shale exploration

Unconventional Methods For Unconventional Plays:

Using Elemental Data To Understand Shale Resource Plays
Ken Ratcliffe and Milly Wright, Chemostrat; Dave Spain, BP USA

Part 2

(Pearce et al., 2005, Ratcliffe et al., 2010), but it

also forms the basis for calculating mineralogical
compositions from elemental data. Simplistically,
where an R2 value between a mineral and
an element exceeds 0.8, the linear regression
between the two can be used to model mineral
abundances. For example on Figure 1A, the
regression equation y=0.47x+2.97 (where y is
the concentration of Al2O3 [known] and x is
the amount of illite [unknown]) can be used to
determine illite abundances in samples where
Al2O3 concentrations have been determined,
but where XRD data have not been acquired.
However, the regression lines are formation
specific and more often than not, the relationship
is more complex than a simple high R2 linear
relationship. There are several public domain
software packages that exist for automatically
calculating mineralogy from geochemistry
(ModAn, Sednorm and Minlith) (Paktunc, 2001,
Rosen et al., 2004), which form the basis for
in-house algorithms used by Chemostrat. Figure
2 displays a comparison of directly measured
mineralogy obtained by XRD analyses and
calculated mineralogy, using an in-house mineral
modelling. It is clear that for most of the common
minerals, the calculated mineralogy closely
matches the directly measured mineralogy.
While this method may not be able to determine

Introduction
In the last edition of PESA News Resources (No
116), the use of chemostratigraphy to help refine
stratigraphic understanding of shale resource
plays was discussed. Following on from Part 1,
we will here look at what additional information
can be gathered about shale reservoirs from
the same elemental dataset that provides
chemostratigraphic correlations.

Mineral modelling in the

Haynesville formation
An important aspect to understanding shale
reservoirs is determining their mineralogy and
TOC (total organic carbon) content, since these
will essentially control reservoir quality. Typically,
this is achieved using X-Ray diffraction (XRD) and
LECO TOC analysis respectively. XRD, in particular,
is expensive and results can take a considerable
time to acquire. From Figure 1, however, it is
clear that a close relationship exists between
major elements and common rock forming
minerals. This relationship is to be expected, since
the primary control on whole rock inorganic
geochemical data is the component mineralogy

y=0.47x+2.97
R2=0.92

25

Na2O(%)

20

Al2O3 (%)

In addition to mineralogy, it is possible to

approximate TOC levels using geochemical data.
Figure 1E shows the relationship between Mo
concentrations and TOC values. When samples
with over 10 ppm Mo are removed, the two
variables have an R2 values of 0.88, i.e. a highly
significant relationship. The samples with Mo
values greater than 10 ppm also have high
values of Ni and V, which suggests that the
high Mo in these samples are associated with
minerals derived from authigenic enrichment
under anoxic conditions (sensu Tribovillard et
al., 2006, see discussion below), rather than a
direct relationship to organic carbon. Using
the regression equation shown on Figure 2E
(TOCcalc = 0.3325*Mo +0.3108), the TOCcal values
shown on Figure 2 are generated. Apart from
samples with Mo >10 ppm, there is a close
correspondence between TOCmeasured and TOCcalc.
This may not be as precise as direct measurement
of TOC using LECO, but it provides a quick and
inexpensive reasonable approximation of TOC.

B
CaO (%)

more subtle mineralogical variations, such as

the illite-smectite abundances, it does provide
a reasonable indication of bulk mineralogy and
this is achieved quickly and inexpensively from
the dataset that has been obtained primarily for
stratigraphic correlations.

1.5

R =0.89
2

R2=0.85

1.4

20

1.3

15% CaO = 21% calcite

15

1.2

15

1% Na2O = c. 7.5% plagioclase

1.1
1.0

10
10

0.9
0.8

0.7
5

10

20

30

40

50

10

15

20

Illite (%)

30

35

Calcite (%)

0.6

10

12

Plagioclase (%)

E
TOC (%)

10

Fe2O3 (%)

25

R =0.90
2

8% Fe2O3 = 17% Chlorite

8

3.5
3.0
2.5
2.0

6
1.5
1.0

0.5
2

10

15

20

Chlorite (%)

