rosity logs in subtidal cycles of HFS
L2.2. Exxon FCU 6122. The systematic
relationship between cycle top facies
and higher porosity permits porosity
logs to be used for cycle and flow unit
correlation.
lateral changes in facies stacking in
L2.1 cycles suggest local variations in
sediment accumulation patterns
that may have been caused by topo-
graphic relief on the platform during
L2.1 flooding. The resultant com-
plex vertical and lateral facies distri-
bution patterns make accurate defi-
nition of cycles difficult. As a result,
cycles defined for L2.1 may actually
reflect combinations of cycles or cy-
cle sets.
Wichita Reservoir Architecture
Rigorous cycle definition is not
possible for the Wichita because
of the preponderance of very low-
accommodation, tidal-flat facies.
Both core and outcrop studies show
that rocks deposited in such settings
uncommonly display systematic
trends in vertical facies stacking and
generally exhibit very low lateral facies continuity.
Accordingly, patterns of depositional facies do not
define extrinsic controls (e.g., sea level rise and fall)
but, instead, local controls on sediment accumulation
(e.g., paleotopography and climate). Thus, for the
Wichita, it is necessary to use diagenetic features to
define the reservoir framework.
Key diagenetic features used to define Wichita
cycles are mineralogy and porosity. In the northern
part of the field, the upper Wichita contains multiple
intervals of low-porosity limestone that we have in-
terpreted to be cycle bases (Figure 18a). Although in-
terpreted largely from diagenetic relationships, cycles
are closely tied to depositional surfaces and, thus, rep-
resent a good approximation of time surfaces. These
surfaces are also especially useful from a reservoir
point of view because they define layers of low and
high porosity and permeability. The top four cycles
in the Wichita are defined on the basis of these cyclic
dolostone (high-porosity)-limestone (low-porosity)
couplets.
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 387
However, no limestone is present in the upper
Wichita in the southern part of the field area nor in
the lower part of the Wichita in any part of the field
(Figure 18a). In these areas, it was necessary to use
porosity trends alone to construct the reservoir frame-
work. Although it is still likely that porosity variations
in these areas are closely tied to diagenesis associated
with depositional surfaces (i.e., cycle tops), we have
little independent evidence (i.e., mineralogical varia-
tions) with which to demonstrate this. Accordingly,
the framework established for these parts of the Wi-
chita is far less geologically robust than that defined
for other parts of the reservoir. For this part of the
reservoir, we defined and correlated 12 surfaces on
the basis of porosity. We have confidence that these
surfaces are subparallel to time surfaces on the basis
of their parallelism to overlying Lower Clear Fork
cycle-top surfaces and to a middle Wichita marine
flooding surface. However, we cannot tie them rig-
orously to cyclicity.
Abo Reservoir Architecture
The Abo is dominated by clinoformal bedding
typical of outer-ramp carbonate deposits; this is dem-
onstrated both by outcrop studies (Kerans et al., 2000)
and by 3-D seismic data at Fullerton (Figure 28). It is
therefore certain that the Abo architecture differs
significantly from the generally subparallel character
of the depositional surfaces of the platform-top Wi-
chita and Lower Clear Fork successions. Outcrop
studies and seismic studies at Fullerton and elsewhere
(Kerans et al., 2000; Ruppel et al., 2000) also dem-
onstrate that clinoformal outer platform successions
like the Abo do not contain readily correlatable cyclic
successions. This is caused by changes in sources and
distributional patterns and to extreme variations in
lateral textures and fabrics of these transported
sediments. Accordingly, it is not possible to establish
an accurate internal architecture for these deposits.
For purposes of reservoir modeling, we created a series
of conceptual clinoform surfaces to constrain the
reservoir architecture. Although these surfaces do
not accurately describe the architecture of the Abo,
they do illustrate the nonparallel and nonhorizontal
nature of these deposits and provide more realistic
constraints for modeling of reservoir attributes.
Reservoir Model
The architecture of the reservoir framework de-
veloped for the Clear Fork reservoir at Fullerton is
depicted by Figure 32. Key aspects of this model are
(1) the subhorizontal and subparallel nature of Clear
Fork and upper Wichita (L2) surfaces; (2) the effec-
tive pinch-out of lower Wichita (proximal L1) sur-
faces at the facies boundary from Wichita tidal flats
to Abo subtidal; and (3) the clinoformal nature of
Abo (distal L1).
Fieldwide Patterns of Porosity Distribution
Core and log data from wells that penetrate the
Abo indicate that the Abo locally contains high po-
rosity. However, well control is too sparse and in-
complete to accurately map porosity distribution.
Wang and Lucia (2004) presented a realization of the
3-D distribution of Abo porosity using a full-field
model based on available porosity data and a con-
ceptual geological framework. However, the accura-
cy of this porosity distribution must be considered
relatively low.
Mapping of phih (porosity  thickness) from wire-
line logs demonstrates that the Wichita Formation
contains higher total phih (and higher average po-
rosity) than either the Abo or the Lower Clear Fork. In
general, maps show that the highest phih lies along
the field structural crest, suggesting that porosity de-
velopment is related to structure. However, a more de-
tailed examination of phih by stratigraphic horizon
shows that phih is related more to platform position
than structure. The upper Wichita (HFS L2.0 TST)
displays a well-defined, arcuate belt of higher phih
(Figure 30) that generally parallels but is slightly
landward of the inner-outer platform boundary (the
Wichita– Lower Clear Fork facies boundary). Decrease
in phih to the east is a function of the decreasing
thickness of the Wichita near the facies transition
(Figure 19). The decrease in phih to the northwest,
however, is not associated with any change in depo-
sitional facies; Wichita rocks comprise similar peri-
tidal tidal-flat facies across the entire area. Accordingly,
the cause of porosity development must be more a
function of diagenesis than deposition. A model for
this diagenesis is presented in the following section.
Lower Clear Fork HFS L2.1 also displays an arcuate
trend in porosity development that is very similar to
that seen in the upper Wichita (Figure 33; see also
Figure 25b). In contrast with the Wichita, there is
evidence that this porosity trend is, at least in part,
associated with facies. Areas of highest porosity are
associated with middle-ramp, grain-rich packstones.
Lower porosity areas are associated with outer-ramp
fusulinid-rich wackestones downdip and mud-rich
packstones updip. It is probable that the high-phih
belt represents a low-energy ramp crest. However, the
388 / Ruppel and Jones

FIGURE 32. Diagrammatic cross section showing framework used for reservoir model construction and general distribution
of major rock-fabric classes at Fullerton Clear Fork field. Class 1 rock fabrics display the highest permeability relative to
porosity, whereas class 3 fabrics display the lowest. It should be noted that the Lower Clear Fork displays far more complex
lateral variations in rock fabrics than depicted here. Datum is sea level.

change in phih appears also to be significantly re-
lated to diagenesis. Areas of high porosity are also
areas of high limestone abundance. An almost 1:1
relationship exists between the presence of limestone
and high-porosity areas (compare Figures 18b, 33).
No similar trend of high phih is apparent in HFS
L2.2. However, like L2.1, areas of highest porosity are
associated with limestone. For L2.2, this is domi-
nantly in the southern end of the field; however, a
small area of limestone also exists in the northwest
corner of the field (Figure 18b).
Models for Porosity Development
Porosity is developed in all parts of the Abo –
Wichita– Lower Clear Fork reservoir section. Howev-
er, there are obvious areal variations in porosity that
reflect a combination of depositional, diagenetic, and
structural controls on reservoir development. Most
reservoir intervals exhibit spatial variations in poros-
ity development that parallel platform paleotopog-
raphy. This is particularly apparent in the Wichita
HFS L2.0 and in the Lower Clear Fork HFS L2.1
(Figures 30, 33). In each of these cases, belts of high
porosity are situated immediately up depositional
dip (landward) from the position of the underlying
inner-outer platform margin (Wichita – Abo facies
boundary). However, the mechanism for porosity
development must differ for the Wichita, which com-
prises peritidal facies that display no systematic varia-
tions across the area, and the Lower Clear Fork, which
consists of subtidal rocks of three facies tracts.
The absence of depositional facies variations in
the Wichita indicates that porosity formation is the
result of diagenetic processes. Stable isotope data
suggest that high-porosity dolomites in the Wichita
are the result of seawater-dominated diagenesis
(Figures 25a, 34a). These rocks display light d18O
values (average: 0.0% PDB) that are consistent with
seawater dolomitization (based on an approximate
3% fractionation from penecontemporaneous sea-
water calcite; Land, 1980). The restriction of these
high-porosity rocks to the outer margin of the Wichita
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 389

FIGURE 33. Map of wire-line phih showing distribution of

porosity in the Lower Clear Fork HFS 2.1. Areas of high
porosity in HFS L2.1 correlate with areas of abundant
limestone. Compare with Figure 17.

tidal flat suggests that porosity development may be

the result of early seawater-dominated dolomitization
and stabilization along the seaward margin of the tidal
flat. Dolomitization may have been favored along this
seaward margin by the development of somewhat
shallower water conditions, perhaps like those seen in
tidal-flat island complexes (Figure 34a).
The Lower Clear Fork HFS 2.1 also displays a trend
in porosity development that closely parallels the
underlying Wichita– Abo and Wichita– Clear Fork
facies transitions (Figures 25b, 33). In this case, how-
ever, both mineralogy and porosity development are
related to facies. Core descriptions imply that the
trend of high porosity is coincident with a platform FIGURE 34. Models depicting three major stages of
diagenesis and porosity formation in the upper Wichita
ramp crest, an area characterized by more common
and the Lower Clear Fork (L2 sequence). (a) Early seawater
ooid-bearing, grain-rich packstones and grainstones dolomitization: Wichita, L1-L2. Wichita porosity develop-
(e.g., Figure 13C) and by locally more abundant tidal- ment is associated with early seawater dolomitization
flat caps. Ramp-crest development was probably con- along the outer margins of the tidal-flat complex. (b) Early
trolled by inherited paleotopography over the L1 seawater cementation and stabilization: Lower Clear Fork,
L 2.1. Porosity is related to preservation of primary po-
(Wichita –Abo) inner to outer ramp margin and/or
rosity in Lower Clear Fork ramp-crest limestones. (c) Reflux
differential subsidence over deep-seated faults in the dolomitization. Most of the Wichita and Lower Clear Fork
same way that the position of the Wichita– Abo facies was dolomitized by evaporatively concentrated, reflux
transition is most likely controlled by such deep brines. See text for discussion.
390 / Ruppel and Jones
structures. The ramp-crest trend is an area of abun-
dant calcite; lower porosity areas to the east (inner
ramp) and west (outer map) are nearly entirely dolo-
mite (see Figure 18). Strontium and oxygen isotope
data ( J. Kaufman, 1991, personal communication;
this study) from the calcites indicate that these rocks
are essentially unaltered original marine precipitates.
Oxygen isotopes (d18O = 3.0%) are virtually iden-
tical to best estimates of marine precipitates of middle
Permian age (d18O = 2.8%; Given and Lohmann,
1985). Collectively, these data suggest the HFS L2.1
porosity trend is the result of early marine calcite
cementation in well-agitated and well-oxygenated
conditions along the ramp crest (Figure 34b). Lower
porosity in updip dolomites is probably the result of
predolomitization compaction of these muddier sedi-
ments and possibly of porosity occlusion by dolomite
and anhydrite cementation. Compaction and dolo-
mitization, and, thus, porosity loss, were probably
limited in the ramp crest because of early calcite ce-
mentation and stabilization and the grain-rich tex-
tures that are dominant there.
Most Wichita and Lower Clear Fork rocks outside
these two high-porosity corridors exhibit lower po-
rosities and a very different stable isotopic character.
These rocks contain much heavier oxygen isotopes
(d18O average 3.0% PDB) that are indicative of do-
lomitization by evaporatively concentrated refluxing
brines (Saller and Henderson, 1998; Ruppel, 2002).
Such brines may have been generated at multiple
times during the Leonardian. Conditions necessary
for reflux were probably developed at each of the
documented sequence boundaries (Figure 34c). How-
ever, it seems likely that the greatest potential for
major reflux was at or soon after the L1-L2 (Wichita)
or L2-L3 (Lower Clear Fork– Tubb) sea level lowstands
(Figure 34c). The similarity in isotopic character
throughout both Wichita and Lower Clear Fork re-
flux dolomites suggests that all were dolomitized by a
single massive reflux event, presumably at about the
L2-L3 sea level fall and rise.
Although most reflux dolomites at Fullerton have
low porosities, grain-rich, Lower Clear Fork dolomite
facies typically exhibit good porosity and perme-
ability (e.g., Figure 13D). This demonstrates the key
function of depositional texture in porosity preser-
vation in dolomites. The lack of porosity in updip
Wichita reflux dolomites, as well as in Lower Clear
Fork mud-rich dolomite facies, probably reflects pre-
dolomitization compaction and/or porosity reduc-
tion by dolomitization and sulfate emplacement dur-
ing reflux.
Rock Fabrics and Permeability Distribution
In addition to a geologically defined stratigraphic
framework and an accurate distribution of porosity, a
robust reservoir model must contain a realistic dis-
tribution of permeability. To calculate and distribute
permeability at Fullerton, we used the rock-fabric
approach developed by Lucia (1995). This approach
is based on defining and mapping key rock fabrics.
Figure 32 displays the stratigraphic distribution of
rock fabrics and associated petrophysical classes
(groupings of rock fabrics having similar relationships
between porosity and permeability) in the Fullerton
reservoir. Abo outer platform skeletal packstones are
dominated by rock fabrics with interparticle pores
representative of petrophysical class 1 (which exhib-
its the highest permeability for a given porosity).
Wichita tidal-flat facies, by contrast, comprise petro-
physical class 2 rock fabrics (which have the lowest
permeability for a given porosity). Rocks of the Lower
Clear Fork are more complex than either the Abo or
Wichita in terms of type and distribution of rock
fabrics. As suggested by Figure 32, subtidal Lower
Clear Fork deposits include both petrophysical class 1
and 2 rocks (and locally contain mappable class 3
tidal-flat facies). The distribution of class 1 and 2 rock
fabrics is dominantly a function of diagenesis (prin-
cipally mineralogy and pore type) and secondarily
related to depositional facies. For example, most ramp-
crest grain-rich limestone packstones are class 2 rocks;
thus, their distribution can be mapped with wire-line
logs that define mineralogy (e.g., dual-porosity log
suites, PE log). Other variations in rock fabrics and
petrophysical class in the Lower Clear Fork are a func-
tion of dolomite crystal size, anhydrite cementation,
development of moldic pores, and depositional tex-
ture. We examined 950 thin sections and obtained
705 new measurements of core porosity and perme-
ability and 30 special core analysis measurements to
define these properties across the field. A detailed dis-
cussion of the methodology and results of the char-
acterization and mapping of these rock fabrics and
calculation of permeability at Fullerton is available in
Jones and Lucia (2004).
SUMMARY AND CONCLUSIONS
Both the procedures used to develop the reservoir
framework at Fullerton field and many of the at-
tributes of this framework offer important guidelines
for the characterization of shallow platform carbon-
ate reservoirs and for the development of predictive
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 391
models of the distribution of reservoir properties.
Key geological findings from this study include the
following.
The Leonardian reservoir succession in the Perm-
ian basin consists of three formations or facies suc-
cessions, each of which is characterized by distinctive
facies, cyclicity, depositional architecture, porosity
development, and petrophysical relationships.
The Abo consists largely of porous outer-ramp,
open-marine, fusulinid and crinoid facies whose clino-
formal architecture is clearly expressed on seismic
data. Lateral facies continuity is poor, and as a result,
porosity distribution is complex and probably highly
discontinuous. The relationship between permeabil-
ity and porosity, however, is among the best in the
reservoir.
The Wichita tidal-flat rocks comprise both high-
stand systems tract updip equivalents of the Abo
(Leonardian sequence L1) and TST updip equivalents
of the basal Lower Clear Fork (Leonardian sequence
L2). Cyclicity is poorly developed; reservoir architec-
ture is more controlled by diagenesis than by deposi-
tional facies. Although these deposits locally contain
high porosity, they exhibit relatively low permeabil-
ity and display poor, small-scale continuity. Areas of
high-porosity development are a result of early dolo-
mitization and stabilization along the outer margins
of the inner-platform, tidal-flat complex. Cycle-base
limestones display very low porosity and permeabil-
ity and act as local fluid-flow baffles between higher
porosity cycle-top dolostones.
The Lower Clear Fork consists of a dominantly
subtidal succession of three high-frequency sequences,
each of which records sea level rise (transgression)
and fall (regression). Reservoir development is large-
ly restricted to subtidal facies (late transgression
and early highstand). Highest porosity and perme-
ability are associated with undolomitized grain-rich,
subtidal packstones and grainstones of the ramp
crest that have been relatively unaffected by major
diagenesis.
Karst modification of the reservoir (including in-
clined beds and monomict and polymict cave-fill
breccias) is widespread at and around the L1-L2 se-
quence boundary. However, karst features are re-
solvable only with cores or image logs, and neither
wire-line log nor core analysis data reveal definitive
differences in petrophysical properties. Neverthe-
less, well-performance data suggest that karst devel-
opment may affect fluid flow.
Wire-line logs must be used for constructing the
reservoir framework. However, gamma-ray logs are
unreliable for this purpose because of the presence
of highly variable volumes of uranium that show no
consistent relationship to facies. We found porosity
logs best suited for high-resolution, cycle-scale corre-
lations, but only if they have been properly calibrated
to facies-stacking patterns and facies.
Wire-line logs do not always provide the best reali-
zation of porosity distribution. Where good-quality
well logs are lacking (e.g., between wells and in areas
with older logs) simple amplitude extractions from
3-D seismic data provide superior resolution of both
reservoir architecture and the distribution of reser-
voir porosity.
All of the forgoing findings have significant im-
pact on reservoir heterogeneity, architecture, and fluid
flow. It is thus critical that all be incorporated into
reservoir models for accurate imaging of reservoir
properties and meaningful simulation of fluid flow.
ACKNOWLEDGMENTS
The results presented in this chapter are part of
continuing research into the styles and causes of
heterogeneity in shallow-water platform carbonate
reservoirs in the Permian basin by the Bureau of
Economic Geology. Frequent, sometimes vigorous,
discussions with colleagues at the Bureau, including
Charlie Kerans, Robert Loucks, Jerry Lucia, and Jim
Jennings, have been especially helpful in formulat-
ing the interpretations presented herein. Principal
funding for the study was provided by the U.S. De-
partment of Energy, ExxonMobil Corporation, and
the University of Texas System. Additional support
was provided by member sponsors of the Bureau’s
Carbonate Reservoir Characterization Research Lab-
oratory, including Anadarko, Aramco, BP, Chevron,
ExxonMobil, Great Western Drilling, Kinder Mor-
gan, Marathon, Occidental Petroleum, Petroleum De-
velopment Oman, Shell International, Statoil, and
TotalFinaElf. Special thanks are extended to David
Smith, Terry Anthony, Steve Krohn, and Amy Po-
well of ExxonMobil and Jeff Simmons, Craig Kemp,
and John Stout of Oxy for their contributions to
organizing the project and providing data. David
Smith and Terry Anthony have been especially help-
ful in proving insights into the field geology and
engineering issues. We also express our apprecia-
tion to Stephen Hartman and Tim Hunt of the
University of Texas System West Texas Operations
Office for providing both collaborative funding and
data.
392 / Ruppel and Jones
REFERENCES CITED
Atchley, S. C., M. G. Kozar, and L. A. Yose, 1999, A pre-
dictive model for reservoir characterization in the
Permian (Leonardian) Clear Fork and Glorieta forma-
tions, Robertson field area, west Texas: AAPG Bulletin,
v. 83, p. 1031 – 1056.
Bebout, D. G., F. J. Lucia, C. R. Hocott, G. E. Fogg, and G. W.
Vander Stoep, 1987, Characterization of the Grayburg
reservoir, University Lands Dune field, Crane County,
Texas: University of Texas at Austin, Bureau of Eco-
nomic Geology, Report of Investigations 168, 98 p.
Bein, A., and L. S. Land, 1982, The San Andres carbonates
in the Texas Panhandle: Sedimentation and diagen-
esis associated with Mg-Cl brines: University of Texas
at Austin, Bureau of Economic Geology, Report of
Investigations 121, 48 p.
Dutton, S. P., E. M. Kim, R. F. Broadhead, C. L. Breton, W. D.
Raatz, S. C. Ruppel, and C. Kerans, 2005, Play analysis
and digital portfolio of major oil reservoirs in the
Permian basin: University of Texas at Austin, Bureau of
Economic Geology Report of Investigations No. 271,
287 p., CD-ROM.
Fischer, A. G., and M. Sarntheirn, 1988, Airborne silts and
dune-derived sands in the Permian of the Delaware
basin: Journal of Sedimentary Petrology, v. 58, p. 637 –
643.
Fitchen, W. M., M. A. Starcher, R. T. Buffler, and G. L.
Wilde, 1995, Sequence stratigraphic framework and
facies models of the Early Permian platform margins,
Sierra Diablo, west Texas, in R. A. Garber and R. F.
Lindsay, eds., Wolfcampian – Leonardian shelf margin
facies of the Sierra Diablo — Seismic scale models for
subsurface exploration: West Texas Geological Society
Publication 95-97, p. 23 – 66.
Garber, R. A., and P. M. Harris, 1990, Depositional facies of
the Grayburg/San Andres dolomite reservoirs; Central
Basin platform, Permian basin, in D. G. Bebout and
P. M. Harris, eds., Geologic and engineering ap-
proaches in evaluation of San Andres/Grayburg hydro-
carbon reservoirs — Permian basin: University of Texas
at Austin, Bureau of Economic Geology, p. 1 – 20.
Given, R. K., and K. C. Lohmann, 1985, Derivation of the
original isotopic composition of Permian marine car-
bonates: Journal of Sedimentary Petrology, v. 55, p. 430–
439.
Hardie, L. A., and P. Garrett, 1977, General environmental
setting, in L. A. Hardie, ed., Sedimentation on the
modern carbonate tidal flats of the northwest Andros
Island, Bahamas: Baltimore, Johns Hopkins University
Press, p. 12 – 50.
Holtz, M. H., and C. M. Garrett, 1990, Geologic and
engineering characterization of Leonardian carbonate
oil reservoirs: A framework for strategic recovery prac-
tices in four oil plays (abs.), in J. E. Flis and R. C. Price,
eds., Permian basin oil and gas fields: Innovative ideas
in exploration and development: West Texas Geolog-
ical Society Publication 90-87, p. 76.
Holtz, M. H., S. C. Ruppel, and C. R. Hocott, 1992, In-
tegrated geologic and engineering determination of
oil-reserve-growth potential in carbonate reservoirs:
Journal of Petroleum Technology, v. 44, no. 11, p. 1250–
1258.
Jones, R. H., and F. J. Lucia, 2004, Integration of rock fabric,
petrophysical class, and stratigraphy for petrophysical
quantification of sequence-stratigraphic framework,
Fullerton field, Texas, in S. C. Ruppel, ed., principal
investigator, Multidisciplinary imaging of rock prop-
erties in carbonate reservoirs for flow unit targeting:
University of Texas at Austin, Bureau of Economic Ge-
ology, final contract report to Department of Energy,
Contract DE-FC26-01BC1535 1, p. 125 – 162.
Jones, R. H., and S. C. Ruppel, 2004, Evidence of post-
Wolfcampian fault movement and its impact on Clear
Fork reservoir quality: Fullerton field, west Texas, in
R. C. Trentham, ed., Banking on the Permian basin:
Plays, field studies, and techniques: West Texas Geo-
logical Society Fall Symposium: West Texas Geological
Society Publication 04-112, p. 207.
Kerans, C., and W. M. Fitchen, 1995, Sequence hierarchy
and facies architecture of a carbonate ramp system: San
Andres Formation of Algerita escarpment and western
Guadalupe Mountains, west Texas and New Mexico:
University of Texas at Austin, Bureau of Economic
Geology, Report of Investigations 235, 86 p.
Kerans, C., and K. Kempter, 2002, Hierarchical stratigraph-
ic analysis of a carbonate platform, Permian of the
Guadalupe Mountains: University of Texas at Austin,
Bureau of Economic Geology (AAPG Datapages Dis-
covery Series 5), CD-ROM.
Kerans, C., and S. C. Ruppel, 1994, San Andres sequence
framework, Guadalupe Mountains: Implications for
San Andres type section and subsurface reservoirs, in
R. A. Garber and D. R. Keller, eds., Field guide to the
Paleozoic section of the San Andres Mountains: Perm-
ian Basin Section SEPM Publication 94-35, p. 105 – 115.
Kerans, C., F. J. Lucia, and R. K. Senger, 1994, Integrated
characterization of carbonate ramp reservoirs using
Permian San Andres Formation outcrop analogs: AAPG
Bulletin, v. 78, no. 2, p. 181 – 216.
Kerans, C., K. Kempter, J. Rush, and W. L. Fisher, 2000,
Facies and stratigraphic controls on a coastal paleo-
karst: Lower Permian, Apache Canyon, west Texas, in
R. Lindsay, R. Trentham, R. F. Ward, and A. H. Smith,
eds., Classic Permian geology of west Texas and
southeastern New Mexico, 75 years of Permian basin
oil and gas exploration and development: West Texas
Geological Society Publication 00-108, p. 55 – 82.
Land, L. S., 1980, The isotopic and trace element geo-
chemistry of dolomite: The state of the art, in D. H.
Zenger, J. B. Dunham, and R. L. Ethington, eds., Con-
cepts and models of dolomitization: SEPM Special Pub-
lication 28, p. 87– 110.
Leary, D. A., 1985, Diagenesis of the Permian (Guadalu-
pian) San Andres and Grayburg formations, Central
Basin platform, Permian basin, west Texas: Master’s
 Key Role of Outcrops and Cores in Carbonate Reservoir Characterization and Modeling / 393
thesis, University of Texas at Austin, Austin, Texas,
125 p.
Leary, D. A., and J. N. Vogt, 1990, Diagenesis of the San
Andres Formation (Guadalupian) reservoirs, Univer-
sity Lands, Central Basin platform, in D. G. Bebout and
P. M. Harris, eds., Geologic and engineering ap-
proaches in evaluation of San Andres/Grayburg hydro-
carbon reservoirs— Permian basin: University of Texas
at Austin, Bureau of Economic Geology, p. 21 – 28.
Lohman, K. C., and J. C. G. Walker, 1989, The d18O record
of Phanerozoic abiotic cements: Geophysical Research
Letters, v. 16, p. 319 – 322.
Longacre, S. A., 1990, The Grayburg reservoir, north Mc-
Elroy unit, Crane County, Texas, in D. G. Bebout and
P. M. Harris, eds., Geologic and engineering ap-
proaches in evaluation of San Andres/Grayburg hy-
drocarbon reservoirs — Permian basin: University of
Texas at Austin, Bureau of Economic Geology, p. 239 –
273.
Loucks, R. G., 1999, Paleocave carbonate reservoirs:
Origins, burial-depth modifications, spatial complex-
ity, and reservoir implications: AAPG Bulletin, v. 83,
p. 1795 – 1834.
Lucia, F. J., 1995, Rock-fabric/petrophysical classification
of carbonate pore space for reservoir characterization:
AAPG Bulletin, v. 79, no. 9, p. 1275 – 1300.
Lucia, F. J., and J. W. Jennings Jr., 2002, Calculation and
distribution of petrophysical properties in the South
Wasson Clear Fork field, in F. J. Lucia, ed., Integrated
outcrop and subsurface studies of the interwell envi-
ronment of carbonate reservoirs: Clear Fork (Leonar-
dian-age) reservoirs, west Texas and New Mexico:
University of Texas at Austin, Bureau of Economic
Geology, final technical report to the Department of
Energy, Contract DE-AC26-98BC15105, p. 95 – 142.
Major, R. P., G. W. Vander Stoep, and M. H. Holtz, 1990,
Delineation of unrecovered mobile oil in a mature do-
lomite reservoir: East Penwell San Andres unit, Univer-
sity Lands, west Texas: University of Texas at Austin,
Bureau of Economic Geology, Report of Investiga-
tions 194, 52 p.
Mazzullo, S. J., 1982, Stratigraphy and depositional mosaics
of lower Clear Fork and Wichita groups (Permian),
northern Midland basin, Texas: AAPG Bulletin, v. 66,
p. 210 – 227.
Mazzullo, S. J., and A. Reid, 1989, Lower Permian platform
and basin depositional systems, northern Midland
basin, Texas, in P. D. Crevello, J. J. Wilson, J. F. Sarg,
and J. F. Read, eds., Controls on carbonate platform
and basin development: SEPM Special Publication 44,
p. 305 – 320.
Presley, M. W., 1987, Evolution of Permian evaporite basin
in Texas Panhandle: AAPG Bulletin, v. 71, p. 167 – 190.
Presley, M. W., and K. A. McGillis, 1982, Coastal evaporite
and tidal-flat sediments of the upper Clear Fork and
Glorieta formations, Texas Panhandle: University of
Texas at Austin, Bureau of Economic Geology, Report
of Investigations 115, 50 p.
Ruppel, S. C., 1992, Styles of deposition and diagenesis in
Leonardian carbonate reservoirs in west Texas: Impli-
cations for improved reservoir characterization: Soci-
ety of Petroleum Engineers Annual Exhibition and
Technical Conference, SPE Paper 24691, p. 313 – 320.
Ruppel, S. C., 2002, Geological controls on reservoir
development in a Leonardian (Lower Permian) car-
bonate platform reservoir, Monahans field, west Texas:
University of Texas at Austin, Bureau of Economic
Geology, Report of Investigations 266, 58 p.
Ruppel, S. C., and E. E. Ariza, 2002, Cycle and sequence
stratigraphy of the clear fork reservoir at South Wasson
field: Gaines County, Texas, in F. J. Lucia, ed., Inte-
grated outcrop and subsurface studies of the interwell
environment of carbonate reservoirs: Clear Fork (Leo-
nardian-age) reservoirs, west Texas and New Mexico:
University of Texas at Austin, Bureau of Economic
Geology, final technical report to the Department of
Energy, Contract DE-AC26-98BC15105, p. 59 – 94.
Ruppel, S. C., and D. G. Bebout, 2001, Competing effects
of depositional architecture and diagenesis on car-
bonate reservoir development: Grayburg Formation,
South Cowden field, west Texas: University of Texas
at Austin, Bureau of Economic Geology, Report of
Investigations 263, 62 p.
Ruppel, S. C., and H. S. Cander, 1988a, Effects of facies and
diagenesis on reservoir heterogeneity: Emma San Andres
field, west Texas: University of Texas at Austin, Bureau
of Economic Geology, Report of Investigations 178,
67 p.
Ruppel, S. C., and H. S. Cander, 1988b, Dolomitization of
shallow-water platform carbonates by sea water and
seawater-derived brines: San Andres Formation (Gua-
dalupian), west Texas, in Sedimentology and geo-
chemistry of dolostones: SEPM Special Publication 43,
p. 245 – 262.
Ruppel, S. C., W. B. Ward, E. E. Ariza, and J. W. Jennings Jr.,
2000, Cycle and sequence stratigraphy of Clear Fork
reservoir-equivalent outcrops: Victorio Peak Forma-
tion, Sierra Diablo, Texas, in R. Lindsay, R. Trentham,
R. F. Ward, and A. H. Smith, eds., Classic Permian
geology of west Texas and southeastern New Mexico,
75 years of Permian basin oil and gas exploration and
development: West Texas Geological Society Publica-
tion 00-108, p. 109 – 130.
Saller, A. H., and N. Henderson, 1998, Distribution of po-
rosity and permeability in platform dolomites: Insights
from the Permian of west Texas: AAPG Bulletin, v. 82,
no. 8, p. 1528 – 1550.
Tyler, N., and N. J. Banta, 1989, Oil and gas resources re-
maining in the Permian basin: Targets for additional
hydrocarbon recovery: University of Texas at Austin,
Bureau of Economic Geology, Geological Circular 89-4,
20 p.
Vogt, J. N., 1986, Dolomitization and anhydrite diagenesis
of the San Andres (Permian) Formation, Gaines
County, Texas: Master’s thesis, University of Texas at
Austin, Austin, Texas, 202 p.
394 / Ruppel and Jones
Wang, F., and F. J. Lucia, 2004, Reservoir modeling and
simulation of Fullerton Clear Fork field, Andrews
County, Texas, in S. C. Ruppel, ed., principal investi-
gator, Multidisciplinary imaging of rock properties in
carbonate reservoirs for flow unit targeting: University
of Texas at Austin, Bureau of Economic Geology, final
contract report to Department of Energy, Contract DE-
FC26-01BC1535 1, p. 219 – 304.
Ye, Q., and S. J. Mazzullo, 1993, Dolomitization of lower
Permian platform facies, Wichita Formation, north plat-
form, Midland basin, Texas: Carbonates and Evapo-
rites v. 8, no. 1, p. 55 – 70.
Zeng, H., 2004, Construction and analysis of 3-D seismic
porosity inversion models, in S. C. Ruppel, ed., prin-
cipal investigator, Multidisciplinary imaging of rock
properties in carbonate reservoirs for flow unit tar-
geting: University of Texas at Austin, Bureau of Eco-
nomic Geology, final contract report to Department
of Energy, Contract DE-FC26-01BC1535 1, p. 305 –
342.
Zeng, H., and C. Kerans, 2003, Seismic frequency control
on carbonate seismic stratigraphy: A case study of the
Kingdom Abo sequence, west Texas: AAPG Bulletin,
v. 87, no. 2, p. 273 – 293.
 11
Weissenberger, J. A. W., R. A. Wierzbicki, and N. J. Harland, 2006, Carbonate
sequence stratigraphy and petroleum geology of the Jurassic deep
Panuke field, offshore Nova Scotia, Canada, in P. M. Harris and L. J.
Weber, eds., Giant hydrocarbon reservoirs of the world: From rocks
to reservoir characterization and modeling: AAPG Memoir 88/SEPM
Special Publication, p. 395 – 431.

Carbonate Sequence Stratigraphy

and Petroleum Geology of the
Jurassic Deep Panuke Field,
Offshore Nova Scotia, Canada
John A. W. Weissenberger
Husky Energy Inc., Calgary, Alberta, Canada

Richard A. Wierzbicki
EnCana Corporation, Calgary, Alberta, Canada

Nancy J. Harland
EnCana Corporation, Calgary, Alberta, Canada

ABSTRACT

P
anCanadian Petroleum (now EnCana Corporation) discovered the deep Pa-
nuke field in 1998 with the drilling of the PP-3C well. The well was drilled in
90 m (295 ft) of water, 250 km (155 mi) southeast of Halifax, Canada. Subsequent
delineation and development drilling has proven a significant gas accumulation.
The gas is trapped, by a combined structural-stratigraphic configuration, in
the Upper Jurassic reefal and oolitic limestones and dolomites of the Abenaki
Formation. The Jurassic carbonate platform on the Scotian Shelf was attached to a
metamorphic hinterland, so that the sediments contain varying amounts of silic-
iclastics. Abundant secondary porosity was encountered, ranging from leached
matrix and intercrystalline to vuggy and/or cavernous. Textural, petrographic,
and isotopic evidences suggest that deep burial and hydrothermal diagenetic pro-
cesses caused the porosity. The gas is believed to have been sourced from adjacent
Verrill Canyon Formation shales, whereas small amounts of hydrogen sulfide have
been isotopically linked to synrift evaporites underlying the Abenaki.
The Abenaki is divided into seven third-order depositional sequences, the
Abenaki V being the primary gas zone. These sequences have been regionally cor-
related using geology and a grid of two-dimensional seismic data. Three-dimensional
seismic data was used in delineation drilling and reservoir characterization. Deep
Panuke is the first, and remains at writing, the only significant hydrocarbon dis-
covery in the Mesozoic carbonates of the continental shelf of eastern North America.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215883M883275

395
396 / Weissenberger et al.
INTRODUCTION
A recent round of exploration drilling for hydro-
carbons on the Scotian Shelf has provided abundant
new data on the Jurassic carbonates and associated
strata of the region. Beginning in 1998, EnCana Cor-
poration, then PanCanadian Petroleum Ltd., initiated
drilling for deep targets below existing production of
the Panuke field, discovered in 1986. The prospective
reservoir was the Jurassic carbonates of the Abenaki
Formation, below oil-bearing siliciclastic sands of the
Cretaceous Panuke Formation. The drilling program
led to the discovery and delineation of the giant deep
Panuke gas field. This chapter proposes new stratigraph-
ic, depositional, and paleogeographic interpretations
of the Nova Scotia Jurassic based on the new seismic
and well data.
Regional Geology
The structural and stratigraphic development of
the Scotian Shelf Jurassic is strongly influenced by its
location on the western edge of the North Atlantic
(Figure 1). The Nova Scotia Jurassic was at the north-
ern end of the Bahama –Grand Bank gigaplatform as
described by Poag (1991). This refers to the carbonate
depositional system that dominated the eastern mar-
gin of North America.
Extensional and transtensional tectonics, associ-
ated with the opening of the Atlantic Ocean, were
dominant during the progressive widening of the
Atlantic, from north to south, beginning in the Late
Triassic (Jansa and Pe-Piper, 1988). The tectonic set-
ting was essentially a Late Triassic to Early Jurassic
rift basin, followed by a passive margin from the Mid-
dle Jurassic through today.
The basement rocks of the Scotian Shelf are of
continental origin, consisting of Devonian granites
and late Precambrian to Ordovician metasediments
(Welsink et al., 1989), which were then intruded by
Devonian-aged granites. Figure 2 shows some of the
main structural elements of the Scotian Shelf, includ-
ing major, fault-bounded structural highs and inter-
vening basinal areas. Three-dimensional (3-D) seismic
data from the deep Panuke area suggest that these
larger structural elements are broken into smaller
horsts and grabens. Jurassic strata are present only in
the outer part of the Scotian Shelf in the form of an
onlapping wedge of sediments (Figure 3).
The gross stratigraphy of this wedge reflects the
tectonic history, from initial rift to essentially passive-
margin sedimentation. The early, basin-fill sediments
are Late Triassic to Early Jurassic (Sinemurian–Pliens-
bachian) in age. These comprise synrift siliciclastics
and evaporites of the Eurydice and Argo formations
(McIver, 1972; Jansa and Wade, 1975). Overlying these
are microcrystalline, restricted-marine, occasionally an-
hydritic dolomites of the Iroquois Formation (McIver,
1972). The Iroquois is overlain by a siliciclastic succes-
sion, comprising red beds, immature sands, and asso-
ciated strata. These are assigned to the Mohican For-
mation (Given, 1977).
Normal-marine carbonate deposition began with
the Abenaki Formation (McIver, 1972; revised by Eliuk,
1978). It has been lithostratigraphically subdivided
into several members: the Scatarie, Misaine, Baccaro,
and Artimon (and the Roseway unit; Wade, 1977). For
further discussion, see Eliuk (1978). However, for the
Misaine Member, which is a marine shale, these com-
prise the bulk of the Jurassic carbonate strata of the
Scotian Shelf. Regionally, these are time equivalent
to shallow-marine and deltaic sediments of the Mic
Mac Formation. The lateral relationships between
these lithofacies belts will be discussed below. The
corresponding Jurassic deep-water shale succession
is termed the Verrill Canyon Formation, thought to
be time equivalent to sediments from the Mohican
Formation to Early Cretaceous strata. This lithostra-
tigraphy (Figure 4) will be discussed from a sequence-
stratigraphic perspective below.
The most comprehensive early work on the Abe-
naki was presented by Eliuk (1978), with later up-
dates (Eliuk et al., 1986; Eliuk and Levesque, 1989).
Other interpretations of the older wells included Jansa
and Wade (1975), Jansa and Wiedmann (1982), and
Ellis (1984). We and our colleagues have presented
preliminary aspects of this project (e.g., Weissenber-
ger et al., 2000; Harland et al., 2002; Hogg and Ena-
chescu, 2003), which is augmented and amended
herein.
Background and Methodology
Prior to the drilling of the PP3C discovery well in
1998, 53 wells had penetrated the Abenaki Formation
on the Scotian Shelf. Only 17 of these wells were cored.
The initial stratigraphic interpretation and explora-
tion program was therefore based on these data; wire-
line logs, cores, and drill cuttings from the aforemen-
tioned wells and a regional grid of two-dimensional
seismic data. A 3-D survey, originally shot to image
shallower, oil-prone Cretaceous strata, covered part of
the Panuke area itself. This was used to pick the
drilling location of the discovery well (PP-3C).
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 397
FIGURE 1. Paleogeography of the Late Jurassic, approximately 150 m.y. Modified from Scotese (1997).
Drilling of the appraisal wells, which followed
PP-3C, provided the additional data presented here-
in. Some additional conventional cores were cut
(see Appendix 1). The distribution of all these data is
shown in Figure 5. Because of the prohibitive costs of
cutting conventional core and the additional risk of
encountering lost circulation while coring, a system
of rotary sidewall coring combined with formation
image logs (and other wire-line logs), was devised to
interpret lithofacies and stratigraphy. Sufficient core
data are now available to confidently construct facies
models and correlate stratigraphy, as described in de-
tail below.
Special analyses to determine diagenetic history,
reservoir, and source rock quality were also under-
taken. These include detailed petrographic studies
of drill cuttings and conventional and sidewall core
samples; fluid-inclusion and stable isotope analyses,
isotopic analyses of gas samples, and geochemical
analyses of hydrocarbon and potential source inter-
vals, from wells that are no longer confidential, will
be described in detail below.
LITHOFACIES AND
DEPOSITIONAL ENVIRONMENTS
Deposition on the Jurassic continental margin of
Nova Scotia comprised a mixed carbonate-siliciclastic
depositional system. The climate was tropical to sub-
tropical (Golonka et al., 1994; Scotese, 1997). Deposi-
tional environments and their component lithofacies
(belts) reflect this overall paleogeographic setting.
A generalized facies model for the Abenaki For-
mation is shown in Figure 6. The depositional envi-
ronments and component lithofacies are described
398 / Weissenberger et al.

FIGURE 2. Tectonic elements of the Scotian Shelf (modified from Welsink et al., 1989).

FIGURE 3. Regional seismic dip line across the Scotian Shelf. Note the wedge-shaped geometry of the Jurassic succession.
Colored lines indicate Abenaki third-order depositional sequences from Abenaki (AB) I at base (red), to AB II (purple),
AB III (brown), AB IV (orange), AB V (yellow), and AB VI/VII (black) at top. Depth is travel time in milliseconds.
FIGURE 4. Stratigraphy of the Scotian Shelf (left) (modified from CNSOPB, 1997), with detailed sequence and chronostratigraphy of the Abenaki Formation and
related strata (right). Chronostratigraphy is from Haq et al. (1987); Abenaki biostratigraphy from van Helden (see Appendix 2).
Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 399
400 / Weissenberger et al.
FIGURE 5. Location map of the deep Panuke pool showing
Abenaki penetrations and core control.
below, from the most landward to the most basin-
ward. More detailed facies analysis in the literature
includes Eliuk (1978), Ellis (1984), Eliuk and Lev-
esque (1989), and Pratt and Jansa (1989). The most
recent summary, with examples from deep Panuke,
was presented by Wierzbicki et al. (2002).
Because a correct facies model was imperative in
building an accurate sequence-stratigraphic reservoir
model, the exploration and development teams re-
viewed numerous potential analogs for deep Panuke.
Outcrop studies from similar basins (e.g., Leinfelder,
FIGURE 6. Facies model for the Abenaki Formation with basic depositional environments (modified from Weissenberger
et al., 2000).
1994; Wenzel and Strasser, 2001; Blomeier and Reij-
mer, 2002, respectively), which were most helpful,
were supplemented with our own field studies in Por-
tugal, the Swiss Jura, and Morocco.
The team was also able to effectively tie formation
image logs to specific lithofacies. This was done by
cutting numerous rotary sidewall cores in the pene-
trated formations. Two complimentary methods were
undertaken: cutting the cores before or after logging.
The former method would pick core points from
sample descriptions. This did not allow precise pick-
ing of sample points, but sample points were clearly
visible on the image logs, giving absolute certainty as
to which log facies was sampled. The latter method
allowed examination of the entire log suite and sub-
sequent sidewall coring of specific formation image
log (Formation MicroImager; FMI) facies. Overall,
more than 200 sidewall core samples were taken in
wells with FMI log suites.
Figure 7 shows the interpreted paleogeography of
the Scotian Shelf during the Late Jurassic. The dis-
tribution of the facies belts is by definition general-
ized, because, as discussed below, facies distribution
changed with changing stands of sea level. That said,
a carbonate shelf margin to shelf interior coexisted
with siliciclastic inner shelf deposits. Siliciclastics
were also passed into the basin at points along the
platform margin, notably the large Sable delta (Mic
Mac Formation) in the northeast part of the study
area. The shelf was generally carbonate dominated to
the southwest and more siliciclastic rich to the north-
east of the delta.
A brief description of the depositional environ-
ments and the constituent lithofacies is given be-
low. Estimates for the relative bathymetries of the
FIGURE 7. Paleogeography of the upper Abenaki Formation, modified from Wierzbicki and Harland (2004).
Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 401
402 / Weissenberger et al.

FIGURE 8. Representative Abenaki lithofacies; (A) burrowed, fine- to medium-grained sandstone with a micritic matrix
(reworked lowstand shelf; I-100 8295.5 ft [2528.4 m]); (B) oncolite packstone (lagoon proximal to margin; K-62, 3384.2 m
[11,103.01 ft]); (C) oolitic grainstone (shoal margin; B-13, 2514 m [8248 ft]; coin is 18 mm [0.8 in.] in diameter).

lithofacies are derived from their position in shoaling- interior can also have a significant siliciclastic com-
upward successions; facies identification in wells that ponent. These are interpreted as deeper lagoon de-
are tied to seismic reflectors traceable from platform to posits, representing as much as 20 m (66 ft) bathym-
basin; and from the literature cited above. etry (the moat of Eliuk, 1978).

Siliciclastic Coastal Plain Platform Interior and Lagoon

Light-gray, very fine to medium quartz sand with
a micritic matrix is found in association with several
other sediment types on the carbonate platform.
These associations are with ooids and oncoids; with
red and green shales and coal; and with oncolite-ooid-
megalodont packstone to wackestone and cryptalgal
mudstones. The sediments vary from pure siliciclastic
to mixed carbonate-siliciclastic deposits depending
on local conditions. The respective subenvironments
of deposition are lowstand sandstones preserved in
lows or karst features on the platform; shallow la-
goon; and nearshore biotopes, deltas, shoreface sands,
and/or estuaries. These lithologies are all interpreted
to have been deposited in less than 10 m (33 ft) of
water. Figure 8A shows a bioturbated sandstone in-
terpreted as lowstand deposits on the shelf, reworked
during the ensuing transgression.
Fossiliferous (gastropod-pelecypod-ostracod) wacke-
stones and dark-gray shales found in the platform
A variety of lithofacies characterize this environ-
ment. Peloidal wackestones to packstones have a mi-
nor coral, sponge, and stromatoporoid component
and are commonly burrowed and stylolitic. These are
interpreted to represent shallow lagoonal deposits
proximal to the margin. Oncolitic facies are also com-
mon in this setting (Figure 8B) and comprise wacke-
stones to packstones containing varying amounts
of reefal or oolitic material, depending on the proxi-
mity to the margin. Argillaceous facies are observed in
the platform interior, dominantly dark-brown skeletal
wackestones, shales, and sands. Oncoids, gastropods,
bivalves, ostracods, and ooids are the most common
clasts. These argillaceous lithologies are interpreted
as deeper shelf deposits (5 – 20 m [16 – 66 ft] water
depth; the moat, Eliuk, 1978, p. 447), deposited during
transgressions or in parts of the shelf that were pe-
riodically not filled in a given fourth- or fifth-order
depositional cycle.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 403

FIGURE 9. Representative Abenaki lithofacies (continued from Figure 8); (A) coral-stromatoporoid packstone to
boundstone (reefal platform margin; G-32, 12,710 ft [3874 m]); (B) skeletal packstone with grainy matrix (proximal
foreslope; G-32, 11,839 ft [3608 m]); (C) thrombolitic mudstone (open marine, basin; G-32, 14,414 ft [4393 m]).