0.0

R2=0.88
y = 0.3325x + 0.3108
0

10

15

20

25

Mo (ppm)

Figure 1. 1A-D: Cross plots of selected elements and minerals, with regression coefficient (R2) displayed. Mineral data were acquired using XRD. 1E: Cross-plot of Mo
and TOC values, with R2 and regression line equation displayed.FIGURE
TOC values1obtained from LECO analysis. All ICP, XRD and LECO TOC analyses were carried out on a
subsample of homogenised powder.
April/May 2012 | PESA News Resources |

55

shale exploration
3 shows that within parts of the Haynesville
formation, EF far in excess of 1 are seen for Mo, U,
Ni and V, all elements that becomes authigenically
enriched in sediments deposited under anoxic
conditions. Figure 3 also demonstrates in that
EFs in Package 4 are all low, implying conditions
where oxic to suboxic during the deposition
of the Bossier formation, whereas EF values in
Package 2 are >1 implying anoxic conditions
during deposition of the Haynesville formation,
with notably high values recorded toward the
bottom of Package 2 (unit 2.1).

Paleoredox modelling in the

Haynesville formation
Understanding paleoredox conditions is of
paramount importance to shale gas exploration
because high TOC values are typically found
in sediments deposited where bottom water
condition were anoxic. Oceanic anoxic events
have long been recognised and studied
(Schlanger and Jenkyns, 1976) and in recent years,
much has been written on the use of elemental
geochemistry in sediments and water columns
as a proxy for depositional redox conditions (e.g.
Tribovillard et al., 2006, Turgen and Brumsack,
2006, Tribovillard et al., 2008, Negri et al., 2009,
Jenkyns, 2010).

Watson-4 and Gaslpie Ocie-10, EF values

are lower in Package 2 than they are in
Johnston Trust-1-2H and Elm Grove Plantation,
implying lateral changes in anoxia. Figure 4
displays the lateral variations in average EF of
vanadium for the entire Haynesville formation
(Chemostratigraphic Package 2) in all study
wells. It is apparent from this figure that during
deposition of the Haynesville formation there
was greater authigenic enrichment of V in the
eastern parts of the basin than in the westerly
areas. This implies that there was greater anoxia
in the east and therefore greater potential for
organic matter preservation.

From Figure 3 it is also apparent that in wells

George T.W. A-8H

As demonstrated in Part 1 of this article, a group

of elements (Mo, U, Ni, V, Cu, Zn and Co) all plot
in close proximity to one another on a cross plot
of Eigen vector 1 vs Eigen vector 2, implying
that they are all somehow relate to one another
within the sediments. These elements are typically
associated with authigenic enrichment within the
sediment under anoxic conditions (Tribovillard
et al., 2006) and as such are potentially important
indicators of anoxia.

Depth (ft)

0
0

10900

Quartz XRD

Illite XRD

Calcite XRD

Chlorite XRD

80 0
Quartz Calc.
80 0

60 0
Illite_calc.
60 0

50 0
Calcite calc.
50 0

25 0
15 0
Chlorite calc.
Plagioclase calc.
25 0
15 0

Plagioclase XRD

Pyrite XRD

TOC

10 0
Pyrite calc.
10 0

TOC_calc.
-

5
5

10950

11000

11050

When dealing with element enrichments, the

level of enrichment is typically expressed by
comparison against standard marine shales.
These are referred to as enrichment factors (EF)
and are calculated as follows (Tribovillard et al.,
2006): ElementEF = (Elementsample/Al203sample) /
(ElementSMS / Al2O3SMS), where EF = enrichment
factor, SMS = standard marine shale (SMS
values taken from Tribovillard et al., 2006). Once
calculated, EF values of 1 indicate no enrichment
of an element relative to a standard marine
shale that was deposited in oxic conditions,
whereas values significantly greater than 1 imply
the element has been enriched although the
mechanism for that enrichment will vary. Figure

11100

11150

11200

11250

Mo >10ppm

11300

11350

Figure 2. Vertical distribution of minerals in well George T.W. A-8H. Bar charts are data derived from XRD and
LECO TOC analyses. Line graphs are mineral data calculated from elemental data.