Platform Margin margin or as backreef patch reefs, all deposited in from

This environment comprises two dominant com-
ponents, the skeletal or reefal (coral-stromatoporoid)
and the oolitic. Despite limited core data and rela-
tively sparse well spacing, the platform crest may still
be interpreted to be dominated by oolitic grainstones
(Figure 8C), whereas coral-stromatoporoid bound-
stones and packstones (Figure 9A) largely occur in
somewhat deeper water.
The reefal component comprises coral-stromatoporoid
wackestones to packstones, with common chaetetids,
sponges, bryozoa, pelecypods, gastropods, and cri-
noids. The matrix is typically light brown and grainy.
These lithofacies are interpreted as high-energy de-
posits found anywhere from the immediate backreef
through the reef crest to the immediate forereef. These
would represent water depths from 0 to 10 m (33 ft).
Similar lithologies can be found as much as 50 m (164 ft)
down the foreslope, representing allochthonous ma-
terial shed downslope. Coral-stromatoporoid bound-
stone, with chaetetid and algal components, forms
the other major reefal lithology. Light to medium
brown in color with a grainy skeletal matrix and typ-
ically displaying massive to graded bedding, its minor
components include bryozoa, pelecypods, gastropods,
and crinoids. These may be associated with dark-gray
skeletal wackestones with the same fauna. These li-
thologies represent bound reef proper, either at the
2 to 10 m (6.6 to 33 ft) of water. The darker wacke-
stones are interpreted as shelter cavity deposits be-
cause of the dominantly grainy texture, suggesting
overall high-energy conditions.
Where the margin is dominantly oolitic, the ooids oc-
cur in massive to cross-bedded grainstones (Figure 8C).
These are light brown in color and have minor oc-
currences of oncoids, solitary corals, pelecypods, and
gastropods. The lithology can vary to include ooid
grapestone and oncoid wackestone fabrics. These li-
thologies are interpreted to represent high-energy
shoal deposits, deposited in less than 5 m of water,
alternating with reefal boundstone lithofacies at the
platform margin or near-margin shoals where waves
are predominantly wind generated.
Proximal Foreslope
This environment is characterized by coral-sponge-
demosponge wackestones to packstones with a light-
gray or buff to dark-gray skeletal or micritic matrix
(Figure 9B). Other components include tubiphytes,
lithistid sponges, stromatoporoids, chaetetids, cri-
noids, gastropods, brachiopods, and pelecypods. This
lithofacies may be associated with medium- to dark-
gray, laminated siltstones and the aforementioned,
coarser reef-derived material.
404 / Weissenberger et al.
Grey to brown coral-sponge-demosponge pack-
stones to boundstones also occur in this setting.
They are typically massive to poorly bedded, with an
occasionally silty to sandy, mud-dominated matrix.
Fauna is deeper water scleractinian corals, sponges,
chaetetids, stromatoporoids, and tubiphytes.
The wackestones and packstones are interpreted
as coral-dominated deposits deposited in 10 –70 m
(3.3–230 ft) of water. The packstones to boundstones
are coral-dominated reefs. This facies is the main reefal
facies seen in core and FMI images at deep Panuke. Al-
though the Abenaki margin is relatively steep, abun-
dant reefal development occurred there, and the plat-
form crest is dominated by grainstones.
Distal Foreslope
This environment is dominated by sponge coral
and tubiphytes wackestone, with minor amounts
of stromatoporoid, chaetetid, and ooids. Microbia-
lite textures and a cyanobacterial micritic matrix are
common. Dark-gray to black laminated shales and
medium- to dark-gray siltstones are associated li-
thologies. In addition, lithistid sponge- and coral-
stromatoporoid packstones to boundstones occur in
this setting. These are medium gray in color, with
a micritic to thrombolitic matrix. Tubiphytes, Micro-
selena (and other deep-water scleractinian corals), bryo-
zoans, pelecypods, brachiopods, and crinoids are asso-
ciated components. Light- to medium-gray skeletal
wackestone and gray to black shales are associated
lithologies.
The wackestone and shaly lithofacies are inter-
preted as background sediments, in-situ sponge, and
deep-water coral meadows, deposited in 30 –70 m
(100 – 229 ft) of water. The boundstones are sponge-
dominated mounds, typically forming in oxygenated
waters, 30 –100 m (100 – 330 ft) deep.
Basin
Dark-gray to black shales and mudstones with lam-
inated to thrombolitic textures (Figure 9C), laminated
siltstones, and marls constitute the dominant in-
situ sediment. These are interpreted to represent
deposition in 100 – 200 m (330 – 660 ft) of water.
Sponge material, pelagic foraminifera, radiolaria,
and ammonites are the most common fossils. Hexa-
tinellid sponge-tubiphytes boundstones to pack-
stones, with associated calcareous shales, are also
observed in this setting. Stromatoporoid and coral
wackestones to packstones, with associated sponges,
demosponges, crinoids oolites, and oncolites, also
occur.
The boundstones are interpreted as hexatinellid
sponge-microbialite reefs, the shales, wackestones, and
packstones as interreef material deposited in greater
than 100 m (330 ft) water depth. The stromatoporoid-
coral deposits are shelf derived (turbidites) occurring
in depths from 10 to more than 100 m (33 to more
than 330 ft).
Discussion
There has been disagreement over whether the
Abenaki Formation represents a dominantly reef- or
ramp-style depositional system (Eliuk, 1978; versus
Ellis, 1984; Ellis et al., 1985, 1990, respectively). Our
new drilling, including new core data and FMI logs
obtained in the deep Panuke program, supports a
modified reefal model. Wells such as M-79, which
essentially penetrate the entire Abenaki section, show
thick aggradational packages of coral-stromatoporoid
rudstone to boundstone as stacked third-order se-
quences evidenced on FMI logs and in sidewall cores.
These boundstones and rudstones contain signifi-
cant amounts of micrite in the matrix (which has
been recrystallized; J. Dravis, 2001, personal com-
munication), suggesting that they were not subjected
to continuous wave action; i.e., they were likely de-
posited below fair-weather-wave base.
By contrast, the Abenaki (AB) I (Scatarie Member)
has a lithofacies composition and depositional pro-
file observed on seismic, suggesting more of a ramp,
or open platform setting, with no well-developed reef
margin. Eliuk (1978) interpreted the AB I to have
more of a rimmed shelf depositional style south of the
G-32 well.
Seismic profiles (as discussed below) support the
geological evidence. Dramatic platform-to-basin re-
lief is developed from the AB II upward, with steep
depositional dips, to a maximum of approximately
508, which clearly indicates a steep reef profile. The
uppermost sequence (AB VII) has a gentler deposi-
tional profile, suggesting it is less reefal. Limited side-
wall core and FMI images, along with drill-cutting
examination, show a deep-water sponge, mound-
dominated argillaceous environment. Further, our
observations of facies relationships elsewhere, partic-
ularly of the Kimmeridgian Ota reef in Portugal, sug-
gest that the ooid shoal and reef environments can
be laterally and vertically adjacent, as is seen in the
Abenaki Formation. This, as well as the presence of
coarse winnowed matrix in some coral-stromatoporoid
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 405
boundstone intervals, suggests that the reefs did pe-
riodically grow above fair-weather-wave base, form-
ing high-energy patch reefs in bathymetries similar to
the ooid shoals. They may have formed an armored
reef front at the platform edge but no well has yet
tested this zone.
Near the Abenaki platform margin, southwest of
the Sable delta, siliciclastics are found only as rela-
tively thin, isolated sands. However, core from the
lower Abenaki (sequence I; Scatarie Member) from
the M-79 well has well-rounded gravel- to cobble-
size metasediments in a lime mud matrix, with abun-
dant marine fauna such as crinoids and bivalves.
This suggests that a much more proximal siliciclas-
tic source existed during lower Abenaki deposition,
as no such coarse siliciclastics are known from any
other cores, nor cuttings examined in this study.
Eliuk (1978, p. 435) interpreted the Scatarie to be
absent (because of nondeposition) from several ba-
sin ridges penetrated by at least two wells. These
may have been the source for the coarse siliciclastics
cored in M-79.
As discussed above, FMI images were successfully
matched to lithofacies, particularly through the use of
sidewall coring. Interpretation of some images was
relatively straightforward. Figure 10 shows a cored in-
terval with a large phaceloid coral in growth position
(foreslope environment) from the L-97 well. It also
FIGURE 10. Comparison of
formation image log (FMI) of
a large phaceloid coral (left)
from M-79 (3865 – 3866 m;
12,680 – 12,683 ft) and core of
the same from L-97 (3417 m
[11,210 ft]; coin is 18 mm
[0.8 in.] in diameter). The coral
is in growth position above
(and overlain by) interpreted
stromatoporoid-coral pack-
stone to boundstone.
shows an FMI image from
M-79 with a similar coral,
roughly 1 m (3.3 ft) in height,
also in growth position, in a
succession of nodular fore-
slope mud to wackestones.
Figure 11 compares an FMI
image of a vuggy reefal dolo-
mite from M-79 to core from
the vuggy upper Abenaki res-
ervoir in the H-08 well. The
textural similarities of leached and dolomitized
decimeter-size vuggy rock to the unaltered reefal
boundstone are clear.
SEQUENCE STRATIGRAPHY
The predrill sequence-stratigraphic framework for
deep Panuke primarily reflects drill-cutting descrip-
tions and geophysical log interpretations because
of the paucity of conventional core in the central
Scotian Shelf. Gamma-ray logs commonly showed a
higher response in muddy lithologies, whereas con-
versely, clean gamma readings indicated higher en-
ergy, winnowed sediments. Higher gamma response
was therefore indicative of either muddy foreslope
or platform interior deposits. Drill cuttings or side-
wall cores were typically used to distinguish between
these environments.
Three basic platformal environments were distin-
guished using older cuttings descriptions, i.e., basic
descriptions of lithology and fossil content: dark lime-
stone representing the deep-water and foreslope;
light-colored limestones representing shallow-water
carbonates (typically reef and shoal); and calcareous
sandstones and sandy and silty carbonates repre-
senting reworked siliciclastic coastal-plain environ-
ments. Alternation of these environments allowed
406 / Weissenberger et al.
FIGURE 11. Comparison of formation image log (FMI) of a vuggy reefal boundstone from the M-79 well (left) to core of
the same lithofacies from PI-1A (4029.4 m; 13,219.8 ft).
the identification of the basic transgressive and re-
gressive trends in any given well.
On the dominantly carbonate part of the shelf,
southwest of the Sable delta, siliciclastics were found
to be concentrated in fairly thin, discrete beds. Con-
versely, in the region approaching Sable and farther
northeast on the Banquereau Bank, the dominantly
siliciclastic section was broken by only a few carbon-
ate intervals. Based on the relative landward position
of the siliciclastic and the seaward position of the
carbonate environments in the depositional model,
the dominance of siliciclastics or carbonates was in-
terpreted to represent relative falling and rising sea
level, respectively.
Stratigraphic Architecture of the
Second-order Sequence
A schematic illustration of the depositional se-
quence on the Scotian Shelf near deep Panuke is
shown in Figure 12. The gross stratigraphic pattern
shows the Abenaki platform initiated by a relative
transgression above the widespread Mohican Forma-
tion siliciclastic shelf. Following relative aggradation
in the deep Panuke area, carbonate deposition be-
came more aerially restricted and replaced by argilla-
ceous foreslope and basinal deposits (e.g., at the G-32
well), suggesting that the Abenaki shelf was initially
terminated by transgression before it was inundated
by Lower Cretaceous siliciclastics. This stratigraphy
is herein interpreted as a second-order depositional
sequence.
Jansa (1993) discussed the possible causes for the
termination of the Abenaki platform, favoring con-
tamination by siliciclastics. This cannot be discounted,
but, as discussed below, the carbonate platform co-
existed with siliciclastics throughout the Jurassic. This
suggests that the ultimate termination of upper Abe-
naki deposition in the deep Panuke area likely had
other causes, such as relative sea level change.
Higher order sequences, herein interpreted as third-
order depositional sequences, were initially defined
by the presence of thin siliciclastic sands and silts at
the sequence boundaries and corresponding deepen-
ings in the middle of the sequences, represented by
deeper water carbonates. The latter were interpreted
as the transgressive maximums of the sequences. These
surfaces were initially correlated in the area illustrated
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 407

FIGURE 12. Schematic block diagram of the Jurassic carbonate platform, deep Panuke area. Maximum flooding surface
of the AB V indicated by an asterisk. Time-equivalent basinal (Verrill Canyon Formation) sediments are not shown. For
horizontal scale, refer to Figure 7.

in Figure 12 and later extended along the entire Sco-
tian Shelf. In the region north of the Cohasset L-97
well, the upper Abenaki is dominated by siliciclastics.
There, the sequences were correlated primarily using
the flooding surfaces, represented by extensive corre-
latable limestone beds in the siliciclastic succession.
New well data largely supported the sequences de-
fined before PP-3C was drilled. Flooding surfaces (and
sequence boundaries) were better constrained by the
better log data in the new wells tied to sidewall cores,
as well as detailed drill-cuttings descriptions. Inter-
pretation of these new data led to minor adjustments
of the sequence picks in the regional data set and the
definition of an additional third-order sequence (at
the top of the Abenaki) in the deep Panuke area, the
AB VII.
Third-order Sequences
The nature of third-order sequences in the defined
Abenaki second-order depositional sequence was in-
fluenced by their position in the larger sequence. The
AB I, which generally corresponds to the Scatarie
Member, is a broad carbonate ramp. It typically has
three recognizable shoaling-upward cycles, compris-
ing outer-ramp fossil wackestones to packstones and
inner-ramp oolitic and oncolitic facies; identifiable
by cleaning-upward profiles on the gamma-ray log.
These shoaling patterns are also apparent in the
cuttings descriptions, as in the M-79 well. In the deep
Panuke area, the AB I is roughly 140 m (459 ft) thick.
The AB I is interpreted as the early transgressive
part of the second-order sequence. The base of the
AB I should likely be placed in the upper part of the
Mohican siliciclastics, but lack of deep well penetra-
tions precludes the identification of this surface. Sim-
ilarly, part of the Mohican siliciclastic shelf would
correspond to the lowstand of the second-order se-
quence, but (again) the lack of deep penetrations and
reduced seismic resolution at depth hampers the in-
terpretation of these strata.
The AB II comprises two readily identifiable com-
ponents, the regionally extensive Misaine Forma-
tion shale and the progradational reefal platform
408 / Weissenberger et al.
above. The base of the sequence is taken at the top of
the AB I carbonate ramp. If a lowstand systems tract is
present, it may consist of shelf-interior siliciclastics by-
passed into the basin. These have only been pro-
visionally identified on seismic data and not pene-
trated by wells.
The aforementioned, recognizable components
of the AB II are interpreted as the transgressive
and highstand parts of the sequence, respectively.
The Misaine represents a major transgression, with
basinal shale mapped seismically at least 15 km
(9 mi) landward of the underlying AB I shelf mar-
gin (G. Syhlonyk, 2005, personal communication).
It is on the order of 100 – 150 m (330 – 492 ft) thick
near the shelf margin. The furthest landward extent
of the Misaine is also interpreted as the maximum
flooding surface (MFS) of the Abenaki second-order
sequence. This transgression is recognized along
the North Atlantic continental margins (Jansa and
Wiedmann, 1982) to the southeast margin of Tethys
(Gradstein et al., 1999).
The AB II highstand comprises approximately 220 m
(721 ft) of relatively clean carbonates that prograde
over the Misaine. Seismic data shows downlap of intra-
highstand reflectors onto the Misaine (discussed and
illustrated below in Figure 13). Cores and drill cut-
tings show the AB II to consist of skeletal wackestones
immediately above the Misaine, with a gamma-ray
reading transitional with the underlying shales. The
remainder consists of oolitic packstones to grainstones
or reefal boundstones. The AB II platform margin is
typically set back from that of the AB I.
The AB III is the thickest of the interpreted third-
order sequences. It ranges from 400 to 450 m (1300
to 1470 ft) thick in the deep Panuke area. Formation
MicroImager, sidewall core, and cuttings data indicate
that it consists of coral-stromatoporoid packstones
to boundstones, i.e., reefal strata. This composition,
reflected also in a moderate gamma-ray signature, is
fairly consistent through the sequence, suggesting
an aggradational stacking pattern.
The northeast part of the schematic (Figure 12)
shows the first appearance of abundant siliciclastics
in the lower part of the AB IV, between the L-30 and
L-57 wells. This is interpreted as the lowstand and
early transgressive part of the sequence. The MFS is
interpreted to be represented by a distinctive clean
carbonate that can be correlated from L-97 to L-30.
Southwest of L-97, the AB IV is a platformal carbonate
as much as 150 m (492 ft) thick with a slightly greater
gamma response overall than the AB III. A higher gam-
ma response in the middle to upper part of the se-
quence is observed in three of the four penetrations of
the sequence in the pool itself. This is interpreted to
reflect the same deepening observed in the siliciclas-
tic realm to the northeast.
The AB V has a moderate gamma-ray signature,
similar to the AB III and IV. It is as much as 150 m
(492 ft) thick in the deep Panuke area. The lithofacies
at the base of the sequence are muddier and have a
higher gamma response, with a forereef macrofauna.
This is interpreted as the transgressive part of the
sequence. A lowstand systems tract for the sequence
is interpreted only from seismic data. Basinal mud-
stones and shales attributed to the AB V were pene-
trated in the M-88 well, but these are interpreted as
transgressive and highstand deposits. The upper AB V
consists of coral stromatoporoid packstones to bound-
stones, representing a reef margin environment.
Figure 12 again shows a correlatable carbonate
bed in the lower and middle part of the sequence
extending northeast of L-97. This is interpreted as
representing the MFS of the AB V, which temporar-
ily interrupted siliciclastic deposition.
The AB VI consists of 150 – 200 m (492 – 660 ft) of
calcareous mudstones and shales with a minor silic-
iclastic component. They display a moderate to high
gamma-ray response in the deep Panuke area. These
are interpreted to represent foreslope and basinal de-
posits derived from a backstepped carbonate plat-
form. Skeletal-peloidal packstones encountered in
the upper AB VI in the F-09 well are interpreted to be
part of this retreated carbonate platform. Approxi-
mately 150 m (492 ft) of basinal calcareous shale are
interpreted as equivalent to the AB VI in the M-88 well.
Wells penetrating the AB VI on the platform northeast
of L-97 encountered an entirely siliciclastic section.
The AB VII is a thin (roughly 50-m [164-ft]-thick)
interval of shales and mudstones with a high gamma-
ray response. Where carbonate-rich intervals have
been encountered in this interval (e.g., G-32 and D-42)
they consist of sponge-rich, argillaceous limestones.
These are interpreted to represent deeper foreslope to
basinal environments (the Artimon Member of Eliuk,
1978). A basinally restricted, 20-m (66-ft) interval of
sandstone encountered in M-88 is interpreted as the
lowstand systems tract of the AB VII.
Regional correlations undertaken in this study sug-
gest the AB VII at deep Panuke correlates with skeletal
and oolitic grainstone shoals drilled in the upper-
most Abenaki in the southwest part of the Scotian
Shelf, e.g., in the Acadia K-62 and Bonnet P-23 wells.
These have been termed the Roseway unit (Wade,
1977) and are of Lower Cretaceous Neocomian age.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 409

FIGURE 13. Two dip-oriented seismic lines extracted from 3-D data in the deep Panuke area. (A) Line 420, approximately
half-way between H-08 and the older well G-32. It shows a strong peak (green) picked as the top of the Abenaki I; the
trough above (blue) represents the Misaine Formation shale. It shows an offlapping Abenaki II succession, the top-lap
geometry suggesting a sequence boundary at the top of the unit. (B) On trace 3239, the Abenaki III displays an offlapping
geometry. A fairly strong trough-peak defines the top of the Abenaki II. Top-lap geometries again support the sequence
boundary at the top of the unit.
410 / Weissenberger et al.

FIGURE 14. Idealized fourth-order depositional sequence, Abenaki Formation. Constituent systems tracts and lithofacies are indicated (modified from Wierzbicki
The AB VII also has a shalier, deep-water
character northeast of L-97, but appears
less calcareous than in the south, suggest-
ing persistent proximity to siliciclastic
input.
An idealized fourth-order depositional
sequence of the Abenaki Formation,
shown in Figure 14, illustrates the approx-
imate thickness and facies distribution.
It assumes some subaerial exposure and
epikarstic dissolution at the basal se-
quence boundary. In this model, sili-
ciclastics were deposited on the shelf and
infilled dissolution pipes during relative
lowstand of sea level. Where favorable
conditions existed, a lowstand reef, limited
in lateral and vertical extent, would grow
on the preexisting slope. During the
transgressive phase, normal-marine car-
bonate deposition resumed on the carbon-
ate platform. The shelf-interior silic-
iclastics would be reworked during the
transgression and commonly preserved
only in topographic lows. A varied suite
of depositional environments as discussed
above would develop, including rela-
tively thick beds of basinal mudstones
and shale. The highstand would develop a
similar suite of facies, shifted somewhat
basinward of the underlying transgressive
deposits.
Figure 15 is a stratigraphic cross sec-
tion correlating the fourth-order parase-
quences and lithofacies of the AB V in
the deep Panuke area. The fourth-order
boundaries (AB Va–Vg) were defined dur-
ing the delineation and predevelopment
phase of the project (see below) based on
log signature and detailed drill-cuttings
descriptions (strip logs in Figure 15).
Immediately evident is the dominance
of reefal lithofacies in most of the se-
quence, except for the two wells drilled
most landward, B-90 and F-09, which are
dominated by oolitic shoal and lagoonal
environments. The reefal intervals dis-
and Harland, 2004).

play well-developed cyclicity from fore-

slope to reef, occasionally shoaling into
oolitic facies (e.g., top AB V in M-79A).
The downward dip of the fourth-order
boundaries from the landward toward the
margin wells (e.g., PI-1B, M-79) suggests
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 411
ongoing subsidence at the margin during AB V
deposition. The margin wells also show significantly
more foreslope lithologies in the lower AB VI, suggest-
ing incipient drowning of the platform margin.
Seismic Stratigraphy
Predrill geological interpretation of well data estab-
lished the basic sequence-stratigraphic framework,
tied to seismic data prior to drilling the deep Panuke
discovery. This was followed, after completion of the
PP-3C well, by a regional seismic-stratigraphic study.
The objective of this work was to determine whether
the interpreted geological sequences could be corre-
lated on seismic data and, further, whether seismic-
stratigraphic criteria confirmed their interpretation as
(third-order) depositional sequences.
Figure 13 illustrates two arbitrary seismic lines
from 3-D seismic data in the deep Panuke area. They
show several aspects of seismic geometry that substan-
tiate the geologically interpreted sequence boundaries.
Figure 13A is a line (420) from the Panuke 3-D
survey. Reflectors above the Misaine are clearly in-
clined, basinward dipping, suggesting progradation
of AB II carbonate over and downlap onto the un-
derlying shale. This is consistent with expected seis-
mic geometries above an MFS; the Misaine is the
interpreted MFS of the second-order depositional se-
quence. Reflectors higher in the sequence display
flat-lying top sets on the interpreted platform, a dis-
cernable change in slope and dipping foresets at the
basin margin. Finally, the most basinward-dipping
reflectors of the AB II do not appear to have any
time-equivalent top sets on the shelf. This offlapping
seismic geometry suggests an accommodation min-
imum at the interpreted top AB II surface. This sur-
face was tied to the geological sequence boundary
interpreted from regional well control and regionally
correlated on the seismic grid.
Figure 13B is a dip line (trace 3239) extracted
from the Musquodoboit 3-D survey. It shows similar
seismic geometries in the AB III to those seen in the
AB II on line 420. Reflectors in the lower AB III can
be traced from the platform interior, across the break
in slope at the margin, into the basinward-dipping
foreslope. Higher reflectors in the upper AB III are
inclined near the margin and can be traced upward
(landward) where they appear to terminate against
flat reflectors of the overlying sequence. This off-
lapping geometry suggests a lack of accommoda-
tion on the platform at the interpreted sequence
boundary capping the AB III. This surface was also
first interpreted from well data and correlated to
seismic data.
The two lines also show other noteworthy strati-
graphic relationships. The overall aggradational na-
ture of the carbonate platform is evident on line 420,
although the break in slope of each third-order
margin shifts from the AB II to IV. The AB V margin
is clearly stepped back several hundreds of meters
from that of the AB IV. The lower Abenaki is more
difficult to interpret in trace 3239 because con-
tinued basin subsidence causes significant basin-
ward dips in the reflectors. The AB I margin likely
is somewhat more than 2 km (1.2 mi) basinward of
the younger margins. A strong, steeply dipping peak
develops in the foreslope, regionally traceable to the
top of the AB III. We interpret this as a submarine
hardground representing a depositional hiatus in the
basin during the post-AB III lowstand. Basinal re-
flectors onlap this surface. The upper Abenaki plat-
form margin is slightly stepped back from the AB II
margin.
Similar seismic geometries were observed at the
other sequence boundaries first interpreted using
geological (well) data. This provided some confidence
that regionally correlatable chronostratigraphic units
could be identified and form a reliable stratigraphic
framework for exploration and development.
Discussion
Regional seismic data (Figure 3) suggests that the
Abenaki is a depositional wedge on the Scotian Shelf.
This line suggests increasing coastal onlap in the up-
per Abenaki (AB V –VII). Relative sea level data that
incorporate other regions (Figure 4) suggest, overall, a
relatively constant onlap. Decreasing onlap in the Mic
Mac and lower Missisauga formations, along with the
change from carbonate to dominantly siliciclastic
deposition, is the main criteria for picking a second-
order sequence boundary above the Abenaki.
It is important to note that Figure 12 depicts the
dominantly aggradational nature of the shelf in the
deep Panuke area. The time-equivalent shales and
basinal mudstones of the Verrill Canyon Formation
are not shown. When the pool was discovered, po-
tential reservoir bodies were thought to include the
high-energy platform margin, pinnacle, or possibly
atoll reefs, and siliciclastic sands bypassed into the
basin during relative sea level lowstand.
Figure 16 shows two isochron maps of the platform
margin south of deep Panuke. They display several
noteworthy features of the platform margin in that
412 / Weissenberger et al.

FIGURE 15. Stratigraphic cross section of the upper Abenaki Formation, deep Panuke area (wells are indicated in Figure 5).
Third-order depositional sequences and constituent fourth-order sequences (parasequence sets) are indicated.
Modified from Wierzbicki and Harland (2004).

part of the shelf. The isochrons at the shelf margin of
the AB I are more widely spaced than at the platform
margin of the AB III, suggesting a gentler, ramp-style
depositional profile. The configuration of the margin
in the two different sequences is quite similar. For
example, the embayments just southwest of the H-08
and G-32 are prominent in both time slices, as is the
westward turn of the margin at the southern end of the
survey. This partly reflects the aggradational nature of
the platform in this region, but also the inheritance of
paleogeography with time, possibly because of the per-
sistent influence of underlying structural elements.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 413

We interpret the position of embayments and prom-
ontories on the platform margin to largely be con-
trolled by the configuration of the underlying (syn-
rift) horst and graben network.
The designation of the Abenaki depositional se-
quences as second and third order is done in broad
adherence to the scheme of Vail et al. (1977), where
second-order sequences have a duration of 10 –
100 m.y. and third-order sequences are 1–10 m.y. The
Abenaki sequences also have seismic-stratigraphic
characteristics (as described above), supporting their
designation as depositional sequences. The eustatic
414 / Weissenberger et al.

FIGURE 16. Time structure on Abenaki Formation horizons, deep Panuke, and southwest. (A) Top Abenaki sequence I
time structure. Relatively widely spaced isochrons indicate ramp profile at margin. (B) Top Abenaki III time structure.
Tighter isochrons near the margin indicate a steeper depositional profile. A discernable similarity exists in the position of
the platform margin from the AB I to the AB III (e.g., the embayment near G-32 and south of H-08), showing an
inheritance of paleogeography through the section, likely reflecting deeper structural control.

sea level curve of Haq et al. (1987; partly reproduced
in Figure 4) defines six supercycles (19 third-order
sequences) throughout a similar time span. Therefore,
the Abenaki sequences may more prudently have been
termed ‘‘composite sequences,’’ or supercycles, but
their ultimate importance to the exploration and
development program at deep Panuke was their util-
ity in defining correlatable time lines across the Sco-
tian Shelf and within the pool.
The sequence-stratigraphic model postulates ex-
posure of the platform at the third-order sequence
boundaries. None of these boundaries are cored, so
other data (e.g., FMI images and internal facies archi-
tecture) are used to define the third-order surfaces.
Eliuk (1978) proposed periodic exposure of the
platform, with the development of freshwater lenses,
citing evidence of karst from the G-32 well (which he
linked to dolomitization).
Three-dimensional seismic slices of the carbonate
platform in the deep Panuke area commonly show
circular depressions, 100 – 700 m (330 –2300 ft) in
diameter, which have been interpreted as karst
sinkholes (G. Syhlonyk, 2005, personal commu-
nication). We have examined the similar aged Can-
deiros formation carbonates in central Portugal. These
have well-developed karst sinkholes, tens of meters
deep and wide, as well as bedding-parallel dissolu-
tion (paleowater table?). These data support the
model of periodic subaerial exposure of the Abenaki
platform.
The nature of the alternation between siliciclastic-
and carbonate-dominated deposition on the Scotian
Shelf is seen in two cored wells to the northeast
of deep Panuke: Peskowesk A-99 and Citadel H-52
(Figure 17). The former displays a change from dom-
inantly carbonate to siliciclastic deposition. The base
of the core consists of a dark-gray oolitic packstone
with crinoids and shell hash, representing a middle
foreslope setting. Above are fissile calcareous shales
becoming nodular and fossiliferous toward the top
(crinoids and corals), representing a transition from
basin to lower foreslope environments. The top of the
latter unit is extensively burrowed and mineralized,
as is the top of the overlying thin crinoid-bivalve
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 415

FIGURE 17. Mixed carbonate-siliciclastic core from two wells northeast of deep Panuke. Cores show switching from one
dominant sediment type to another after major depositional hiati.
416 / Weissenberger et al.
wackestone. This succession is interpreted as trans-
gressive, with the borrowed and bored contacts at the
tops of the two shoaling-upward cycles representing
significant depositional hiati at two marine flooding
surfaces. The base of the overlying succession is a
black, fissile, occasionally sideritic shale. This is over-
lain by a thin, fine-grained sandstone, followed by shales,
silts, and finally, fine- to medium-grained sand. This
is interpreted as a basinal to lower shoreface-offshore
transition. This is sharply overlain by planar to tabular,
parallel-laminated, medium- to coarse-grained sand-
stone with common Skolithos, interpreted as upper
shoreface deposits above a sequence boundary.
The succession in Citadel H-52 begins with silic-
iclastics, shales, and silts passing upward into fine- to
medium-grained, parallel-laminated sandstone, with
a burrowed upper surface. Above is a dark-gray to
black, very fine silt to sandstone, with Planolites and
Terebellina. The top of the unit is irregular and miner-
alized. This succession is interpreted as an offshore
transition to middle (and finally, lower) shoreface
setting with a major submarine hardground at the
top. Above the hardground is a fossil lag with abraded
coral fragments, and bivalve fragments, overlain by
fissile to nodular calcareous shale. This is, in turn,
overlain by a thick branching coral-abraded stro-
matoporoid wackestone, with crinoids, bivalves, and
sponges, capped by a massive, medium gray-brown,
peloidal wackestone to packstone, with crinoids and
bivalve fragments. The youngest interval is inter-
preted to represent basinal to middle-foreslope en-
vironments, initiated by an initial marine flooding
event.
These two cores demonstrate that carbonate and
siliciclastic environments coexisted on the Abenaki
shelf. However, although there are certainly mixed
lithologies (e.g., ooids with quartz nuclei, calcareous
sandstones), there is commonly a switching of dom-
inant sediment type at a major event, such as a depo-
sitional hiatus, significant transgression, or drop in
sea level. Prather (1991) describes in detail similar
mixed carbonate-siliciclastic environments from the
Jurassic of the Baltimore Canyon.
Such a surface is interpreted at a fourth-order cycle
(likely parasequence set, sensu Van Wagoner et al.,
1988) from the M-79 well, illustrated in Figure 18.
An irregular surface, observed at the top of a coral-
stromatoporoid packstone, is interpreted to represent
a platform margin environment. This is abruptly over-
lain by 25 cm (10 in.) of fine-grained quartz sand-
stone, interpreted to have been deposited during ex-
posure of the platform at a sea level lowstand and
reworked during the subsequent transgression. This is
overlain by 30 cm (12 in.) of oolitic grainstone coral-
stromatoporoid-sponge packstone, reflecting trans-
gression above the boundary.
DIAGENESIS
Understanding diagenetic processes is critical for
the correct evaluation of the deep Panuke reservoir.
Examination of drill cuttings from PP-3C, immedi-
ately above the cavernous porosity zone, revealed the
first Abenaki dolomite in coral-stromatoporoid pack-
stones, suggesting that dolomitization may have
been involved in vuggy porosity creation. However,
logs from PP-3C indicated that the cavernous po-
rosity was mostly in limestone. Detailed petrogra-
phy, isotope, and fluid-inclusion analyses were
conducted on the subsequent wells and presented
by R. A. Wierzbicki, J. J. Dravis, I. Al-Aasm, and N. J.
Harland (2005, personal communication). This work
and some additional data are discussed below.
Early diagenesis is characterized by grain suturing
in grainstones and grainy packstones, suggesting lit-
tle cementation before compaction. Stylolites and
wispy microstylolites are common. However, abun-
dant secondary limestone porosity is preserved in the
Abenaki at relatively deep burial depths. This poros-
ity was observed to cut stylolites. The porosity is pre-
served along stylolites, and stable skeletal grains (e.g.,
echinoids, calcispheres, corals, stromatoporoids, etc.)
are partially dissolved, and calcite cement in healed
fractures does not extend into adjacent secondary mi-
croporosity. These fabrics suggest the secondary lime-
stone porosity formed after pressure solution and at
least some of the fracturing. The observed porosity was
dominantly vuggy and moldic, the latter typically
reflecting dissolution of micritic peloids or micritized
skeletal grains.
Dolomites are dominantly fine to medium crys-
talline, nonferroan, and subhedral to anhedral in
shape. This dolomite appears to be mostly fabric selec-
tive, replacing grainy limestone textures (packstones-
grainstones) proximal to the platform margin. Sad-
dle dolomite was observed as a replacive mineral and
pore-filling cement. The matrix dolomites overlie
pressure-solution seams and massively replace relict
sutured grain fabrics. Porosity in the dolomite section
is vuggy and intercrystalline. Vuggy porosity cross-
cuts pressure solution seams in dolomites, with many
pores appearing to be molds or partial molds. Saddle
dolomites and dolomitized echinoids were also dis-
solved. This suggests that the pores may have formed
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 417

FIGURE 18. Formation image log (FMI) of a fourth-order sequence boundary from the Abenaki V of the M-79 well.
Irregular surface capping grainy coral-stromatoporoid packstone (3884.9 m; 12,745.7 ft) is overlain by sandstone lag,
oolitic grainstone, and finally, argillaceous coral-stromatoporoid-sponge packstone.

as a result of the dissolution of dolomitized relict fractures and swarms of short macrofractures), whereas
grains or cement, instead of relict limestone. horizontal natural and crosscutting fractures sug-
Fracturing is common in the Abenaki, normal ver- gest active wrench faulting. Vertical fractures com-
tical burial and tectonic fractures (including micro- monly cut stylolites, suggesting their burial origin.
418 / Weissenberger et al.

FIGURE 19. Thin-section photomicrographs of Abenaki Formation reservoir, porosity, and diagenetic fabrics. (A) Secondary
vuggy porosity associated with emplacement of replacement dolomite, as commonly observed in the Abenaki dolomitic
limestones (H-08; 3448.6 m [11,314.3 ft]; plane polarized light, 5.5 mm [0.2 in.] field of view); (B) Well-developed
micromoldic porosity in ooids and peloids, interpreted to result from deep burial conditions (F-09, 3520 m [11,548 ft];
diffused plane polarized light, 3.0 mm [0.12 in.] field of view); (C) Partial dissolution of saddle dolomite (replacing
recrystallized limestone, with partial occlusion of porosity in saddles by nonferroan calcite (arrows). Such partial dis-
solution of dolomites is common in Abenaki dolomitic limestones (H-08, 3448.6 m [11,314.3 ft]; plane polarized light,
3.0 mm [0.12 in.] field of view); (D) A curved (saddle) dolomite rhomb adjacent to a pinpoint vug (PI-1A, 4029.5 m
[13,220.1 ft], 3.5-mm [0.13-in.]-wide field of view). Original descriptions by J. Dravis.

Dissolution of dolomite crystals along fracture planes
suggests that the fractures delivered diagenetic fluids.
Well tests from deep Panuke suggest an effective frac-
ture network, with fractured dolomites in the AB V as
much as 1000 m (3300 ft) wide.
Figure 19 illustrates some of the key diagenetic
textures of the Abenaki in a series of photomicro-
graphs. The samples show evidence for secondary po-
rosity development related to partial dolomitization,
small moldic porosity in limestones, and saddle do-
lomites (19A, B, and C/D, respectively).
Geochemical analyses (R. A. Wierzbicki, J. J. Dravis,
I. Al-Aasm, and N. J. Harland, 2005, personal commu-
nication) of the dolomites revealed carbon isotopes,
which were not depleted, and oxygen isotopes that
were relatively highly depleted. This suggests a lack of
organic influence on dolomitization and recrystalli-
zation and cementation at depth caused by higher
burial temperatures. Matrix dolomites were enriched
in Sr87, whereas saddle dolomites were nonradiogenic,
implying mixing of a radiogenic (juvenile) and non-
radiogenic (basinal) source. Fluid inclusions in cal-
cite cements showed homogenization temperatures
ranging from 60 to 1818C, with salinities of less than
0 to 13.9 wt.% NaCl, i.e., saline to nonsaline fluids
acting in a warm to hot regime. Dolomite cement
(saddles and others) inclusions reveal temperatures
from 85 to 1478C, with salinities from 7.6 to 12.0 wt.%
NaCl. This suggests that warm to hot, moderately
saline waters were involved in dolomitization.
Evidence for faulting during diagenesis includes
vertical stylolites, natural horizontal fractures, and
twinned calcite. Helium is also present (0.02–0.03%)
in the deep Panuke gas, and minor sphalerite is observed
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 419
FIGURE 20. Map of limestone and dolomite porosity trends in the central part of the deep Panuke field. Interpretation
made from amplitudes extracted from the 3-D seismic volume.
in the reservoir. As mentioned above, the presence of
juvenile fluids is suggested by the strontium isotope
data. Vitrinite reflectance data from the M-79 well
also showed a marked increase from 0.8 to 1.10 at
about 4260-m (13,976-ft) drill depth, suggesting a
thermal anomaly. Regionally, evidence for the pres-
ence of hot, fault-borne fluids includes zebra dolo-
mite textures in the core of the Iroquois Formation
from the P-23 well.
These data suggest that deep burial and hydro-
thermal fluids were responsible for limestone and
dolomite porosity at deep Panuke. High-porosity
vuggy limestones in the PP-3C and H-08 wells occur
as linear trends behind the platform margin, ap-
parently influenced by wrench faulting. Vuggy do-
lomites occur as curvilinear trends subparallel to the
platform margin (Figure 20). Figure 21 shows the
stratigraphic distribution of these same porous lime-
stone and dolomite intervals.
PETROLEUM GEOLOGY
The PP-3C discovery well was drilled on the basis
of the aforementioned predrill stratigraphic model;
the location picked from a 1990 3-D seismic survey. An
arbitrary dip line from that survey, near the discovery
location, is shown in Figure 22A, with a schematic in-
terpretation in Figure 22B. The apparent topography
in the basement is interpreted as a basement horst
and associated half graben to the west, the latter
filled by synrift deposits. These are overlain by Mohi-
can Formation siliciclastics. Three basic seismic facies
are discernable from the data: stratified platform-
interior (lagoon), chaotic reflectors at the platform
margin, and basinward-dipping reflectors in the
foreslope.
The chaotic margin facies is between 500 and
1200 m (1640 and 3900 ft) wide in a dip sense. Seismic
modeling indicated that some of the reflectors in the
 420 / Weissenberger et al.

FIGURE 21. Cross section of the deep Panuke wells, showing main occurrences of vuggy limestone and dolomite porosity. Color legend: light blue is porous
limestone; purple is porous dolomite; dark blue is tight limestone; orange stippling reflects no significant vuggy porosity. Diagram is modified from Wierzbicki
and Harland (2004).
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 421

FIGURE 22. Arbitrary dip line from the Panuke 3-D seismic survey (A) and schematic interpretation of the same,
illustrating the end of the Abenaki sequence V (B). Interpretation after G. Syhlonyk.
422 / Weissenberger et al.

margin facies represent a porosity anomaly. This is

corroborated by a velocity sag in reflectors below
the margin, representing the effect of slower velocity
through the overlying porosity. Bright reflectors in
the adjacent foreslope were (later) interpreted as sands,
which had bypassed the platform during sea level
lowstands.

Exploration Drilling Phase

The PP-3C well was intended to test such a
velocity anomaly. Beneath the Cretaceous (Mic
Mac–Missisauga Formation) siliciclastics, the well-
encountered gas shows in calcareous shales and mud-
stones of the upper Abenaki. Upon penetrating less
than 20 m (66 ft) of clean limestone, a lost circu-
lation zone occurred at 3934 m (12,906 ft) drill depth.
Although only 0.1 m (0.3 ft) of porosity had been
drilled, the formation continued to take drilling fluid,
including two plugs of lost circulation material (LCM).
Gas and condensate were flared. Introduction of a
third LCM plug sealed off the formation, and the
well was suspended.
Given the likelihood of penetrating further cav-
ernous porosity, it was decided to drill out using the
annular velocity control (AVC) drilling method, which
had not previously been attempted on the Scotian
Shelf. This entailed pumping sufficient sea water
into the wellbore while drilling to control the for-
mation. No drill cuttings can be recovered using
this method. Once drilling recommenced, several bit
drops of decimeters to more than 1 m (3.3 ft) were
encountered, suggesting further significant porosi-
ty. More than 200 m (660 ft) of Abenaki carbonate
was drilled in this manner, to a final total depth of
4163 m (13,658 ft).
Logging revealed a porosity zone 44 m (144 ft)
thick in the AB V, which was interpreted from logs to
represent predominantly vuggy dolomitic limestone
porosity. As mentioned above, dolomite had been
observed in the drill cuttings immediately above the
lost circulation zone (a coral-stromatoporoid pack-
stone; Weissenberger et al., 2000). The variability of
the porosity seen on logs (Figure 23) was interpreted
to reflect a complex pore system, likely including
vugs, caverns, and matrix.
Testing yielded a maximum flow rate of 69 mmcf/
day, limited by the equipment employed. Wet gas
was produced, with 600 ppm of hydrogen sulfide
(H2S). Also observed was a relatively slow buildup that
may reflect formation damage and possible reser- FIGURE 23. Well log of the Panuke PP-3C discovery well
voir phase changes caused by the introduction of showing cavernous porosity in the Abenaki V.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 423

Table 1. Abenaki Formation penetrations, deep Panuke.*

Well Name and Spud Top AB V Total Depth Net Pay Maximum mmcf/
Number Total Vertical (Depth and Test Rate day
Depth, Subsea Formation) (106 m3/day)
Measured Depth

Panuke F-09 August 23, 2000 3255.1 m 3815 m Abenaki 4 0m Flowed gas, TSTM 0
Panuke H-08 May 24, 2000 3348.5 m 3682 m Abenaki 4 104.7 m 1.55 55
Panuke M-79 July 10, 2000 3351.5 m 4597 m Abenaki 1 0m No test NT
Panuke M-79A October 11, 2000 3331.2 m 3934.2 m Abenaki 5 11.4 m TVD 1.75 62
Panuke PI1A-J99 August 27, 1999 3323.5 m 4033 m Abenaki 4 1.5 m No test NT
Panuke PI1B-J99 November 11, 1999 3322.2 m 4046.4 m Abenaki 4 13.9 m 1.49 53
Panuke PP3C J-99 July 17, 1998 3315 m 4163 m Abenaki 4 44.6 m 1.96 69
Margaree F-70 May 21, 2003 3330 m 3677 m Abenaki 4 70 m 1.5 53
Marcoh D-41 August 28, 2003 3306 m 3625 m Abenaki 4 74.2 m No test NT
Queensland M-88 December 14, 2001 3737.4 m (eq) 4443 m Abenaki 2 0m No test NT

*Well information summary table. Porosity cutoff for net pay calculation is 3%.

abundant LCM into the reservoir, as well as more
than 300,000 bbl of seawater pumped into the for-
mation during AVC drilling. Reservoir temperature
dropped simultaneously, from 130 to 808C in the
wellbore.
Delineation and Predevelopment Drilling Phase
The PI-1 well tested the extent of the deep Panuke
discovery to the northeast (Figure 5). It was drilled
on a seismic porosity anomaly analogous to PP3-C
interpreted in the AB V. The well (PI-1A) encountered
significant gas shows, some dolomite, but no cavern-
ous porosity. Instead of testing the AB V, the seismic
data was reinterpreted, and a sidetrack wellbore (PI-IB)
drilled 70 m (230 ft) to the northeast. PI-1B penetrated
moderate, dolomitic limestone porosity in the AB V.
The well was completed and tested gas at significant
rates (Table 1).
The dimensions of the pool to the southwest were
then tested by the H-08 well. The well encountered a
thick section of cavernous porosity in the AB V. An-
nular velocity control drilling was again necessary to
control the formation. Ultimately, logs showed 104 m
(341 ft) of pay at a slightly lower structural eleva-
tion than PP-3C and PI-1A (Figure 24) and tested gas
(Table 1). A further delineation well at M-79 had a
similar result to PI-1A (no significant AB V porosity in
the initial wellbore). However, M-79A, drilled 600 m
(1968 ft) to the southwest, encountered extremely
productive gas pay (Figure 24). The seismic section
(Figure 25) shows M-79 drilled behind the interpreted
platform margin, M-79A penetrating a significant
interval of porous dolomite closer to the margin. The
M-79 wellbore did encounter significant, wet porosity
lower in the Abenaki.
Two further wells investigated the northeast ex-
tent of the pool: F-70 is another gas well, analogous
to M-79A; and D-41, which was drilled in 2004, was
still confidential at the time of writing. Two other
wells, F-09 and M-88, were drilled off the main AB V
bank margin during the delineation phase of the proj-
ect. F-09 drilled 70 m (230 ft) of interbedded carbon-
ate muds and siliciclastics in the AB VI (Figure 24).
The underlying oolitic AB V interval had some gas
shows, but was too tight to be economic. The M-88 well
was drilled basinward of the Abenaki platform mar-
gin. It encountered bypassed siliciclastic sands in the
interpreted lowstand of the AB VII (Figure 26).
A new 3-D seismic survey, 450 km2 (173 mi2) in
area, was shot across the deep Panuke area in 2002.
These data were used to locate the subsequent wells
and to model the porosity in the pool for reserves
determination and depletion planning.
Predevelopment geoscience and engineering
work on deep Panuke was presented in Brown et al.
(2004). To summarize, shear sonic logs from eight
wells, vertical seismic profile data, high-quality seis-
mic processing of the new 3-D data set, including
Post-stack Time Migration, Post-stack Depth Migra-
tion (PSDM), and Compressional + Shear Wave imped-
ances and Lambda-Mu-Rho volumes, were used to
image porosity in the pool. Seismic reservoir charac-
terization used a supervised neural network approach
with seismic attributes from PSDM and amplitude-
versus-offset cubes as input nodes to analyze basin,
backreef, and reefal parts of the pool. Accuracy of
porosity prediction was then tested by matching well
424 / Weissenberger et al.

FIGURE 24. Structural cross section, deep Panuke pool, indicating distribution of conventional and sidewall core data.
Diagram modified from Wierzbicki and Harland (2004).
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 425
FIGURE 25. Seismic dip line through the M-79 and M-79a wells, deep Panuke pool. The platform margin is aggradational
from the AB II upward.
porosity to seismic data and defining a set of sig-
nificant seismic attributes. The attributes were first
extracted from a 3  3 trace area around each well-
bore and then from the entire volume and fed into
the trained neural net.
Numerous field-scale porosity-permeability mod-
els were developed by integrating geological and geo-
physical interpretations in a geostatistical framework.
These data were integrated with engineering infor-
mation into a ‘‘shared earth model, containing the
full 3D representation of the reservoir’’ (Brown et al.,
2004, p. 3). Numerous property models, generated
using seismic attributes, well log, and core data, were
simulated and compared to well test data. This was an
iterative process, testing the validity of each reservoir
model. The five wells in the pool with production
tests provide large-scale permeability data, which was
used to history match the different reservoir models,
thereby reducing uncertainty, particularly in perme-
ability estimates for the pool. Figure 27 shows one
such 3-D porosity model for the deep Panuke pool,
with an extracted seismic profile.
A geostatistical approach was also used to estimate
the likely reserve size distribution of the pool (Deutsch
et al., 2003). Geological and geophysical inputs (the
latter from the 2002 3-D data set) were used to real-
istically represent the reservoir heterogeneity.
Critical Elements
The deep Panuke pool is interpreted as a com-
bined stratigraphic-structural trap. A common gas-
water contact exists at 3504 m (11,496 ft) subsea
(Figure 24). Top seal is provided by argillaceous car-
bonates and shales in the upper Abenaki (Abenaki
426 / Weissenberger et al.

FIGURE 26. Geological dip

cross section interpreting plat-
form to basin transition be-
tween the M-79 and M-88
wells. Note bypassed quartz
sands in M-88, just below
3600 m (11,811 ft).

the Abenaki platform. Two

hypotheses for the origin of
the gas were contemplated:
biogenic production caused
by the prolonged interac-
tion of sea water with the
reservoir (while suspended
and AVC drilling) and mi-
gration from anhydrite-rich
strata where thermochemi-
cal sulfate reduction (TSR)
had occurred. Sulfur isotope
analyses were undertaken
on a sample of the gas and
on a sample of anhydrite
from the Eurydice Forma-
tion of the Mohican I-100
well (Figure 4). The analyses
(I. Hutcheon, 1999, personal
communication) indicate
that the sulfur in the deep
sequence VI and VII) and Mic Mac–lower Missisauga Panuke gas matches the isotopic signature in the
formations. Lateral seal is provided by these same Eurydice Formation anhydrite. Subsequent gas anal-
formations and shale- and mud-filled reentrants in ysis of wells where there was no significant fluid
the platform at the northeast end of the pool, as well influx showed the same H2S concentrations. It is
as tight, platform interior oolitic limestones. A re- likely that the observed sulfur concentrations are
gional structural doming also exists along the Abe- related to TSR processes.
naki carbonate margin, which supports the east–west
closure of the pool.
The source of the gas is believed to be organic CONCLUSIONS
shales in the Verrill Canyon Formation (Williamson
and Des Roches, 1993). The principal reservoir zones Figure 28 illustrates the tectonostratigraphic evo-
are reefal boundstones and associated sediments, hav- lution of the Jurassic carbonate hydrocarbon system
ing both intercrystalline and vuggy porosity in dolo- on the Scotian Shelf. During the Early Jurassic, sea
mites and secondary and vuggy porosity in limestones. level rise allowed the deposition of the Iroquois For-
The diagenesis responsible for this porosity was dis- mation, a broad, shallow-water carbonate shelf over
cussed above. synrift deposits of the Argo and Eurydice formations
(Figure 28A). The latter was largely restricted to topo-
graphic lows (half grabens) between tilted fault blocks.
Discussion Overlying the Iroquois, a thick section of siliciclastics
When small amounts of H2S were encountered in was deposited in the transition from the Lower to
tests of the PP-3C well, it raised the immediate ques- Middle Jurassic (Mohican Formation; Toarcian to Ba-
tion of provenance and possible extent of sour gas jocian). These strata are provisionally defined as a
in any further wells drilled or discoveries made on second-order depositional sequence.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 427

FIGURE 27. Derived porosity distribution, Abenaki sequence V, deep Panuke pool. Seismic section shown is taken just
north of the M-79 well. The red and black colors indicate amplitudes interpreted as the highest porosity zones in the
porosity model. Interpretation by R. Tonn.