FIGURE 2

10600
2.4

11300

2.3

2.4

10650

Gamma
0

API 250 0

EF Mo

EF U

EF Ni

- 100 0.5 -

31

EF V
5 0.5 -

11650

Chemostratigraphic
Units

31

Elm Grove Plantation-63

EF V
5 0.5

API 250 0

EF Mo

EF U

EF Ni

- 100 0.5

31

EF V
5 0.5

Pa c k a g e 3

2.2

2.4
11100

11150

2.2
11800

Gamma
0

11050

2.3

2.3

2.2

2.1
2.1

2.1

10850

Package
1

2.1
10800

Package 0

Package 0 Package 1

Pa c k a g e 1

11850

11200

LA
TX

11500

11750

10750

11400

11450

10700
2.2

Package 2

Package 2

11350

EF U

EF Ni

- 100 0.5

Depth (ft)

2.4

11700

2.3

EF Mo

Chemostratigraphic
Units

Depth (ft)

API 250 0

Chemostratigraphic
Packages

Gamma
0

Package 2

5 0.5 -

Package 4 Packages

11600

Pac kage 3

31

Depth (ft)

Package 2

- 100 0.5 -

Glaspie Ocie GU-10

EF V

Chemostratigraphic
Units

API 250 0

EF U

EF Ni

Chemostratigraphic
Packages

Chemostratigraphic
Units

EF Mo

Package 0

11250

Package Chemostratigraphic
Packages
4

11200

Gamma

Pac kage
3

Depth (ft)

Chemostratigraphic

Johnson Trust 1-2H

Watson-4

CLAIBORNE

Elm Grove Plantation 63

Watson GU 4

Glaspie GU 10
Johnson Trust 1-2H
0

10

SCALE IN MILES

Figure 3. Selected element enrichment factors (EF) for wells Watson-4, Glaspie Ocie GU-10, Elm Grove Planation-63 and Johnson Trust 1-2H. (The inset map shows the
position of these wells). EFs for U(ppm), V(ppm), Mo(ppm) and Ni(ppm) have been the plotted for the study intervals in these wells, with Package 2 (the chemically
defined Haynesville formation) correlated for stratigraphic context.
56

| PESA News Resources | April/May 2012

shale exploration

LA

TX

1.65
1.
15

1.4

0
1.6

40
1.

Anoxic

Oxic / dysoxic

5
1.6

1.1

65
1.

1.4

Typically, an upward coarsening interval is

used to define an upward shoaling sequence,
terminating in a maximum regressive surface
(MRS) and an upward fining interval is used to

5
1.65

EFVanadium

1.1

High
Enrichment

15
1.

0
1.40

Low
Enrichment

10

SCALE IN MILES

= Well control locations

Figure 4. Map showing the average values for the enrichment factor of vanadium (EFV). Darker shading
FIGURE
represents a higher average EFV
value for4the Haynesville formation (Package 2). Red circles mark points of well
control in the area used to build this map. Additionally, the Louisiana/Texas border has been superimposed for
geographic context.

(avg.)

Terrigenous
content

Elm Grove Plantation # 63

10850

Zr/Nb (avg.)

18

SiO2/Al2O3
2
5 Terrigenous
SiO2/Al2O3
2

(avg.)

content

11050

11600

11100

Package 2

10750
11650

11150

11700

11200

Package 1

10800

Zr/Nb
- 18

11550

11750

Package 0

10700

Package 2

10650

Depth
(ft)

1st Order T-R

clay packages

Depth
(ft)

2nd Order T-R

Williams 22 3

10600

Zr/Nb
- 18

Zr/Nb (avg.)

18

SiO2/Al2O3
2
5 Terrigenous
SiO2/Al2O3
2

(avg.)

content

1st Order T-R

SiO2/Al2O3
2

2nd Order T-R

18

clay packages

Pac kage
3

SiO2/Al2O3
2
5
-

Package 2

Zr/Nb (avg.)