Later in the Middle Jurassic (Bajocian – Callovian;
Figure 28B) the first Abenaki third-order deposition-
al sequence (AB I) was deposited above the Mohican
Formation with continued relative sea level rise. The
second Abenaki sequence (AB II) comprises two parts:
the basinal Misaine Formation (transgressive systems
tract) and the overlying foreslope and shallow-water
carbonates (highstand systems tract). The Misaine rep-
resents the MFS of both the third-order AB II and the
second-order depositional sequence. Figure 28B il-
lustrates some of the stratal geometries that are ap-
parent on seismic data, and the possibility of reefs
initiating on the drowned AB I.
Beginning in the Upper Jurassic, the stratal archi-
tecture of the Abenaki platform was essentially ag-
gradational. The coral-stromatoporoid reef margin
became well established (Figure 28C). During rela-
tive lowstands in sea level, siliciclastics from the plat-
form interior bypassed into the basin. Minor epikarst
developed during these lowstands. Carbonate plat-
form deposition continued through the lowermost
Cretaceous (Berriasian; Figure 28D). The large-scale
stacking pattern was aggradational to the top of the
AB V in the deep Panuke area, then retrogradational
in the AB VI and VII. Northeast of deep Panuke, sedi-
mentation was dominated by siliciclastics.
From the Early to middle Cretaceous, sedimen-
tation was dominated by siliciclastics (Figure 28E).
Tectonic activity in the region began in the Aptian
(Jansa and Pe-Piper, 1988). This caused the reacti-
vation of basement faults in a strike-slip sense and
the movement of hydrothermal fluids into the Abe-
naki carbonate platform. These fluids completed the
creation of secondary limestone and dolomite po-
rosity begun earlier in the burial history (R. A. Wierz-
bicki, J. J. Dravis, I. Al-Aasm, and N. J. Harland, 2005,
personal communication). The Verrill Canyon source
rock entered the gas-generation window in the Ap-
tian (Williamson and Des Roches, 1993), likely filling
the deep Panuke trap by the Albian.
Drilling since 1998 has proven a large accumula-
tion of natural gas at deep Panuke, extending 26 km
(16 mi) along the Jurassic carbonate platform margin.
EnCana has not published a revised reserves estimate
since 2002, despite the drilling of several delineation
wells. It remains the first and only significant hydro-
carbon discovery in the Mesozoic carbonates on the
entire continental shelf of eastern North America.
428 / Weissenberger et al.

FIGURE 28. Tectonostratigraphic

evolution of the carbonate plat-
form, deep Panuke area. (A) Early
Jurassic; Iroquois Formation car-
bonate platform deposited during
transgression over Triassic synrift
deposits. Early dolomite formation
in semiarid climate. (B) Bathonian–
Callovian; maximum flooding
surface of second-order deposi-
tional sequence (Misaine Forma-
tion) and progradation of Abenaki
II highstand carbonate. (C) Oxfor-
dian; exposure during a relative
sea level lowstand at the base of
an Abenaki third-order sequences
(e.g., AB IV) yields meteoric karst
and cross-bank transport (and
foreslope deposition) of siliciclas-
tics. (D) Kimmeridgian–Berriasian;
relative transgression in Abenaki VI
and VII sequences. Platform ar-
chitecture changes from domi-
nantly aggradational to retrogra-
dational in the deep Panuke area.
(E) Valanginian – Cenomanian;
Missisauga to Logan Canyon sili-
ciclastics deposited; subsidence of
forereef; regional wrench tecto-
nism moves hot diagenetic fluids
up reactivated normal faults.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 429

APPENDIX 1: NEW CORE AND SIDEWALL CORE DATA, DEEP PANUKE

Location Core Type Core Interval Number Recovery Analysis Done

of Side-wall (Whole Core)
Core
PP-3C None

PI-1A Whole core Whole core 4029.28 – 4040 m 45 1.15 m (3.77 ft) Routine
and sidewall (13,219.42 – 13,254 ft)

Side-wall core 3895 – 4034 m Routine

(12,778 – 13,234 ft)

PI-1B None

M-79 Whole core Whole core 4532.7 – 4538.7 m 99 5 m (16 ft) None
and sidewall (14,871.06 – 14,890.7 ft)

Side-wall core 3512.5 – 4591 m

(11,523.9 – 15,062 ft)

M-79A None

H-08 Whole core Whole core 3446 – 3460 m 3.2 m (10.4 ft) Routine and
(11,305 – 11,351 ft) special core

F-09 Sidewall Side-wall core 3264 – 3798 m 31 Routine

(10,708 – 12,460 ft)

F-70 Whole core Whole core 3434 – 3461 m 45 24.65 m (80.8 ft) Routine
and sidewall (11,266 – 11,354 ft)

Side-wall core 3350 – 3636 m Special core

(10,990 – 11,929 ft) analysis

APPENDIX 2: BIOSTRATIGRAPHIC SUMMARY, M-79 WELL, DEEP PANUKE

Depth (m) Age Assignment Cycle

3180 (top sample) to 3220 Early Cretaceous, Neocomian AB 7

3240 – 3360 Portlandian, anguiformis or older AB 6

3380 – 3440 Late Kimmeridgian, fittoni or older AB 6

3460 – 3820 Late to early Kimmeridgian, elegans or mutabilis or older AB 5/4/3

3840 – 4020 Late Oxfordian, rozenkrantzi or older AB 3

4040 – 4240 Middle Oxfordian, glosense to Callovian AB 2

4260 – 4460 Middle Callovian, coronatum and jason AB 2/Misaine

4480 – 4600 (total depth) Callovian AB 1 Scatarie

B. G. T. van Helden
430 / Weissenberger et al.
ACKNOWLEDGMENTS
We thank EnCana Ltd. for giving permission to
publish our work on deep Panuke and Husky Energy
for additional support in completion of this publica-
tion. This paper would not have been written without
the contribution of our many colleagues and associ-
ates who worked on various aspects of the project, and
we beg the pardon of those we do not mention. J. A.
W. Weissenberger thanks John Hogg and Ian DeLong
for the opportunity to work with them on the project
from the beginning. We must single out the contri-
butions of our geophysicists, Garth Syhlonyk, Robert
Riddy, and Rainer Tonn, who worked through the
delineation and development phases, as well as en-
gineers Tom Craig and John Slade. We must also ac-
knowledge several colleagues who supported this proj-
ect through their specialized work: Ihsan Al-Aasm
(isotopes and fluid inclusions), Jeff Dravis (petrogra-
phy and diagenesis), Les Eliuk and Bev Harris (drill-
cuttings description), Bert van Helden (micropaleon-
tology), and Ian Hutcheon (sulfur isotopes). Thanks
again to John Hogg for critically reading the manu-
script and to Phil Argatoff of Penta Graphix Ltd., who
prepared many of the diagrams. The authors also ap-
preciate the helpful comments of reviewers Sherry
Becker, Niall Toomey, and Mitch Harris, which greatly
improved the manuscript. Please note that the refer-
ence to proven reserves in the AAPG 2006 conference
abstracts is withdrawn because it was included in error.
REFERENCES CITED
Blomeier, D. P. G., and J. J. G. Reijmer, 2002, Facies archi-
tecture of an Early Jurassic carbonate platform slope
(Jbel Bou Dahar, High Atlas, Morocco): Journal of Sedi-
mentary Research, v. 72, no. 4, p. 462 – 475.
Brown, S., R. Riddy, R. Tonn, and R. A. Wierzbicki, 2004,
The integration of geology, geophysics and reservoir
engineering for field appraisal (abs.): Canadian Society
of Exploration Geophysics Annual Meeting Abstracts
and Program, 4 p.
Deutsch, C., R. A. Wierzbicki, R. Riddy, and J. Slade, 2003,
Quantification of uncertainty in gas resources of deep
Panuke (abs.): AAPG Annual Meeting Program, v. 12,
6 p., CD-ROM.
Eliuk, L. S., 1978, The Abenaki Formation, Nova Scotia
Shelf, Canada — Depositional and diagenetic model
for a Mesozoic carbonate platform: Bulletin of Cana-
dian Petroleum Geology, v. 26, p. 424 – 514.
Eliuk, L. S., and R. Levesque, 1989, Earliest Cretaceous
sponge reef mound, Nova Scotia shelf (Shell Demas-
cota G-32), in H. H. J. Geldsetzer, N. P. James, and G. E.
Tebbutt, eds., Reefs, Canada and adjacent areas: Cana-
dian Society of Petroleum Geologists Memoir 13,
p. 713 – 720.
Eliuk, L. S., S. C. Cearley, and R. Levesque, 1986, West
Atlantic Mesozoic carbonates: Comparison of Balti-
more Canyon and offshore Nova Scotian basins: AAPG
Bulletin v. 70, p. 586 – 587.
Ellis, P. M., 1984, Upper Jurassic carbonates from the
Lusitanian basin, Portugal and their subsurface coun-
terparts in the Nova Scotian Shelf: Ph.D. thesis, Open
University, England, 283 p.
Ellis, P. M., P. D. Crevello, and L. S. Eliuk, 1985, Upper Ju-
rassic and Lower Cretaceous deep-water buildups, Abe-
naki Formation, Nova Scotian Shelf, in P. D. Crevello
and P. M. Harris, eds., Deep-water carbonates: SEPM
Core Workshop 6, p. 212 – 248.
Ellis, P. M., R. C. L. Willson, and R. R. Leinfelder, 1990,
Controls on Upper Jurassic carbonate buildup devel-
opment in the Lusitanian basin, Portugal: Interna-
tional Association of Sedimentologists Special Publi-
cation 9, p. 169 – 202.
Given, M. M., 1977, Mesozoic and early Cenozoic geology
of offshore Nova Scotia: Bulletin of Canadian Petro-
leum Geology, v. 25, p. 63 – 91.
Golonka, J., M. I. Ross, and C. R. Scotese, 1994, Phanerozoic
paleogeographic and paleoclimatic modeling maps, in
A. F. Embry, B. Beauchamp, and D. J. Glass, eds., Pangaea,
global environments and resources: Canadian Society
of Petroleum Geologists Memoir 17, p. 1 – 47.
Gradstein, F. M., M. A. Kaminski, and F. P. Agterberg, 1999,
Biostratigraphy and paleoceanography of the Creta-
ceous seaway between Norway and Greenland, in F. M.
Gradstein and G. J. van der Zwaan, eds., Ordering the
fossil record; challenges in stratigraphy and paleontol-
ogy: Selected papers from a symposium held in honor
of the 75th birthday of Cor Drooger: Earth Science Re-
views, v. 46, no. 1 – 4, p. 27 – 98.
Haq, B. U., J. Hardenbol, and P. R. Vail, 1987, Chronology
of fluctuating sea levels since the Triassic: Science,
v. 235, p. 1156 – 1167.
Harland, N. J., J. R. Hogg, R. Riddy, G. Syhlonyk, G.
Uswak, J. A. W. Weissenberger, and R. A. Wierzbicki,
2002, A major gas discovery at the Panuke field, Ju-
rassic Abenaki Formation, offshore Nova Scotia (abs.):
Canadian Society of Petroleum Geologists, Annual
Meeting, Calgary, p. 154.
Hogg, J. R., and M. E. Enachescu, 2003, An overview of the
Grand Banks and Scotian Shelf basins development
and exploration, offshore Canada (abs.): AAPG Interna-
tional Conference and Exhibition, Barcelona, Spain,
CD-ROM.
Jansa, L. F., 1993, Early Cretaceous carbonate platforms of
the northeastern North American margin, in J. A. T.
Simo, R. W. Scott, and J. P. Masse, eds., Cretaceous car-
bonate platforms: AAPG Memoir 56, p. 111– 126.
Jansa, L. F., and G. Pe-Piper, 1988, Middle Jurassic to Early
Cretaceous igneous rocks along eastern North Amer-
ican continental margin: AAPG Bulletin, v. 72, no. 3,
p. 347 – 366.
 Sequence Stratigraphy and Petroleum Geology of the Jurassic Deep Panuke Field / 431
Jansa, L. F., and J. A. Wade, 1975, Paleogeography and sedi-
mentation in the Mesozoic and Cenozoic, southeastern
Canada, in W. J. M. van der Linden and J. A. Wade, eds.,
Offshore geology of eastern Canada: Geological Survey
of Canada Paper 74-30, p. 51– 105.
Jansa, L. F., and J. Wiedmann, 1982, Mesozoic – Cenozoic
development of the eastern North American and north-
west African continental margins — A comparison, in
U. von Rad, K. Hinz, M. Sarnthein, and E. Seibold, eds.,
Geology of the northwest African continental margin:
New York, Springer-Verlag, p. 215 – 269.
Leinfelder, R. R., 1994, Distribution of Jurassic reef types:
A mirror of structural and environmental changes
during the breakup of Pangaea, in A. F. Embry, B. Beau-
champ, and D. F. Glass, eds., Pangaea: Global environ-
ments and resources: Canadian Society of Petroleum
Geologists Memoir 17, p. 677 – 700.
McIver, N. L., 1972, Cenozoic and Mesozoic stratigraphy
of the Nova Scotia Shelf: Canadian Journal of Earth
Sciences, v. 9, p. 54 – 70.
Poag, C. W., 1991, Rise and demise of the Bahama – Grand
Bank gigaplatform, northern margin of the Jurassic
proto-Atlantic seaway: Marine Geology, v. 102, p. 63 –
103.
Prather, B. E., 1991, Petroleum geology of the Upper Juras-
sic and Lower Cretaceous, Baltimore Canyon Trough,
western North Atlantic Ocean: AAPG Bulletin, v. 75,
p. 258 – 277.
Pratt, B. R., and L. F. Jansa, 1989, Late Jurassic shallow wa-
ter reefs of offshore Nova Scotia, in H. H. J. Geldsetzer,
N. P. James, and G. E. Tebbutt, eds., Reefs, Canada and
adjacent areas: Canadian Society of Petroleum Geolo-
gists Memoir 13, p. 741 – 747.
Scotese, C. R., 1997, Paleogeographic atlas: PALEOMAP
Progress Report 90-0497, University of Texas at Arling-
ton, 20 p.
Vail, P. R., R. M. Mitchum Jr., and S. Thompson, 1977,
Seismic stratigraphy and global changes in sea level:
Part 4. Global cycles of relative changes of sea level, in
Seismic stratigraphy — Applications to hydrocarbon
exploration: AAPG Memoir 26, p. 83 – 97.
Van Wagoner, J. C., H. W. Posamentier, R. M. Mitchum,
P. R. Vail, J. F. Sarg, T. S. Loutit, and J. Hardenbol, 1988,
An overview of the fundamentals of sequence stratigra-
phy and key definitions, in C. K. Wilgus, B. S. Hastings,
C. G. St. C. Kendall, H. W. Posamentier, C. A. Ross, and
J. C. Van Wagoner, eds., Sea level changes: An inte-
grated approach: SEPM Special Publication 42, p. 125 –
154.
Wade, J. A., 1977, Stratigraphy of the George’s Bank basin—
Interpreted from seismic correlation to the western Sco-
tian Shelf: Canadian Journal of Earth Sciences, v. 14,
p. 2274 – 2283.
Weissenberger, J. A. W., N. J. Harland, J. R. Hogg, and G.
Syhlonyk, 2000, Sequence stratigraphy of Mesozoic
carbonates, Scotian Shelf, Canada (abs.): GeoCanada
2000 Convention ( Joint Meeting of the Canadian
Geophysical Union, Canadian Society of Exploration
Geophysicists, Canadian Society of Petroleum Geolo-
gists, Canadian Well Logging Society, Geological
Association of Canada, and the Minerological Associ-
ation of Canada) CD disk (5 p., 4 figures).
Welsink, H. J., J. D. Dwyer, and R. J. Knight, 1989,
Tectono-stratigraphy of the passive margin off Nova
Scotia, in A. J. Tankard and H. R. Balkwill, eds., Exten-
sional tectonics and stratigraphy of the North Atlantic
margin: AAPG Memoir 46, p. 215 – 231.
Wenzel, A., and A. Strasser, 2001, Sedimentology, paleo-
ecology and high resolution sequence stratigraphy of
a carbonate-siliciclastic shelf (Oxfordian, Swiss Jura
Mountains): Field trip guidebook A3, International Asso-
ciation of Sedimentologists, Davos, Switzerland, 19 p.
Wierzbicki, R. A., and N. J. Harland, 2004, Diagenetic mod-
el: deep Panuke reservoir, offshore Nova Scotia, Canada
(abs.): AAPG Annual Convention Abstracts and Pro-
gram, CD-ROM.
Wierzbicki, R. A., N. J. Harland, and L. S. Eliuk, 2002, Deep
Panuke and Demascota core from the Jurassic Abenaki
Formation, Nova Scotia: Facies models, deep Panuke,
Abenaki Formation (abs.): Canadian Society of Petro-
leum Geologists Annual Meeting Core Convention
Extended Abstracts, p. 71 – 101.
Williamson, M. A., and K. Des Roches, 1993, A maturation
framework for Jurassic sediments in the Sable subba-
sin, offshore Nova Scotia: Bulletin of Canadian Petro-
leum Geology, v. 41, p. 244 – 257.
 12
Ramon, J. C., and A. Fajardo, 2006, Sedimentology, sequence stratigraphy,
and reservoir architecture of the Eocene Mirador Formation, Cupiagua
field, Llanos Foothills, Colombia, in P. M. Harris and L. J. Weber, eds.,
Giant hydrocarbon reservoirs of the world: From rocks to reservoir
characterization and modeling: AAPG Memoir 88/SEPM Special
Publication, p. 433 – 469.

Sedimentology, Sequence
Stratigraphy, and Reservoir
Architecture of the Eocene
Mirador Formation, Cupiagua
Field, Llanos Foothills, Colombia
Juan Carlos Ramon
BP Colombia, Bogotá, Colombia

Andres Fajardo
Chevron-Texaco, Bogotá, Colombia

ABSTRACT

T
he stratigraphic architecture and facies distributions in a high-resolution
time-space framework define the three-dimensional (3-D) reservoir zonation
of the Mirador Formation in the Cupiagua field. A high-resolution genetic
sequence stratigraphy study, using more than 731 m (2400 ft) of core and 40 well logs,
is integrated with petrophysical information to populate a static structural model
based on the interpretation of a 312-km2 (120-mi2) 3-D seismic volume. Dynamic data
(pressure, gas tracers, gas-oil ratio behavior) are integrated with the geological model
to better define sandstone bodies’ lateral continuity. Production logs (production
logging tools) complement petrophysical data in the definition of fluid-flow units.
Three scales of stratigraphic cycles are recognized in the Mirador Formation.
Short-term (high-frequency) cycles correspond to progradational-aggradational
units. Six intermediate-term cycles are identified by the stacking patterns of their
component short-term cycles and by the general trend of facies successions, in-
dicating increasing or decreasing accommodation-to-sediment supply (A/S) ratios.
Two long-term cycles are defined from the stacking pattern of the intermediate-
term cycles and by the general trend of facies successions.
The lower half of the Mirador Formation consists of coastal-plain facies tracts
and is composed of channel, crevasse splay, and swamp and flood-plain facies
successions. A bay facies tract occurs in the upper half of the Mirador Formation
and is composed of bay-fill, bay-head delta, and channel facies successions.

Copyright n2006 by The American Association of Petroleum Geologists.

DOI:10.1306/1215884M883276

433
434 / Ramon and Fajardo

The lower Mirador Unit was deposited over a wide flood-plain sequence.
Each intermediate-term cycle is composed of aggradational channel deposits, pro-
gradational and aggradational crevasse splay bodies, and aggradational swamp
and flood-plain facies successions. The first two intermediate-scale cycles (I, II)
show a seaward-stepping pattern, and then the next cycle (III) shows a landward-
stepping stacking pattern. The fall-to-rise turnaround is located at the base of
cycle II. The upper Mirador shows the continuous landward-stepping pattern and
places prograding bay-head delta and bay-fill facies successions over the alluvial-
plain setting of the Lower Mirador. This upper unit consists of three onlapping
cycles composed of a succession of aggradational channel deposits, prograda-
tional bay-head delta, and bay-fill deposits with a landward-stepping stacking
pattern. During this cycle, deepening-up bay-fill facies successions were deposited
in the area as a consequence of the increasing accommodation conditions that
prevailed in the area. Finally, the Mirador is capped by restricted-marine shales of
the Carbonera Formation.
The Cupiagua structure is a large, east-verging, asymmetric anticlinal fold
that trends north-northeast in the hanging wall of the frontal fault. Average
length and width of the Cupiagua structure are 25 and 3 km (15 and 1.8 mi),
respectively. The original oil in place in the Cupiagua field is estimated between
1000 and 1100 MMSTB of oil and 3000 to 4500 mmcf of gas. The Mirador For-
mation accounts for approximately 51% of the recoverable oil in the Cupiagua
field.

INTRODUCTION The Cupiagua field is a rich, near-critical gas conden-

The Cusiana and Cupiagua fields, discovered in
1988 and 1992, respectively, are located about 150 km
(93 mi) northeast of Bogota, Colombia, South Amer-
ica (Figure 1). The fields lie in the foothills trend on
the edge of the eastern Cordillera. The first oil pro-
duction from Cusiana began in September 1992 as
part of an early production scheme via the existing
pipeline systems. Commerciality of Cusiana and a small
central area of Cupiagua were declared in June 1993.
Phase I of the development of Cusiana allowed pro-
duction to increase to 185 MBOPD during 1995. A
second phase of the project was sanctioned during
1995. This consisted of an expansion of the Cusiana
facilities, the construction of an 800-km (497-mi) pipe-
line between Cusiana and the oil terminal at the
Atlantic Coast (Coveñas) and the construction of the
Cupiagua central processing facilities. This second de-
velopment phase allowed production to progressively
ramp up from 185 MBOPD to reach a peak through-
put in the first half of 1998 of 341 MBOPD, with gas
handling rates in excess of 1.4 MMSCFD. During 1997,
the Cupiagua South accumulation was discovered.
The Cusiana and Cupiagua field developments are
now close to complete, and the fields have entered
their decline phase. At mid-2005, the field has 48 wells,
36 producers, and 12 gas-injector wells (Figure 2).
sate reservoir over a very high (1800-m; 6000-ft)
hydrocarbon column. The fluid composition changes
slightly with depth. The condensate yield is about
250 bbl/MMSCF of gas, and the average API gravity
is 408. The initial reservoir pressure was close to
6500 psia. The dew point pressure is 5350 psia.
The main drive mechanism is gas reinjection and
gas expansion. An efficient revaporization recovery
mechanism exists. At mid-2005, the pore volume re-
placement is about 70%. A predicted decline of close
to 30% yr1 is exacerbated by reservoir complexity,
scale, and fluid issues like gas recycling. Recent op-
timization of the gas injection resulted in an average
decline of about 23% yr1.
REGIONAL SETTING
The Llanos Foothills are located between the un-
deformed Llanos foreland grasslands and the high-
elevation, highly deformed eastern Cordillera. The
Llanos Foothills involves a zone of frontal deforma-
tion running northeast for hundreds of kilometers
and about 20 km (12 mi) wide (Cooper et al., 1995).
The foothills are limited by the Guaicaramo and the
frontal-Yopal fault systems. The main outcropping
structural feature is the Nunchia syncline.
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 435
The stratigraphic column (Figure 3) may be divid-
ed into three major depositional cycles, which, in turn,
may be further divided. The first cycle covers the Pa-
leozoic, the second begins in the Albian–Cenomanian
and continues until the Paleocene (mainly Cretaceous
succession), and the third extends from the Eocene to
the present. The rock successions corresponding to
these cycles are separated by the regionally significant
sub-Cretaceous and sub-Paleocene unconformities.
Early Cretaceous extension created a graben where
the Cordillera Oriental mountain range is today, which
was filled with as much as 3 km (1.8 mi) of Lower
Cretaceous marine sediments (Dengo and Covey,
1993). The main Cretaceous transgression, which be-
gan in the Aptian, entered from the north and spread
rapidly across the basin. These sediments were pri-
marily derived from the Guyana shield to the east.
Early Tertiary uplift and erosion have cut back the
present edge of Cretaceous sediments well west of
their original depositional limit. From west to east,
progressively younger Cretaceous strata overlie pro-
gressively older Paleozoic rocks.
FIGURE 1. Location map of the
Cupiagua and Cusiana fields
in the Llanos Foothills. The
map shows the different fields
and their producing fluids.
Cretaceous strata are domi-
nated by shallow-marine, del-
taic, and basal transgressive
sandstones. Interbedded shales
increase in thickness and fre-
quency basinward. The Cre-
taceous cycle (Figure 3) rep-
resents a sedimentary wedge
with a central shale unit (Ga-
cheta Formation) enclosed
by transgressive sands (the
Une and Guadalupe forma-
tions, respectively).
The Tertiary sequence be-
gan with deposition of the
laterally amalgamated chan-
nel-belt systems of the Barco
Formation. The overlying Los
Cuervos Formation consists
of aggradational flood-plain
deposits and isolated single-
story channel belts. Paleocene
deposits are preserved in a
relatively narrow region along the foothills but have
been mostly removed by subsequent erosion in most
of the adjacent Llanos basin. The Eocene Mirador
Formation transitionally overlies the Los Cuervos
mudstones. The Carbonera Formation comprises four
regressive sandy units (labeled C1, C3, C5, and C7)
intercalated with four transgressive shale units (C2,
C4, C6, and C8). The shale units can be correlated
regionally, particularly the uppermost unit. The sandy
units are interpreted to be nearshore, coastal plain,
and predominantly deltaic. The Carbonera Forma-
tion, as with all previous formations, was deposited
in a basin that extended to the west far beyond the
present-day Llanos (Villamil, 1999). The Carbonera
Formation thickens steadily westward toward the
mountain front, reaching a thickness of more than
5000 ft (1500 m). The shaly Miocene Leon Forma-
tion, which overlies the Carbonera Formation, pro-
vides the first indication of the uplift of the Cordillera
Oriental and the isolation of the Llanos basin in the
east from the Magdalena basin in the west. The thick-
est Leon Formation sediments are located east of
436 / Ramon and Fajardo

FIGURE 2. Well base map of

the Cupiagua field. Main struc-
tural features are shown. Pro-
ducing wells are in green, and
injector wells are in red.

more than 4000 m (13,000 ft).

These sediments are mainly
coarse-grained red beds of
continental origin.
The hydrocarbons were
generated toward the west
from the world-class source
rock La Luna – Gachetá For-
mation, buried more than
6000 m (20,000 ft) below the
main thick skin thrusting of
the eastern Cordillera.

CUPIAGUA STRUCTURE

The Cupiagua structure is

the present basin edge, where they reach as much as
1000 m (3300 ft) in thickness. Farther westward, a
thinning trend (facies change into sandstones) is
apparent before the formation reaches the mountain
front. The Leon Formation consists of homogeneous
shale across much of the central Llanos basin but is
rich in sandstone layers along the foothills. Coarser
clastics present in the west and northwest are in-
terpreted as having been derived from the eastward
reworking of Carbonera sands exposed in island ridges
ancestral to the Cordillera Oriental (Cooper et al.,
1995; Villamil, 1999). The Miocene –Holocene Gua-
yabo and Necesidad formations represent very thick
mollasse deposits associated to the uplifting of the
eastern Cordillera. This last cycle is a thick, easterly
thinning wedge of sands and shales, which reaches
an elongated asymmetrical
anticline running north-
northeast for approximately
30 km (18 mi) (Figure 4). This
structure is an east-southeast-
propagating hanging-block
anticline associated with the
east-southeast-verging fron-
tal and core faults (Figure 5).
The regional detachments for
these faults are in the mud-
stone intervals of the Creta-
ceous Gacheta Formation.
The Cupiagua structure lies below the Yopal fault,
which separates the Nunchia syncline from the un-
derlying structural deformation (Coral and Rathke,
1997). The west flank is 2.5 – 3 km (1.5– 1.8 mi) wide
with average dip of 35–388. Locally, it can get as much
as 4 km (2.5 mi) wide. The forelimb is quite variable
from south to north along the structure (Figures 4, 5).
In the south part of the Cupiagua structure, the fore-
limb is very high angle to inverted and is almost as
high as the west flank (Figure 4B). To the north, the
forelimb is structurally simpler, and finally, it disap-
pears in the middle part of the field (Figure 4C). Far-
ther north along the structure, the detachment position
rises stratigraphically, and the whole structure bends
strongly toward the east. Figure 6 is a time slice of the
Cupiagua three-dimensional (3-D) volume showing
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 437

FIGURE 3. Generalized stratigraphic column of the Llanos Foothills. The main reservoir sandstones of the Mirador, Barco,
and Guadalupe formations are shown.
438 / Ramon and Fajardo
the north–south trend of the syncline over the Cu-
piagua structure. Toward the northern portion, the
syncline bends toward the northeast (Figure 6). Along
the area of strike change, a lateral ramp or a later strike-
slip fault is interpreted. This lateral ramp corresponds
to a surface lineament called the Golconda wrench
fault (Coral and Rathke, 1997). This feature affects
the whole sequence from surface to the Cretaceous
sequence, and the Cupiagua structure bends to the
northeast and then continues to the north (Figure 4).
The western limit of the structure is a west-verging
series of faults. In the southern half (south of the
lateral ramp) of the field, there is a major backfault
that seems to cut completely the east-southeast-
verging system of thrusts (Figure 4B). As seen in
Figure 5A, the seismic reflectors below the frontal
fault are continuous, and the frontal fault looks broken
in two by the west-verging backthrust. This fault is
interpreted to be a backthrust of a deeper thrust (deeper
than the frontal thrust) and to cut all overlying thrust
faults. This backthrust seems to be genetically asso-
ciated with the Golconda lateral ramp, as it dies and
does not exist north from the ramp. On the contrary,
in the northern part of the field (north of the lateral
ramp), the western limit of the field is defined by a
series of backthrusts of the frontal fault (Figure 5B).
These west-verging faults are genetically related to
the propagation of the Cupiagua structure toward
FIGURE 4. Details of the Cu-
piagua structure in (A) 3-D
view. Note the northeast bend-
ing in the Cupiagua structure.
This change in strike is associ-
ated to a vertical feature that
affects the whole sedimentary
sequence and in geological
maps is called the Golconda
wrench. Structural cross sec-
tions across (B) the southern
and (C) central-northern parts
of the Cupiagua field show the
change in structural style along
strike (after Martinez, 2003).
In the south, the structure has
a high-angle to inverted im-
bricate. Toward the north, this
imbricate disappears, and the
structure becomes a simple
hanging-block anticline.
the east, and they formed
where the frontal thrust
bends (Figure 4C). Structural
mapping, fault-seal analysis,
well-test data, production data, and gas tracer results
demonstrate that the faults in the anticline are not
compartmentalizing the field. A few exceptions oc-
cur where these faults act as baffles.
Associated with the Golconda lateral ramp or wrench
in the central part of the structure are a series of east-
to southeast-trending lineaments. These very high-
angle faults cause minor displacements of the Car-
bonera strata and seem to extend down and cut the
reservoir section. Cross-line 1510 (Figure 7) running
north-northeast shows several of these almost vertical
faults. The Golconda lineament dipping to the north
is the lateral ramp feature that causes the bend in the
Cupiagua structure. South of this feature, an opposite
thrust exists that breaks the Cupiagua structure.
Several secondary faults occur both at the north and
south but close to the lateral ramp. Seismic quality
in the reservoir section is poor, and it is hard to de-
termine if the vertical faults cut all the way down.
Locally, dynamic data support the presence of some
of these faults and evidence of their seal or baffle
character.
RESERVOIR STRATIGRAPHY
In the Cupiagua and Cusiana fields, the reservoir
sandstones comprise Late Cretaceous to late Eocene
shallow-marine to alluvial sandstones, sourced from
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 439
the Guyana field to the east. The reservoirs display
sheetlike packages of shoreface bodies; laterally and
vertically amalgamated channel fills and overbank
deposits and bay-head delta and bay-fill bodies.
The Mirador Formation is the main reservoir zone
and contains 51% of the hydrocarbons initially in
place. Barco and Guadalupe formations have 28 and
21% hydrocarbons initially in place, respectively.
Cupiagua field has moderate core coverage of the
reservoirs, with six cores taken from Mirador, four
from Barco, and four from Guadalupe Formation, for
a total of about 1100 m (3800 ft).
FIGURE 5. Dip seismic sections
across the Cupiagua field.
Cupiagua wells close to each
line are shown. Inline 1200
(A) runs along the southern
part of the Cupiagua structure.
Note that the dominant struc-
tural feature is the Nunchia
syncline. The Yopal fault sepa-
rates two different structural
styles of deformation. In this
seismic line, it appears that the
back-main west-verging thrust
breaks the Cupiagua structure
and the frontal thrust. This
would suggest that the back-
fault is a younger event. This
back-main thrust extends only
south of the Golconda linea-
ment. Inline 1900 (B) is located
in the northern part of the
Cupiagua structure. Note that
the Nunchia syncline axis and
the crest of the Cupiagua
structure move eastward. The
west-verging faults are, in
this case, associated to the
frontal fault.
This chapter concentrates
on the sedimentology and
the stratigraphy of the Mir-
ador Formation.
FACIES, FACIES
SUCCESSIONS, AND
FACIES TRACTS IN THE
MIRADOR FORMATION
Two facies tracts exist in
the Mirador Formation in
the Cupiagua field. The
coastal-plain facies tract is
located mostly in the lower half of the Mirador and
is composed of channel, crevasse splay, and swamp
and flood-plain facies successions. The bay facies
tract occurs in the upper half of the Mirador For-
mation and is composed of bay-fill, bay-head delta,
and channel facies successions (Figure 8). Each facies
succession consists of a spectrum of facies that are
arranged in regular vertical and lateral successions.
These regular facies successions reflect lateral facies
transitions along the depositional profile and their
vertical superposition through progradation and
aggradation.
440 / Ramon and Fajardo

FIGURE 6. Time slice at 2236 ms

of the Cupiagua 3-D volume
showing the attitude of the
Carbonera strata. The axis of
the Nunchia syncline runs along
the left. Most strata over the
time slice belong to the east
flank of the syncline. Leon (yel-
low) and Carbonera (blue, red)
markers are shown. Note the
bending of the syncline and the
displacement of strata along
the Golconda lineament.

ture are followed by initials

This section presents the results of the facies
analysis and descriptions and interpretations of the
facies, facies successions, and facies tracts recognized
in this study. Interpretations are discussed in terms of
their inferred hydrodynamic and environmental pa-
rameters in a stratigraphic context.
Facies abbreviations are constructed from the
dominant sedimentary structure and the texture of
the rock. Initials of the dominant sedimentary struc-
of the texture of the rock.
Trough cross-stratification is
abbreviated as tx, ripples as
rp, sandstone as Ss, mud-
stone as Md, and granule
sandstone as gSs. For exam-
ple, a facies of trough cross-
stratified granule sandstone
is txgSs, sandstone with hor-
izontal burrows is hbSs, and
burrowed and irregular lam-
inated mudstone is blMd. Figure 9 shows conven-
tions of sedimentary structures, bed contacts, and
other symbols that are used in figures and core de-
scription logs.
Core logs for the Cupiagua A-1, C-3, and H-11
cored wells are presented in Figures 10 –12. These
logs include a graphic representation of the core de-
scription, facies interpretation, inferred stratigraphic
cycles, gamma-ray response, and important remarks.

FIGURE 7. Strike line (cross-

line 1510) along the crest of the
Cupiagua structure. North is
to the right. The main feature in
the center of the figure is the
Golconda wrench (magenta
fault). Several vertical faults oc-
cur associated to this wrench.
North of this feature, the Cupi-
agua structure and the Nun-
chia syncline bend toward the
northeast. Mirador and intra-
Carbonera horizons are shown.
Note the change in thickness
between the C5 and Mirador
markers caused by the Yopal
thrust fault. This fault separates
the Cupiagua structure from
the overriding Nunchia syncline.
Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 441

FIGURE 8. Gamma-ray (GR) log and

interpreted stratigraphic cycles and fa-
cies tracts in the Mirador Formation of
the Cupiagua A-1 well. The coastal-plain
facies tract composes the lower half of
the Mirador and the bay facies tract
constitutes the upper half of the Mira-
dor Formation. Intermediate-term and
long-term stratigraphic cycles are shown.

Coastal-plain Facies Tract

The coastal-plain facies tract oc-
curs in the lower half of the Mirador
Formation. Three facies successions
were recognized: channel, crevasse
splay, and swamp and flood-plain
facies successions. The channel facies
succession occurs at the base of the
intermediate-term cycles and is com-
posed of medium- to coarse-grained
sandstones and sandstones with
floating granule-size clasts. The cre-
vasse splay facies succession overlies
the interval of channel sandstones
and is composed mainly of very fine-
to fine-grained sandstones. The
swamp and flood-plain facies succes-
sion occurs associated with the cre-
vasse splay complex and consists of
massive, laminated, and burrowed
mudstones intercalated with flood-
plain deposits and soils. The change
from channel to crevasse splay to
swamp facies succession defines an
overall increase in accommodation
space. Flood-plain deposits represent
rock deposited during a period of de-
creased accommodation.

CHANNEL FACIES SUCCESSION

The channel facies succession

contains the main reservoir rock in
the coastal-plain facies tract. Eight
facies were identified in the channel
facies succession; however, a single
channel does not necessarily exhib-
it this complete spectrum of facies.
The facies of this facies succession
are pebbly conglomerate (pbCg), mas-
sive granule sandstone (mSs), granule
442 / Ramon and Fajardo

FIGURE 9. Conventions of sed-

imentary structures used in
the core descriptions (used in
the following figures).

sandstone with upper plane bed stratification (uppSs),
trough cross-stratified granule sandstone (txgSs), trough
cross-stratified sandstone (txSs), burrowed sandstone
with relict trough cross-stratification (btxSs), and
ripple-laminated sandstone (rpSs).
The facies present in a single, aggradational chan-
nel succession are strongly dependent on temporal
and spatial changes in accommodation space and
base level. In general, under overall increasing ac-
commodation conditions, there is a transition from
facies pbCg, mSs, and/or uppSs into txgSs and/or txSs
and then into btxSs, cvSs, and/or rpSs.
A facies substitution diagram for the channel facies
succession is presented in Figure 13. This diagram
condenses the natural succession of facies during an
overall increase in accommodation space for the
channel facies succession cored in the coastal-plain
facies tract of the Mirador Formation.
Pebbly Conglomerate
(pbCg)
Facies pbCg appears very
sparsely at the base of the ag-
gradational channel succes-
sion in relatively thick beds
(as much as 1.2 m [4 ft]) or in
thin lag deposits (as much as
7.6 cm [3 in.]). It is overlain
by facies mSs, uppSs, txgSs, or
txSs. The lower contact of
this facies is a scour surface,
and the upper contact is tran-
sitional or sharp with the
overlying facies. Bed thick-
ness normally is between
5 cm (2 in.) and 0.6 m (2 ft);
however, beds as much as
1.2 m (4 ft) occur. Coarse
clasts are well rounded and
generally pebble size, but
some are granule size. Quartz
clasts are the most common
(more than 95%), and the re-
mainder are lithic clasts. Sedi-
mentary structures are un-
common. Figure 14A shows
a photograph of this facies.
In a stratigraphic context,
this facies is common in relatively low accommoda-
tion conditions, where it constitutes a minor propor-
tion of the aggradational channel facies and occurs
in the thickest beds. In contrast, in high accommo-
dation settings, beds of this facies are uncommon.
Two hydrodynamic interpretations are possible.
When this facies contains either normal or inverse
grading, it is interpreted as deposited by a hypercon-
centrated flow. When grading is absent, it is inter-
preted as rapidly deposited by a turbulent current in
the upper flow regime.
Massive Granule Sandstone (mSs)
This facies occupies the basal part of an aggra-
dational channel succession. This facies is overlain
by facies uppSs, txgSs, or txSs. The lower contact, in
some cases, is a scour surface; in other cases, the con-
tact is sharp or transitional with facies pbCg. The
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 443

FIGURE 10. Core description showing interpreted facies tracts and cycles of the Cupiagua A-1 well. The Mirador
Formation is divided into six intermediate-term stratigraphic cycles and two long-term cycles. (See Figure 9 for legend.)
444 / Ramon and Fajardo

FIGURE 11. Core description showing interpreted facies tracts and cycles of the Cupiagua H-11 well. (See Figure 9 for legend.)
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 445

FIGURE 12. Core description showing interpreted facies tracts and cycles of the Cupiagua C-3 well. (See Figure 9 for legend.)
446 / Ramon and Fajardo
upper contact normally is sharp. Bed thickness com-
monly ranges between 7.6 and 20 cm (3 and 8 in.),
but, in exceptional cases, are as much as 0.6 m (2 ft).
FIGURE 13. (A) Channel facies
succession, example from Cupia-
gua H-11 well. Facies that com-
pose these aggradational units
define an overall increase in ac-
commodation space. (B) Facies
substitution diagram of channel
facies succession. The horizontal
axis represents the probability of
the substitution of facies. The
area occupied by each facies is
directly proportional to the fre-
quency of the facies in the chan-
nel facies succession. (See Figure 9
for legend.)
The quartz arenite is upper
medium to coarse grained with
a few scattered granules. Sedi-
mentary structures are absent,
but normal or inverse coarse-tail
grading uncommonly occurs.
In a stratigraphic context,
this facies is present only in
low accommodation condi-
tions. Two hydrodynamic in-
terpretations are considered.
When normal or inverse coarse-
tail grading is present, this fa-
cies is interpreted as a hypercon-
centrated flow deposit. When
grading does not occur, it is in-
terpreted as rapidly deposited
by a turbulent current in the up-
per flow regime.
Granule Sandstone with
Upper Plane Bed
Stratification (uppSs)
This facies occurs in the
lower part of the aggradational
channel successions and is
overlain by facies txgSs or txSs.
The lower boundary is either
sharp or transitional. The up-
per boundary is always sharp.
Bed thickness commonly ranges
between 0.15 and 0.6 m (0.5
and 2 ft). The quartz arenite is
coarse to very coarse grained,
with scattered quartz and un-
common lithic granules. The planar stratification is
defined by alignment of pebble-size quartz clasts in
horizontal or low-angle (less than 58) laminae.
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 447
In a stratigraphic context, this facies occurs in rela-
tively low accommodation conditions. Facies uppSs
was deposited by a turbulent current in the upper flow
regime.
Trough Cross-stratified Granule Sandstone (txgSs)
This facies constitutes as much as 30% of the channel
facies succession cored in the Mirador Formation. It is
commonly overlain by facies txSs and, in very few cases,
by facies btxSs. The quartz arenite is coarse to very
coarse grained with scattered pebbles. Sometimes, the
granule-size grains are arranged along the foresets.
Thickness of trough cross-stratification sets is variable,
commonly from 0.3 to 1.2 m (1 to 4 ft). Sets are com-
monly of constant grain size, but some sets exhibit a
fining-upward trend. The angle of the foreset lami-
nae range between 20 and 308, and the foreset lami-
nae thickness ranges between 5 and 40 mm (0.2 and
1.5 in.). Pebbles are aligned parallel to the foreset
laminae.
This facies occurs in higher accommodation con-
ditions than the previous facies. The bed set thickness
changes with changes in accommodation space; as
accommodation increases, bed set thickness increases.
It was deposited by a turbulent current in the lower
flow regime. Facies txgSs is the stratigraphic record of
migrating 3-D dunes on the channel floor.
Trough Cross-stratified Sandstone (txSs)
Facies txSs is the most common (as much as 60%)
facies of the channel facies succession. It is overlain
by facies btxSs, cvSs, and rpSs. The quartz arenite is
medium to coarse grained. Fining-upward trends
occur in some vertical successions of sets. Thickness
of trough cross-stratification sets is variable, common-
ly between 0.15 and 0.6 m (0.5 and 2 ft); however,
some sets are strongly amalgamated and range be-
tween 5 and 12.7 cm (2 and 5 in.) thick, whereas other
sets are as much as 1.2 m (4 ft) thick. Foreset lami-
nae are inclined between 15 and 258 and are between
3 and 30 mm (0.118 and 1.18 in.) thick. Foreset lami-
nae commonly are locally broken by burrowing.
Figure 14B is a photograph of this facies.
In a stratigraphic context, this facies was deposited
in higher or the same accommodation conditions as
facies txgSs. The bed set thickness changes as a func-
tion of the accommodation space; as accommoda-
tion space increases, the bed set thickness increases.
Facies txSs was deposited by turbulent currents in
the lower flow regime. The trough cross-stratified
sandstone is the stratigraphic record of migrating 3-D
dunes on the channel floor.
Burrowed Sandstone with Relict
Trough Cross-stratification (btxSs)
This facies occupies the upper part of aggrada-
tional channel successions. This facies overlies facies
txSs and substitutes with facies rpSs. Bed thickness is
sometimes difficult to observe, but, in general, is less
than 0.3 m (1 ft). The quartz arenite is medium to coarse
grained. The most characteristic feature is the burrow-
ing. More than 70% of the original trough cross-
stratification is destroyed, and relict foresets with vari-
able angles of inclination are visible through the
burrowing overprint. In some cases, the rock is com-
pletely burrowed, and some irregular laminations occur.
In a stratigraphic context, this facies occurs in high
accommodation conditions. When the burrow inten-
sity is constant between two scour surfaces, this facies
was deposited under low sedimentation rates. When
burrow intensity decreases from top to bottom in
sandstones between two scour surfaces, it is inter-
preted as fluctuations in rates of accumulation; periods
of bed migration were interrupted by a time of non-
deposition when organisms reworked the sediment.
Ripple-laminated Sandstone (rpSs)
This facies is uncommon in the channel facies suc-
cession cored in the Mirador Formation. Facies rpSs,
when present, appears on top of the aggradational
channel successions. Bed thickness is between 5 cm
(2 in.) and 0.3 m (1 ft). The quartz arenite is lower to
upper medium grained. Occasionally, minor burrow-
ing that partially destroyed the ripple laminae is typical.
In a stratigraphic context, this facies occurs in high
accommodation settings. It is interpreted as a waning
flow cap of the hydrodynamic regime responsible
for the deposition of the channel succession.
CREVASSE SPLAY FACIES SUCCESSION
This facies succession was preserved during rela-
tively higher accommodation conditions than the
channel facies succession. Thin muddy beds of the
swamp and flood-plain facies succession are com-
monly interlayered with this facies succession. Five
facies are identified: sandstone with mudstone rip-up
clasts (rupSs), small-scale trough cross-stratified sand-
stone (stxSs), ripple-laminated sandstone (rpSs),
convolute sandstone (cvSs), and burrowed and irreg-
ular laminated sandstone (blSs). The crevasse splay fa-
cies succession generally does not contain good res-
ervoir rock. For similar porosities, the crevasse splay
sandstones have much lower permeability distribu-
tions as compared to channel sandstones.
448 / Ramon and Fajardo

FIGURE 14. Photos of the different facies present in the coastal-plain facies tract in the Lower Mirador in the Cupiagua
field. (A) Pebbly conglomerate. Clast are well rounded and composed of white quartz and some cherts; (B) trough
cross-stratified sandstone; (C) sandstone with rip-up clasts; (D) small-scale trough cross-stratified sandstone; (E) ripple-
laminated sandstone; (F) convolute sandstone (3805 m [12,484 ft] measured depth); (G) burrowed, laminated
sandstone; (H) laminated mudstone with starved ripples; (I) mottled mudstone; (J) rooted and irregular laminated
mudstone; (K) brecciated mudstone. All cores are 10cm (4 in.) wide.
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 449

FIGURE 14. (cont.).

450 / Ramon and Fajardo

Crevasse splay intervals normally occur between

channel facies successions and swamp facies succes-
sions. This succession of facies tracts records increas-
ing wetness and increasing preservation through
time, beginning with a subaerially exposed surface of
unconformity at the channel base.
Facies successions in crevasse splay complexes
constitute a series of commonly asymmetrical, short-
term, base-level fall hemicycles that contain progres-
sively less crevasse-channel facies and progressively
more distal crevasse splay facies (Figure 15). This
means that initial crevasse splay progradations onto
the flood-plain or ephemeral flood-plain swamps were
replaced by crevasse splay progradations into wetter
flood plains and permanent swamps. Thus, individual
short-term stratigraphic cycles represent episodes of
progradation, but each episode brought progressively
more distal facies into the short-term facies succes-
sion. The short-term stratigraphic cycles are stacked
to form an intermediate-term base-level rise cycle re-
flecting increasing accommodation through time.

Sandstone with Rip-up Clasts (rupSs)

Facies rupSs occurs uncommonly and normally
overlies scour surfaces; in some cases, it is interlay-
ered with facies blSs. The upper contact is either tran-
sitional or sharp. Bed thickness commonly ranges
between 12 and 20 cm (5 and 8 in.); however, beds as
much as 0.6 m (2 ft) occur. This quartz arenite is fine
to lower medium grained. Mudstone rip-up clasts are
as much as 3 cm (1.2 in.) long in dimension and oc-
casionally so abundant that the rock is almost a rip-
up clast conglomerate (Figure 14C). Variable degrees
of burrowing and irregular lamination occur in fa-
cies rupSs.
Where this facies overlies a scour surface, it is in-
terpreted as a crevasse channel lag deposit. The mud
clasts were either ripped up because of channel scour-
ing and incorporated in the flow or were derived from
slumps into the channel and incorporated in the flow.
When facies rupSs does not overlie a scour surface, it
is interpreted as related to an upstream scour event.
Facies rupSs occurs in the proximal parts of a cre-
vasse splay. In a stratigraphic context, it represents
relatively lower accommodation conditions in a cre-
vasse splay succession.

Small-scale Trough Cross-stratified

Sandstone (stxSs)
This facies constitutes as much as 30% of the cre-
vasse splay intervals cored in the Mirador Forma-
FIGURE 14. (cont.). tion. It occurs in proximal portions in a crevasse splay
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 451
complex. The lower contact is either a scour surface
or a transitional surface from underlying ripple-
laminated sandstones. The upper contact is sharp
or transitional. Bed thickness commonly ranges be-
tween 5 and 15 cm (2 and 6 in.); uncommonly, beds
FIGURE 15. Crevasse-splay facies suc-
cession in cycle I of the Cupiagua A-1
well. Individual short-term strati-
graphic cycles represent episodes of
progradation, but each episode brings
progressively more distal facies into
the short-term facies succession. The
short-term stratigraphic cycles are
stacked to form an intermediate-term
base-level-rise cycle reflecting in-
creasing accommodation through
time. (See Figure 9 for legend.)
are thicker than 0.3 m (1 ft). The
quartz arenite is lower to middle
fine grained and well sorted. Small-
scale trough cross-bedding is the
most common sedimentary struc-
ture; however, different degrees of
burrowing partially destroy the fore-
set laminae. Foreset laminae are
inclined 10–208 and are 1–2 mm
(0.04–0.08 in.) thick. The type of
burrowing was not identified, but
there is a low diversity of ichnofos-
sils. Figure 14D shows a photograph
of this facies.
The small-scale trough cross-
stratification is the record of mi-
grating 3-D dunes in both channel-
ized and unconfined unidirectional
flow in the lower flow regime. Bur-
rowing is evidence of fluctuations in
flow strength that allowed rework-
ing of sediments by organisms.
Facies stxSs is interpreted as part of
a crevasse channel deposit where
there is a succession from a scour
surface into facies stxSs, then into
rpSS, and finally, into facies blSs.
When the succession is from facies
blSs into rpSs and into stxSs, this
facies is interpreted as a crevasse
splay prograding into a swamp and
flood plain.
Facies stxSs occupies proximal
portions in a crevasse splay com-
plex. This facies commonly appears in the lower
parts of the crevasse splay intervals cored in the
Mirador Formation. In a stratigraphic context, facies
stxSs represents relatively lower accommodation
conditions in a crevasse splay deposit.
452 / Ramon and Fajardo
Ripple-laminated Sandstone (rpSs)
Facies rpSs constitutes as much as 40% of the cre-
vasse splay facies succession. Upper and lower contacts
of the ripple-laminated sandstone are either sharp or
transitional. Bed thickness generally ranges between
2.5 and 15 cm (1 and 6 in.); however, beds as much as
0.45 m (1.5 ft) thick are found. The quartz arenite is
fine to very fine grained and well sorted (Figure 14E).
Differential degrees of burrowing partially destroy the
ripple laminae.
This facies represents a migration of ripples in the
lower flow regime in either channelized or uncon-
fined flow. This facies appears in two different suc-
cessions in a crevasse complex. It occurs in crevasse
channel deposits, where it is interpreted as a waning
flow cap of the crevasse channel succession. It also
appears in the lower half of shallowing-upward suc-
cessions of prograding crevasse splays and is inter-
preted as deposition in unconfined flow.
Facies rpSs is better developed in more distal parts
of a crevasse splay facies succession than facies stxSs
and rupSs. In a stratigraphic context, those intervals
with major proportion of facies rpSs are interpreted
as deposited during higher accommodation condi-
tions than intervals dominated by facies stxSs.
Convolute Sandstone (cvSs)
This facies is uncommon in the crevasse splay fa-
cies succession. Bed thickness ranges between 5 and
25 cm (2 and 10 in.) (Figure 14F). The quartz arenite
is very fine grained and well sorted. Relicts of ripple
laminae are common.
This facies is interpreted as rapidly deposited sedi-
ments that collapsed because of water escape pro-
cesses. This facies is associated with facies rpSs or blSs.
It was deposited in similar accommodation condi-
tions as facies rpSs.
Burrowed, Laminated Sandstone (blSs)
This facies constitutes as much as 10% of the cre-
vasse splay facies succession cored in the Mirador
Formation. It occupies the distal part of a crevasse
splay complex. Upper and lower contacts are either
transitional or sharp. Bed thickness is most common-
ly between 5 cm (2 in.) and 0.6 m (2 ft). The quartz
arenite is very fine to fine grained and well sorted.
Clay matrix is present in some intervals. The most
distinctive feature of this facies is the irregular ar-
rangement of the lamination caused by a high degree
of burrowing. Sometimes, the rock is completely or
almost completely homogenized by organism rework-
ing. Some intervals with lower degree of burrowing
exhibit carbonaceous or muddy irregular lamina-
tions and relict foreset laminae of short-term trough
cross-stratification (Figure 14G).
Facies blSs was deposited in an oxygenated envi-
ronment where benthic organisms reworked the sedi-
ments that were deposited. It is found in distal parts of
prograding crevasse splays and on top of crevasse chan-
nels. This facies was deposited in relatively higher
accommodation conditions than facies stxSs or rpSs.
SWAMP AND FLOOD-PLAIN
FACIES SUCCESSIONS
The swamp and flood-plain facies successions are
dominated by mudstones, and they generally occur
adjacent to the crevasse splay facies succession. They
are more common toward the top of the intermediate
cycles of the Lower Mirador (Figure 10). Most mud-
stones correspond to wet flood plain and swamps.
Locally, these are replaced transitionally by increas-
ing thickness of soils. An unconformity is interpreted
toward the top of the Lower Mirador in the Cusiana
field. This unconformity is underlain by soils, caliche,
and brecciated mudstones and is overlain by strata,
which accumulated in wet flood-plain, marsh, and
swamp to brackish bay environments (Fajardo, 1995).
Three facies were differentiated in the swamp fa-
cies succession: massive mudstone (mMd), laminated
mudstone with starved ripples (lsrMd), and burrowed
and irregular laminated mudstone (blMd). Three fa-
cies where identified in the flood-plain facies succes-
sion: mottled mudstone (mtMd), rooted and irregular
laminated mudstone (rtMd), and brecciated mud-
stone (brMd).
A facies substitution diagram for swamp and flood-
plain facies succession is presented in Figure 16. This
diagram represents the succession of facies under
decreasing accommodation conditions from swamp
to flood-plain facies succession. The left part of the
diagram represents drier conditions. The right part of
the diagram represents wetter conditions or even
brackish conditions.
Massive Mudstone (mMd)
This facies occurs between crevasse splay deposits
or with other swamp facies. Lower and upper con-
tacts are commonly sharp, where it is interbedded
with crevasse splay facies successions, but where it is
interbedded with other swamp facies associations,
the contacts are generally transitional. Bed thickness
is variable. Where this facies is interbedded with cre-
vasse splay sandstones, the bed thickness ranges from
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 453

2.5 to 12 cm (1 to 5 in.). The color is

commonly dark gray and medium
gray. Some mudstones have high or-
ganic matter content.
Facies mMd was deposited from
suspension in open-swamp and open-
lake conditions. This facies is inter-
preted as the deepest swamp facies.

Laminated Mudstone with

Starved Ripples (lsrMd)
This facies consist of laminated
gray mudstones intercalated with
variable amounts of silt and few very
fine sandstone ripples. Upper and
lower contacts are either sharp or tran-
sitional. When it is interbedded with
crevasse splay sandstones, contacts are
sharp, and beds range between 2.5 and
15 cm (1 and 6 in.) thick. Beds of this
facies in open-swamp intervals range
between 0.3 and 1 m (1 and 3 ft) thick;
in exceptional cases, beds as much as
1.8 m (6 ft) thick occur. Texturally,
this facies varies from silty mud to
muddy silt. It is commonly dark to
middle gray and occasionally green-
ish gray or black. Plane parallel hori-
zontal lamination, starved ripples,
and lenticular and wavy laminations
are the characteristic sedimentary
structures (Figure 14H).
This facies was deposited from
suspension in open-swamp environ-
ments. It is interpreted as having
been deposited at the same depth or
even a shallower depth than facies
mMd. The starved ripples and lentic-
ular and wavy lamination are evi-
dence of weak currents.

Burrowed and Irregular

Laminated Mudstone (blMd)
Lower and upper contacts are
either transitional or sharp. Bed thick-
ness ranges from 0.3 to 1 m (1 to 3 ft).
It is generally composed of muddy
siltstone but may be sandy siltstone.
It exhibits different colors, including
FIGURE 16. Swamp and flood-plain facies successions in the Cupiagua A-1
well. The succession from swamp to flood plain defines an overall increase dark to light gray, greenish gray, and
in accommodation space in intermediate-scale cycle III. Cycle IV is composed beige. Burrowing is the most distinc-
of wetter mudstones of marsh and brackish ponds. (See Figure 9 for legend.) tive feature of this facies. Irregular
454 / Ramon and Fajardo
laminations are frequently present. On top of the
lower Mirador, facies blMd changes transitionally to
mottled and rooted mudstones (mtMd or rtMd).
This facies was deposited from suspension in a
swamp or in ephemeral flood-plain ponds. In a strati-
graphic context, where this facies is interbedded with
facies mtMd or rtMd, it is interpreted as the result
of an overall decrease in accommodation space. In
contrast, where it is interbedded with facies mMd
and lsrMd, it is interpreted as having been deposited
during an overall increase in accommodation space.
Mottled Mudstone (mtMd)
This facies’ most distinctive feature is its varicol-
ored aspect. Upper and lower contacts are transi-
tional. Bed thickness ranges between 0.6 and 2.7 m
(2 and 9 ft). The mudstones of some intervals have a
dark-gray background with irregular reddish-brown
patches; other mudstones are light gray or beige
with reddish and yellowish patches, and others are
green or beige with blue and purple patches. A com-
mon feature is the vertical distribution and orienta-
tion of patches and irregular veins (Figure 14I).
This facies is interpreted as deposited from sus-
pension in a flood-plain environment where there
are alternating dry and wet conditions. In a strati-
graphic context, this facies represents a time of de-
creasing accommodation.
Rooted and Irregular Laminated
Mudstone (rtMd)
This facies is beige and light gray; it contains
black to brown root traces composed of carbona-
ceous material and is interbedded with facies mtMd.
The basal and upper contacts are commonly transi-
tional. Bed thickness ranges between 0.3 and 2.1 m
(1 and 7 ft). Burrows are another feature present in
this facies (Figure 14J).
This facies represents the soil alteration of a pre-
vious facies deposited from suspension. Two inter-
pretations of this facies are possible. Where it is in-
terbedded with facies mtMd, it is interpreted as
deposited in a flood-plain setting where drier condi-
tions predominated. Where it contains organic lami-
nations and it is interbedded with facies mMd, it is
interpreted as deposited in a wetter environment like
an ephemeral pond.
Brecciated Mudstone (brMd)
This facies constitutes a small percentage of the
swamp and flood-plain facies successions. The lower
contact is transitional, and the upper contact is sharp.
FIGURE 17. Facies succession and substitution diagram
of the bay-fill and the channel and bay-head facies tract.
Bay-fill and bay-head cycles tend to have a base-level-fall
asymmetry with a thin base-level-rise cap. (See Figure 9
for legend and text for abbreviations.)
Bed thickness ranges between 1 and 1.5 m (3 and 5 ft).
This facies appears brecciated, and it is composed
of angular, irregular masses of brownish color with
chaotic distribution (Figure 14K). In some intervals,
ferruginous pseudonodules of less than 2 mm (0.08 in.)
are present.
This facies results from modification of previously
deposited mudstones by soil processes. The facies
has strong evidence of subaerial exposure.
Bay Facies Tract
The bay facies tract occurs in the upper Mirador
Formation. In this facies tract are channel, bay-head
delta, and bay-fill facies successions (Figures 10 –12).
A transition from channel to bay-head delta to bay-fill
facies successions represents a continuous increase in
accommodation space. A transition from bay-fill to
bay-head delta to channel facies succession defines
a continuous decrease in accommodation space.
A facies substitution diagram of the bay facies tract
is presented in Figure 17. This diagram was construct-
ed from the bay facies tract intervals cored in three
wells in the Mirador Formation in the Cupiagua field.
CHANNEL AND BAY-HEAD DELTA
FACIES SUCCESSIONS
Channel and bay-head delta facies successions con-
tain the main reservoir rock in the bay facies tract.
These successions are dominant in the lower half of
the upper Mirador in the Cupiagua field area. Three
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 455
facies are recognized in these successions: trough
cross-stratified sandstone (txSs); burrowed sandstone
with relict trough cross-stratification (btxSs); and
ripple-laminated sandstone (rpSs). Figures 10–12 show
examples of these facies successions.
Trough Cross-stratified Sandstone (txSs)
Facies txSs is the most common facies of the bay
facies tract. It occurs both in channel and bay-head
delta facies successions. Upper and lower contacts of
trough cross-stratification sets are commonly sharp;
however, when burrows are present, the contacts are
indistinct. The quartz arenite normally is medium
grained, but coarse-grained intervals occur. Fining-
upward or coarsening-upward trends are recognized
in a vertical succession of sets. Thickness of trough
cross-stratification sets is variable; it ranges from few
inches to as much as 1.2 m (4 ft). Foreset laminae are
inclined between 15 and 258 and are between 2 and
8 mm (0.08 and 0.3 in.) thick. Minor burrowing oc-
curs in this facies. With the exception of uncommon
burrows, this facies is virtually identical to facies
txSs of the coastal-plain facies tract (see Figure 14B).
Figure 18A is a photograph of this facies.
Facies txSs was deposited by turbulent currents in
the lower flow regime. The trough cross-stratified sand-
stone is the stratigraphic record of migrating 3-D dunes
on the channel floor or in bay-head deltas.
Bioturbated Sandstone with Relict Trough
Cross-stratification (btxSs)
This facies is the second most common facies of
the bay facies tract. It occurs in the upper part of
aggradational channel successions and in the lower
part of progradational bay-head delta successions.
Trough cross-stratification set thickness is some-
times difficult to observe but, in general, is less than
0.45 m (1.5 ft). The quartz arenite is medium to coarse
grained. The most characteristic feature is the bur-
rowing. More than 60% of the original trough cross-
stratification is destroyed, and relicts of foresets with
variable angle of inclination are visible throughout
the burrowing overprint (Figure 18B). Macaronichnus,
Ophiomorpha, Gyrolithes, Arenicolites, and crab bur-
rows are recognized.
Where the burrow intensity is constant between
two scour surfaces, this facies was deposited under
low sedimentation rates. Where burrow intensity
decreases from top to bottom, it is interpreted as
fluctuations in rates of accumulation; periods of bed
migration are interrupted by times of nondeposition
when organisms reworked the sediment.
Ripple-laminated Sandstone (rpSs)
This facies is uncommon (2%) in the bay facies tract.
It sometimes occurs on top of channel facies suc-
cessions and uncommonly within bay-fill facies suc-
cessions. In channel successions, the thickness is as
much as 0.6 m (2 ft); in bay-fill successions, it is less
than 20 cm (8 in.). The quartz arenite is fine grained
and commonly contains clay matrix. It exhibits sev-
eral degrees of burrowing, but relict ripples can al-
ways be easily identified.
This facies was deposited in the lower flow re-
gime, and it represents the migration of 3-D rip-
ples. The intensive burrowing suggests low rates of
deposition.
BAY-FILL FACIES SUCCESSION
This facies succession is composed of burrowed
sandstones and mudstones. It is more common in
the upper half of the upper Mirador. Four facies are
defined in this facies succession. From deeper to shal-
lower, those facies are laminated mudstone (lMd),
burrowed and irregular laminated mudstone (blMd),
burrowed and irregular laminated sandstone (blSs),
and vertical burrowed sandstone (vbSs) (see Figure 17).
A transition from lMd to blMd to blSs and, later, into
vbSs represents a continuous decrease in accommo-
dation space (Figure 17). In some cases, facies vbSs is
overlain by facies hbSs; this upward transition rep-
resents an increase in accommodation space.
Laminated Mudstone (lMd)
This facies is uncommon in the bay facies tract. It
occurs at the base of shallowing-upward bay-fill suc-
cessions. Upper and lower contacts are normally sharp.
Bed thickness is less than 1 m (3 ft). The mudstone is
dark gray to black and exhibits discontinuous silt lami-
nae or lenses (Figure 18C).
This facies is interpreted as having been deposited
from suspension. The lamination is evidence of weak
currents. The absence of burrowing indicates that
dysaerobic or anaerobic conditions occurred at the
sediment-water interface.
Burrowed and Irregularly
Laminated Mudstone (blMd)
This facies constitutes less than 4% of the bay facies
tract. It occupies the lower part of the shallowing-
upward bay-fill successions. Upper and lower con-
tacts are either sharp or transitional. Bed thickness is
variable, ranging from few inches to as much as 1.2 m
(4 ft). This facies is composed of dark-gray or black
456 / Ramon and Fajardo

FIGURE 18. Photos of the different facies present in the bay facies tract in the upper Mirador in the Cupiagua field.
(A) Trough cross-stratified sandstone; (B) bioturbated sandstone with relict trough cross-stratification; (C) laminated
mudstone; (D) bioturbated and irregularly laminated mudstone (3797 m [12,457 ft] measured depth); (E) bioturbated
and irregularly laminated sandstone; (F) vertically burrowed sandstone. All cores are 10 cm (4 in.) wide.
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 457

silty mudstone. The more distinctive characteristic is

the intensive burrowing. Burrows are mainly hori-
zontal. In some intervals, Thalassinoides burrows are
recognized (Figure 18D). Irregular and discontinuous
laminations normally occur.
This facies was deposited from suspension in open-
bay conditions. The intensive burrowing implies
optimal substrate conditions for organisms to live.

Bioturbated and Irregularly

Laminated Sandstone (blSs)
This facies constitutes 30% of the bay-fill facies
tract. Bed thickness normally is between 15 cm (6 in.)
and 0.3 m (1 ft); beds as much as 1 m (3 ft) thick occur.
It is composed of fine and very fine sandstone with
variables amounts of clay matrix. It has carbonaceous
or muddy irregular laminations and variable degrees
of burrowing (Figure 18E). The major difference with
facies vbSs is that single ichnofossils are not easily
distinguished. Mudstone rip-up clasts occur in some
intervals. This facies was deposited in open-bay en-
vironments at slow rates, which allowed intense re-
working by organisms.

Vertically Burrowed Sandstone (vbSs)

This facies is the most common of the bay facies
tract. It occupies the upper part of shallowing-upward
bay-fill successions. Upper and lower contacts are
normally transitional with overlying and underlying
facies. Bed thickness is variable, ranging from a few
inches to 2 m (6 ft). The quartz arenite is fine and
sparsely very fine or medium grained. The distinctive
feature is the dominance of vertical burrows, such as
Ophiomorpha, Skolithos sp., and Arenicolithes (Figure 18F).
Horizontal burrows are uncommon and include
Teichichnus, Planolites, and Paleophycus. This facies
lacks ripple lamination or cross-stratification, but ex-
hibits carbonaceous or muddy horizontal wispy lami-
nations. This facies was deposited at higher rates than
facies blMd and blss in open-bay settings.
Facies vbSs was deposited in open-bay environ-
ments at slow rates, which allowed reworking by
organisms.

FOUR-DIMENSIONAL STRATIGRAPHIC
ARCHITECTURE OF THE MIRADOR FORMATION

Three scales of stratigraphic cycles are recognized

in the Mirador Formation. Short-term (high-frequency)
cycles correspond to progradational and aggradational
units. Six intermediate-term cycles were identified by
FIGURE 18. (cont.). the stacking patterns of their component short-term
458 / Ramon and Fajardo
FIGURE 19. Correlation of the cored wells in the Cupiagua field showing the long-term stratigraphic cycles defined in
the Mirador Formation. Depth in feet.
cycles and by the general trend of facies successions,
indicating increasing or decreasing accommodation-
to-sediment supply (A/S) ratios. Two long-term cy-
cles were defined from the stacking pattern of the
intermediate-term cycles and by the general trend of
facies successions (Figure 19).
Stratigraphic cycles, regardless of scale, record a
complete base-level cycle (Barrell, 1917; Wheeler,
1964). During a base-level cycle, accommodation-to-
sediment supply ratio increases (as base-level rises) to
a maximum limit and then decreases (as base-level
falls) to a minimum limit along the entire geomorphic
profile connecting linked depositional environments.
Short-term Stratigraphic Cycles
Short-term stratigraphic cycles are progradational
and aggradational units that conform to Walther’s
law. These short-term cycles are the building blocks
of the stratigraphic framework. Aggradation occurs
by deposition on an almost horizontal surface; exam-
ples of aggradational units are channel, lake, and
flood-plain deposits. Progradation occurs by deposi-
tion on an inclined surface; examples of prograda-
tional units are crevasse splay, bay-head delta, and
bay-fill deposits. A single short-term cycle generally
is composed of both aggradational and prograda-
tional components because of changes in inclination
of depositional surfaces along the geomorphic pro-
file. The definition of short-term cycles is based on fa-
cies successions, occurrence of surfaces of stratigraphic
discontinuity, and measurements of sedimentologic
attributes like bed set thickness or burrow orientation,
diversity, and density.
Short-term stratigraphic cycles in channel facies
successions commonly exhibit base-level rise asym-
metry, and less commonly with a thin base-level fall
hemicycle. Asymmetry means that the time of in-
creasing A/S is preserved as rock, whereas the time of
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 459
decreasing A/S is mainly represented by surfaces of
stratigraphic discontinuity. In the Mirador Forma-
tion, typical thickness of short-term cycles composed
of channel facies successions is 3– 6 m (10 –20 ft), and
the range is 2.1 – 9 m (7 – 30 ft).
When channel successions exhibit enough facies
diversity, the recognition of base-level rise deposits
is based on a vertical change of facies representing
reduction in flow concentration and/or flow strength.
A complete succession would include facies depos-
ited as hyperconcentrated flows, followed by facies
deposited in the upper flow regime, followed by fa-
cies deposited in the lower flow regime. This suc-
cession could be indicated by a transition from granule
sandstone with upper plane bed stratification (uppSs)
facies at the channel base to trough cross-stratified
granule sandstone (txgSs) or trough cross-stratified sand-
stone (txSs) facies, followed by trough cross-stratification
with some burrowing (btxSs), ripple-laminated sand-
stone (rpSs), or convolute-laminated sandstone (cvSs)
facies (Figure 13). The base-level-fall hemicycle is
commonly represented by a scour surface. However,
if it is preserved as rock, it is recognized by facies
successions that indicate an increase in sediment
reworking and amalgamation. A complete succession
would be a transition from facies deposited in the
lower flow regime to facies deposited in the upper
flow regime and even to facies deposited as hyper-
concentrated flows.
Short-term stratigraphic cycles in crevasse splay
facies successions exhibit a base-level fall or base-
level rise asymmetry or are symmetric alternating rise
and fall hemicycles (Figure 15). Thickness of these
short-term cycles in the Mirador Formation ranges
from 0.9 to 5.1 m (3 to 17 ft).
Definition of cycles in crevasse splay facies suc-
cessions is based on the recognition of shallowing-
upward intervals developed during the progradation
of a crevasse splay into permanent or ephemeral
flood-plain lakes. Sometimes, the crevasse facies be-
come less frequent and thinner, and swamp and flood-
plain facies becomes thicker and more frequent.
This transition is interpreted as a rise base-level.
Short-term stratigraphic cycles in lake and flood-
plain facies successions are more symmetrical than
in any other facies succession of the coastal-plain
facies tract. Thickness of Mirador lake and flood-
plain short-term cycles ranges from 0.6 to 4.8 m (2 to
16 ft). The definition of short-term stratigraphic cy-
cles in intervals composed only of swamp and flood-
plain facies successions is based on the interpreted
wetness and dryness of the different facies that occur
in the interval. From deep to shallow, the swamp
facies succession is from massive mudstone (mMd),
to laminated mudstone with starved ripples (lsrMd),
to burrowed and irregularly laminated mudstone
(blMd), to gray mudstones with roots to mottled
mudstones. This transition from swamp (wet) to dry
flood plain is interpreted as a rise in base level.
If flood-plain facies successions occur without in-
tervening swamp and lake beds, short-term cycles
are defined by the degree of soil development. A suc-
cession from mtMd to brecciated mudstone (brMd)
facies indicates a base-level-rise hemicycle. In inter-
vals composed only of mtMd facies, short-term cycles
are defined by changes in the degree of mottling. An
increase in mottling (increased duration of soil-forming
processes) is interpreted as a base-level rise, and a
decrease in mottling is interpreted as a base-level fall
as more sediment is aggraded over the flood plain.
Short-term stratigraphic cycles in the bay facies
tract were identified in channel, bay-head delta, and
bay-fill facies successions. A single short-term cycle
can be composed of only one or a combination of
these facies successions.
Channel and bay-head delta facies successions com-
monly coexist at each geographic location in a single
short-term cycle. Cycles are base-level rise or fall asym-
metrical, or they are symmetrical. Channels are pre-
served during both short-term base-level rise and fall.
In contrast, bay-head deltas only occur during short-
term base-level-fall time. Bay-head deltas are probably
replaced by bay-fill facies successions during short-
term base-level-rise time. Cycle thickness is variable:
cycles composed of channel and bay-head delta suc-
cessions range from 3.3 to 9 m (11 to 30 ft), but cycles
composed of either one or the other facies succession
range from 1.2 to 3.6 m (4 to 12 ft).
A short-term base-level-rise hemicycle in channel
facies successions is indicated when a channel scour
surface is overlain by trough cross-stratified sand-
stones (txSs), which constitutes the major part of the
channel deposit and is followed by burrowed sand-
stone with relict trough cross-stratification (btxSs) and,
in some cases, is capped by rippled sandstone (rpSs).
The increase in burrowing toward the top and the
decrease in flow velocity reflect increasing accom-
modation and gradual drowning of the channel.
Short-term base-level-fall hemicycles in bay-head
delta successions are recognized from facies succes-
sions and changes in cross-stratification set thick-
ness. Bay-head delta successions exhibit both tran-
sitional and sharp lower contacts with underlying
facies. A typical facies succession is a change from
460 / Ramon and Fajardo

FIGURE 20. (A) Channel and bay-head

delta facies successions, Cupiagua
A-1 well. Trough cross-stratified sand-
stone in the lower part of bay-head delta
succession is intensively burrowed, and
the burrowing decreases upward. The
core photograph (B) shows the same
interval. Note the better oil saturation
toward the top of the core (right), indi-
cating better reservoir quality. (See
Figure 9 for legend.) Depth in feet.

burrowed sandstone (bSs) at the base

to burrowed sandstone with relict
trough cross-stratification (btxSs) and
then to trough cross-stratified sand-
stone (txSs). Burrowing intensity di-
minishes toward the top, and trough
cross-stratification set thickness in
facies txSs thins upward. These facies
successions record an increase in the
hydromechanical energy from base
to top and increased cannibalization
of migrating dunes during progra-
dation, which are interpreted as an
upward decrease in accommodation
(Figure 20).
Short-term stratigraphic cycles in
bay-fill facies successions tend to
have base-level-fall asymmetry with
a thin base-level-rise cap. Typical
cycle thickness ranges from 1.2 to
3.6 m (4 to 12 ft), but in excep-
tional cases, cycles are as much as
5.4 m (18 ft) thick. Short-term cycles
commonly are recognized from fa-
cies successions of shallowing-up-
ward bay-fill deposits. In a typical
shallowing-upward bay-fill succes-
sion, there is a transition from bur-
rowed and irregularly laminated
mudstone (bMd) at the base to bur-
rowed and irregularly laminated
sandstone (blSs), followed by vertical
burrowed sandstone (vbSs) toward
the top. This succession represents a
continuous decrease in accommo-
dation-to-sediment supply ratio
caused by progradation. The bay-
head delta facies succession some-
times overlies bay-fill successions in
a single short-term cycle (Figure 20).
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 461
Thin base-level rise caps are recognized in bay-fill
deposits. They exhibit a succession from shallower to
deeper facies, from vbSs to blSs facies, and eventually,
to bMd facies. The short-term cycles have a thicker
base-level-fall hemicycle with thin base-level-rise caps.
Intermediate-term Stratigraphic Cycles
Six intermediate-term stratigraphic cycles are de-
fined in the Mirador Formation in the Cupiagua field
(Figures 10 – 12, 19). Intermediate-term stratigraphic
cycles are defined from the stacking pattern of their
component short-term cycles, from the general trend
of facies successions, and from surfaces of strati-
graphic discontinuity or dislocation. These cycles are
designated with Roman numerals (I–VI). Facies as-
semblages of the three basal cycles (I, II, and III) be-
long to the coastal-plain facies tract. Cycles IV, V, and
VI are composed of bay facies tracts.
Cycle I
Cycle I is only cored in the Cupiagua A-1 well. This
cycle exhibits mostly flood-plain and crevasse facies
tracts. This core shows a transition from stacked
flood-plain facies of the Los Cuervos Formation into
crevasse channel and splay facies with some swamp
and wet flood-plain facies successions. This trend of
facies successions suggests a base-level fall to rise
progression (Figure 10). Well-log correlation indicates
that laterally, the crevasse sequences change into
channel-belt sandstones (Figure 21). These channel-
belt sandstones tend to be single story and laterally
discontinuous.
The transition from aggradational flood-plain de-
posits into crevasse sandstone bodies indicates the
progradation of the crevasses over a distal part of the
flood plain. The crevasse splay facies succession at
the upper half of this cycle occurs in six short-term
cycles (Figure 15). Short-term cycles cored in Cu-
piagua A-1 well consist of more distal crevasse splay
facies, so each of these cycles steps landward (away
from the channel belt) with respect to the underly-
ing one defining an overall base-level rise in the
upper part of this intermediate cycle.
Fajardo (1995) showed that this lowermost inter-
mediate cycle in the adjacent Cusiana field is com-
posed of channel to crevasse to flood-plain facies
tracts. His cycle starts with the erosion at the base of
the channel belt and ends with the next channel belt
base. This is interpreted as a base-level-rise cycle.
Fajardo (1995) made an isopach map of his cycle I,
which shows that the cycle occurs only in the south-
ern part of the Cusiana field. Cross sections across the
Cusiana field show that the proportion of swamp and
flood-plain facies successions increases from south-
east to northwest, and the proportion of crevasse
splay facies successions decreases in this direction.
The lower proportion of the channel facies tracts
with respect to the crevasse and flood-plain facies
tracts in the Cupiagua field area suggests that depo-
sitional dip orientation was toward the northwest.
This is consistent with the trend of paleovalleys
mapped by Fajardo (1995).
Cycle II
The lower boundary of intermediate-term cycle II
is a base-level-fall surface. This cycle exhibits a base-
level-rise asymmetry. Cycle II consists of channel and
crevasse splay facies successions. The flood-plain fa-
cies succession is only found in a couple of wells.
Cycle II has a transition from channel to crevasse
splay facies successions (Figure 10), which, in some
cases, are overlain by the swamp and flood-plain fa-
cies succession. This trend of facies successions de-
fines an intermediate-term base-level rise. Short-term
cycles in channel and crevasse splay intervals have a
general landward-stepping pattern in the base-level-
rise hemicycle.
Channel facies successions are dominant in the
lower half of cycle II. Facies that compose these cycles
have a general decrease in amalgamation and an in-
crease in bedform preservation from base to top. Cre-
vasse splay facies successions occur in the upper half
of the intermediate-term base-level cycle II (Figure 10).
These short-term cycles are successively composed
of deeper or more distal crevasse splay facies; hence,
these short-term cycles have a landward-stepping stack-
ing pattern that defines an increasing accommoda-
tion regime.
Intermediate-term cycle II has a higher propor-
tion of channel facies tracts than cycle I. In addition,
the channel facies tracts are thicker and show a much
higher degree of amalgamation. Channel sandstone
bodies are laterally more continuous (Figures 19, 21).
These indicate that cycle II records a seaward step-
ping of the depositional systems of the alluvial plain.
In Cupiagua, this cycle has the thickest proportion of
granule facies. Also, channels at the base of this cycle
are more amalgamated than cycle I (Fajardo, 1995). This
evidences also a seaward-stepping stacking pattern.
Cycle III
This intermediate-term cycle shows the highest
degree of facies variability. In the three wells (A-1, C-3,
and H-11; Figures 10–12) that cored this interval, the
 462 / Ramon and Fajardo

FIGURE 21. Well-log correlation of the A-1, Q-6, and U-23 wells of the Cupiagua field. Cycle I shows lateral facies changes. Crevasse and swamp and flood-plain
deposits cored in A-1 change laterally into channel-belt sandstones. Depth in feet.
Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 463

FIGURE 21. (cont.).

464 / Ramon and Fajardo
cycle has diverse proportions of channel, crevasse,
and flood-plain facies tracts. Cycle III is more sym-
metrical than cycles I and II, but symmetry changes
from well to well. In the A-1 well, the cycle is com-
posed primarily of agradational flood-plain facies
tracts. The base-level-rise hemicycle is composed of
crevasse splay (20%) and swamp-flood-plain (80%)
facies successions (Figure 10). The base-level-fall hemi-
cycle is composed of swamp (65%), flood-plain (15%),
and channel (20%) facies successions. In the C-3 well,
the cycle consists mostly of channel facies successions
(Figure 12). Minor intercalations of channel aban-
donment and crevasse facies successions occur. Well
H-11 is composed of channel (60%), crevasse (20%),
and swamp (20%) facies tracts (Figure 11). The base-
level-rise hemicycle is characterized by a transition
from channel to crevasse splay to swamp and flood-
plain facies successions (Figure 11). This trend of
facies successions defines an overall increase in
accommodation space. The intermediate-term base-
level-fall hemicycle is characterized by a transition
from flood-plain to crevasse to channel facies succes-
sions. This trend defines an overall decrease in accom-
modation/sediment supply ratio. The turnaround point
from intermediate-term base-level rise to fall is located
within the massive mudstone facies of the swamp
and flood plain.
Channel-belt sandstones of this intermediate-term
cycle III are laterally discontinuous and less amal-
gamated as compared to the ones in cycle II. The
highest proportion of flood-plain and crevasse facies
tracts and the single-story character of the channel-
belt bodies indicate that cycle III represents a landward
stepping into more distal, lower energy facies tracts.
Cycle IV
Intermediate-term cycle IV marks a major change
in the depositional setting in the Cupiagua field
area. Cycle IV is composed of channel (30 –60%),
bay-fill (0 10%), and bay-head delta (30 –70%) facies
successions. The lower boundary is generally a base-
level-fall surface, whereas the upper boundaries are
base-level-rise flooding surfaces. This cycle commonly
exhibits base-level-rise and base-level-fall hemicycles
of similar thickness. Mudstones of cycle IV are com-
monly black or dark gray (Figure 18D) and are bur-
rowed as compared to mudstones in cycle III, which
exhibit light- to medium-gray, brown, and purple
mottling (Figure 14J, K). The change from mottled
mudstones to dark-gray mudstones is inferred to re-
flect a change from drier, soil-forming conditions to
wet coastal-plain or bay conditions.
Although facies changes from well to well, the
base-level-rise and base-level-fall hemicycles are pres-
ent in all wells. Well A-1 records the base-level rise as
the transition from channel to bay-fill and the base-
level fall with the change from bay-head delta back
into estuarine-channel facies tracts (Figure 10). Well
H-11 records the rise-to-fall cycle as the transition
from estuarine-channel to bay-head and back into
channel facies tracts (Figure 11). Finally, well C-3
shows a transition from channel to bay-head facies
tracts (Figure 12). The presence of new more distal
brackish to shallow-marine facies tracts indicates
that cycle IV records a major landward stepping of
all facies tracts. The alluvial facies tracts dominant
in the Lower Mirador (cycles I, II, and III) shifted
toward the east at the time of cycle IV.
Cycle V
Intermediate-term cycle V records a continuous
landward shift of facies tracts. This cycle is com-
posed by bay-head delta (40 –70%) and bay-fill (30–
60%) facies successions. Channel facies tracts are
only locally present toward the top of the cycle. The
lower and upper boundaries are turnaround points
(fall to rise) in conformable strata and are located at
the position of maximum thickness of channel and
bay-head delta facies successions. Cycle V is gener-
ally symmetrical.
The base-level-rise hemicycle of cycle V is char-
acterized by a transition from bay-head delta to bay-
fill facies successions. Short-term cycles show landward-
stepping stacking patterns. Progressively, bay-head
facies becomes thinner and more bioturbated, indi-
cating preferential preservation of the distal portions.
In the same direction, bay-fill facies tracts become
thicker. Bay-fill mudstones are better developed to
the middle part of the cycle. This cycle has the highest
proportion of distal bay-fill facies tracts. This suggests
a continuous landward-stepping pattern.
The base-level-fall hemicycle of cycle V records the
transition from bay-fill to bay-head delta facies suc-
cessions, which is sometimes capped by channel facies
successions. This trend of facies successions indicates
decreasing A/S. Short-term cycles have a seaward-
stepping stacking pattern. This transition represents a
shallowing-upward profile produced by progradation.
Cycle VI
This intermediate-term cycle is asymmetrical where
mostly the base-level rise is preserved. This cycle is
only cored at well A-1. The lower boundary is located
within conformable strata at the turnaround point
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 465
from decreasing to increasing accommodation-to-
sediment supply ratio. Cycle VI is composed of bay-
fill (40– 70%) and bay-head delta (30 –60%) facies
successions.
In the rise hemicycle, the proportion of deeper facies
progressively increases, and shallower facies decreases
within short-term cycles. This upward increase in the
proportion of deeper facies means that short-term
cycles have a landward-stepping stacking pattern.
In few wells, the base-level-fall hemicycle is made
up of short-term cycles with a general seaward-stepping
stacking pattern. The upper half of cycle VI shows chan-
nel facies successions directly over the bay-fill facies
tract. This would indicate an anomalous seaward shift
in facies (downhill offset in facies). Above the channel
facies tract, a rapid landward shift in facies without
stratigraphic discontinuity occurs. This landward off-
set places deeper bay-fill facies of the C8 member of the
Carbonera Formation on top of channel deposits.
The higher proportion of channel facies tracts in
cycle VI relative to cycle V suggests a seaward-stepping
stacking pattern. Above the seaward facies offset, the
landward-stepping stacking pattern is maintained up
to the top of the Mirador Formation. Both the sea-
ward and landward shift in facies are interpreted as
anomalies in the stacking pattern and are not con-
sidered boundaries of intermediate-term stratigraphic
cycles. These stratigraphic anomalies were identified
in the cored well (A-1, Figure 10) and might be present
in uncored wells.
RESERVOIR ZONATION OF THE
MIRADOR FORMATION
High-resolution stratigraphic correlation, defined
as establishing temporal equivalency of rock units re-
gardless of rock type, mineralogy, texture, or environ-
ment of deposition is the base for a more accurate and
predictive reservoir characterization. Because the ba-
sic lithological, petrophysical, geometrical, and res-
ervoir continuity attributes of rocks originate during
sediment accumulation, accurate correlation permits
accurate representation of these rock properties in a
four-dimensional (time-space) context. High-resolution
correlation is the best means of identifying stratigraph-
ically controlled fluid-flow compartments, compart-
ment boundaries, and fluid-flow pathways in reser-
voirs. This is because the natural bounding surfaces of
episodically accumulated stratigraphic successions
compartmentalize strata into time-bounded packages
of varying spatial and temporal scales. These natural
stratigraphic boundaries commonly coincide with the
most laterally continuous shifts of facies, and conse-
quently, the most pronounced changes in lithological
properties commonly occur across those bounding sur-
faces. Thus, petrophysically significant boundaries of
fluid-flow compartments commonly are coincident
with time-significant stratigraphic bounding surfaces.
Numerous occurrences of natural stratigraphic bound-
aries are not associated with petrophysical boundaries;
but occurrences of petrophysical boundaries that are
not associated with stratigraphic boundaries are far
less common (Cross et al., 1993).
A second stratigraphic attribute that can contribute
to reservoir analysis is the relation between dynamic
changes in accommodation/sediment supply ratio
and the resulting systematic changes in stratal ar-
chitecture and diversity, successions, and continuity
of facies preserved in the stratigraphic record. For
example, Allen (1978, 1979) and Bridge and Leeder
(1979) suggested that fluvial channel-belt sands de-
posited during conditions of slow subsidence (equiv-
alent to low accommodation) form vertically and
laterally interconnected, blanketlike sandstone bod-
ies that could function as single compartment res-
ervoirs in terms of fluid flow and pressure. By con-
trast, channel-belt sandstones of identical depositional
systems may occur as isolated, stringerlike reservoir
sandstones if deposited during periods of faster sub-
sidence (equivalent to high accommodation). Each
sandstone body would function as a separate fluid-
flow compartment.
The Lower Mirador cycles clearly show these dif-
ferences in channel-belt architecture. Intermediate
cycles I and III deposited under relatively higher ac-
commodation conditions have thinner channel-belt
sandstone bodies associated with thick crevasse and
flood-plain facies tracts. Channel-belt sandstone bod-
ies are laterally discontinuous between wells and show
strong thickness changes. Facies tract proportions
change from well to well, but crevasse and flood-
plain facies tracts are generally thicker than in cycle II
(Figures 10, 19). As observed in Figure 10, the Cu-
piagua A-1 well has only a thin channel sandstone in
cycle I. Close wells show fining-upward, bell-shaped
gamma ray response that is interpreted as channel
successions (Figure 21). Stratigraphic position and
thickness of the channel facies successions change be-
tween wells. In contrast, channel-belt sandstone bod-
ies are thicker and laterally more continuous in cycle II
(Figures 10, 11). The relative proportion of overbank
facies tracts in cycle II is lower as compared to cycles I
and III. Channel sandstone bodies are laterally more
466 / Ramon and Fajardo
continuous and are present in almost every single
well in the Cupiagua field (Figures 19, 21, 22).
Finally, facies tracts control the quality of the fluid-
flow properties of the sandstone bodies. Porosity
versus log permeability as a function of facies suc-
cessions were plotted for intermediate-term cycles I
to III (Figure 23). This plot shows that channel facies
successions have higher porosity and permeability
values than do crevasse splay facies successions in
intermediate-term cycles I, II, and III of the Lower
Mirador. These figures show two distinctive popula-
tions: one with higher porosity and permeability,
which corresponds to the channel facies succession,
and the other with lower porosity and permeability,
which corresponds to the crevasse splay facies succes-
sion. Because both channel and crevasse splay facies
succession intervals are laterally continuous through-
out the field, these differences in petrophysical prop-
erties become critical in the reservoir zonation of the
lower Mirador long-term cycle.
Similar differences are found in the bay facies tract
of the upper Mirador (cycles IV– VI; Figure 24). Chan-
nel and bay-head facies successions have higher per-
meability values for the same range in porosity as
compared to the bay-fill facies successions. The plot
shows that for a given porosity, the channel and bay-
head sandstones have as much as two orders of mag-
nitude higher permeabilities than the equivalent
FIGURE 22. Well-log correlation of the wells in the southern part of the Cupiagua field. Depth in feet.
bay-fill facies successions. In single wells, the differ-
ent facies tracts control the porosity and permeability
vertical distribution. Well H-11 has massive sand-
stones with only minor variations in the gamma-
ray response over most of the Mirador Formation
(Figure 25). Interpreted facies tracts correspond to
major changes in the porosity and permeability val-
ues. These different facies tracts then define differ-
ent units with different fluid-flow capabilities. The
crevasse-splay facies successions typically have po-
rosities in the 3–4% range and permeabilities lower
than 0.1 md. Bay-head facies successions typically
exhibit porosities in the 5–7% range and permeabil-
ities of tens of milidarcys.
Water saturation data from porous plate and Dean
Stark analysis performed in core plugs indicate the
existence of different rock types. Channel of the Lower
Mirador and channel and bay-head facies successions
of the upper Mirador show water saturation values
ranging from 5 to 7%. Crevasse and bay-fill sand-
stones show water saturation values higher than 17%.
FLUID-FLOW UNITS AND STATIC MODEL
OF THE MIRADOR FORMATION
Previous reservoir characterization studies identi-
fied four mega-rock-types for the three producing
reservoir sandstones (Guadalupe, Barco, and Mirador
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 467

FIGURE 23. Plot of porosity versus log permeability as a function of facies successions for the coastal-plain facies tract
(intermediate-term cycles I to III).

FIGURE 24. Plot of porosity versus log permeability as a function of facies successions for the bay facies tract
(intermediate-term cycles IV to VI).
468 / Ramon and Fajardo

FIGURE 25. Vertical variation of porosity and permeability in the Cupiagua H-11 well.

formations): mudstone (nonreservoir), lithoarenites, erate reservoir quality. For each of these reservoir
quartzarenites, and phosphatic lithoarenites. The rock types, a porosity-permeability transform was ob-
lithoarenites have poor reservoir quality; the quart- tained by comparing core with wire-line data. For static
zarenites and phosphatic lithoarenites have mod- modeling purposes, to capture the reservoir quality
 Sedimentology, Stratigraphy, and Architecture of the Eocene Mirador Formation / 469
variability in the quartzarenites, a total of four facies
associations and their four porosity-permeability trends
were used to cover the entire rock quality range of this
mega-rock-type (Figures 23, 24). The facies associa-
tions identified in cored wells were extrapolated to
the rest of the wells using electrical log signatures
(gamma ray, spectral gamma ray, neutron, density,
and resistivity). The interwell static modeling was
generated using geostatistics techniques to populate
the facies, followed by the porosity and permeability
based on the facies tracts.
Net pay is defined at a permeability cutoff equal or
higher than 0.1 md, which is a 3.5% porosity and
about 388 API gamma-ray cutoff for the Mirador For-
mation quartzarenites. These cutoffs have been eval-
uated at different scales from pore scale (mercury
injection data) to centimeter scale in Repeated For-
mation Test flow response, to meter scale using core
fluorescence presence ending with unit flows, pro-
ducing intervals confirmed by production logging
tools (PLTs) response.
As mentioned before, facies associations were pop-
ulated within the gross rock volume using stochastic
methods. The 3-D stochastic reservoir model was built
using the statistics of rock type distribution gained
from the Cupiagua core suite and semiregional depo-
sitional model. Porosity and water saturation distri-
butions were assigned to each of the rock types in the
gross rock volume.
Average porosity and permeability values in the
geocellular static model are consistent with the well
average values in each formation and during the up-
scaling process to generate the reservoir-simulation
model; those averages were maintained representa-
tive of the well average values.
The stock tank barrels of original oil initially in place
(STOOIP) volumes were obtained from the initializa-
tion of the dynamic reservoir model, and they are good
representations of the static stochastic 3-D model.
ACKNOWLEDGMENTS
We are grateful to BP for the support to this work
and permission to publish. The authors thank Steven
Bachtel, Paul (Mitch) Harris, and an anonymous re-
viewer for their helpful suggestions.
REFERENCES CITED
Allen, J. R. L., 1978, Studies in fluviatile sedimentation:
Bars, bar complexes and sandstone sheets (low sin-
uosity braided streams) in the Brownstones (L. Dev.),
Welsh borders: Sedimentary Geology, v. 26, p. 281 –293.
Allen, J. R. L., 1979, Studies in fluviatile sedimentation:
An exploratory quantitative model for the architecture
of avulsion-controlled alluvial suites: Sedimentary Geol-
ogy, v. 21, p. 129–147.
Barrell, J., 1917, Rhythms and the measurement of geo-
logic time: Geological Society of America Bulletin, v. 28,
p. 745 – 904.
Bridge, J. S., and M. R. Leeder, 1979, A simulation model of
alluvial stratigraphy: Sedimentology, v. 26, p. 617 – 644.
Cooper, M. A., et al., 1995, Basin development and tec-
tonic history of the Llanos basin, eastern Cordillera,
and Middle Magdalena Valley, Colombia: AAPG
Bulletin, v. 79, no. 10, p. 1421 – 1443.
Coral, M., and W. Rathke, 1997, Cupiagua field, Colombia:
Interpretation case history of a large, complex thrust
belt gas condensate field: Cartagena, Memorias del VI
Simposio Bolivariano ‘‘Exploración Petrolera en las
Cuencas SubAndinas,’’ Colombia, Tomo 1, p. 119 – 128.
Cross, T. A., et al., 1993, Applications of high-resolution
sequence stratigraphy to reservoir analysis, in R. Eschard
and B. Doligez, eds., Subsurface reservoir characteri-
zation from outcrop observations: Proceedings of the
7th Exploration and Production Research Conference:
Paris, Teichnip, p. 11 – 33.
Dengo, C., and M. Covey, 1993, Structure of the eastern
Cordillera of Colombia: Implications for trap styles
and regional tectonics: AAPG Bulletin, v. 77, p. 1315 –
1337.
Fajardo, A. A., 1995, 4-D stratigraphic architecture and 3-D
reservoir fluid flow model of the Mirador Fm., Cusiana
field, foothills area in the Cordillera Oriental, Colom-
bia: M.Sc. thesis, Colorado School of Mines, Golden,
Colorado, 171 p.
Martinez, J., 2003, Modelamiento estructural 3D y aplica-
ciones en la exploración y explotación de hidrocar-
buros en el Cinturón de Cabalgamiento del Pie-
demonte Llanero, Cordillera Oriental, Colombia: VIII
Simposio Bolivariano — Exploración Petrolera en las
Cuencas Subandinas.
Villamil, T., 1999, Campanian – Miocene tectonostratig-
raphy, depocenter evolution and basin development
of Colombia and western Venezuela: Paleogeography,
Paleoclimatology, Paleoecology, v. 153, p. 239 – 275.
Wheeler, H. O., 1964, Baselevel, lithosphere surface, and
time-stratigraphy: Geological Society of America Bulletin,
v. 75, p. 599 – 610.