1st Order T-R

Zr/Nb
9
- 18

2nd Order T-R

Depth
(ft)

clay packages

Glaspie

Package 1

Superimposed on these 1st order variations

defined by terrigenous content are a series of
shorter-term cyclical changes in Zr/Nb and SiO2/
Al2O3 values (Figure 5). The SiO2/Al2O3 and Zr/
Nb chemical logs display notable similarities on
Figure 5, suggesting they are both controlled by
the same processes. SiO2/Al2O3 typically reflects
the abundance of silt-grade quartz in sediments
and as such is normally a reliable grain size
indicator. However, as discussed below, SiO2 can
also be associated with biogenic quartz in shale
gas plays. The similarity of Zr/Nb chemical logs
to SiO2/Al2O3 logs in the Haynesville suggests
it could provide be a proxy for grain size. Zr
is normally associated with detrital zircon in
sediments, while Nb is often associated with clay
minerals, typically illite. Therefore, the Zr/Nb ratio
is a good grain size indicator that is free from
influence of biogenic silica.

and MFSs are defined using Zr/Nb values. The

correlation on Figure 5 implies that regressive
portions of the cycles are better developed in the
west than in the east, where terrigenous content
is high. Transgressive portions of the cycles are
better developed in the east where terrigenous
content is low, but organic content is high and

1.1

As demonstrated in the preceding section, the

elemental geochemistry of the Haynesville
formation is to a large extent controlled by
authigenic enrichment from sea water. However,
there is undoubtedly a component of the
geochemistry that is associated with land-derived
debris. The amount of terrigenous input can be
modelled from the major element geochemistry
by combining those major elements associated
with Group 3 on the Eigen vector plot presented
in Part 1 (Figure 5 Part 1) of this article. On Figure
5, terrestrial input is calculated by summing
Al2O3, TiO2, Na2O and K2O values; SiO2 is not
incorporated in this summation, since it can
also be associated with biogenic quartz (see
succeeding discussion). When terrigenous
content is plotted against depth, it is apparent
that there is a gradual increase from local minima
at the base of the Haynesville / Smackover
transition (Package 1) to a maximum value 2/3
of the way through Package2 (=Haynesville
formation). From this point, the terrigenous
content remains high within Package 2,
decreasing into Package 3 and increasing again
sharply at the base of Package 4 (=the Bossier
formation [not shown]). These long-term trends in
terrigenous content are interpreted to reflect the
proximity of the paleoshoreline; high terrigenous
content occurs when the paleoshoreline extends
into the basin, i.e. during lowstand periods and
low terrigenous contents are found when the
shoreline has been pushed landward by high
base level. The long term changes in terrigenous
contents seen in Figure 5 are therefore considered
to reflect 1st order sea level fluctuations.

define a transgressive sequence, culminating in

a maximum flooding surface (MFS) (Donovan,
2010, Embry, 2010 and references cited therein).
Therefore, the highest values of Zr/Nb (=coarsest
sediment) correspond to an MRS and lowest
values to a MFS. Figure 5 displays the resultant
correlation between select wells when MRSs

Package 0

Recognition and application of

changes in terrigenous input

Deepening
Shallowing
Max regresssive surface
Max flooding surface
Alternating 2nd Order T-R cycle sets

Figure 5. Terrigenous content, SiO2/Al2O3 and Zr/Nb chemical logs plotted for selected wells to highlight long
term (1st Order) and shorter term (2nd Order) cyclical
fluctuations
in the geochemical data.
FIGURE
5
April/May 2012 | PESA News Resources |

57

shale exploration

As mentioned above, SiO2 is not used in

terrigenous input calculations. This is because
some shale gas reservoirs such as the Muskwa
and the Woodford formations contain high
biogenic quartz contents. Recognition of zones
with high biogenic quartz is critical to reservoir
quality in shales effecting porosity, brittleness
and log response. Although the amount of
biogenic quartz cannot be quantified using
whole rock geochemical data, samples that
contain a significant component of biogenic
silica can be identified from geochemical data
using relative proportions of SiO2 and Zr (Figure
6). SiO2 is proportional to the amount of quartz
in the shale and Zr is associated with the heavy
mineral zircon. Zircon represents a proxy for all
silt-sized terrestrial input and, as such, displays a
linear association with detrital quartz (terrestrial
trend on Figure 6). However, in samples where
SiO2 increases, but Zr decreases, i.e. a negative
trend between quartz and terrestrial input, it
can be implied that biogenic rather than detrital
quartz is providing the SiO2 (biogenic trend
on Figure 6). Haynesville formation samples
all plot with a terrestrial trend, implying that
biogenic quartz is not present in the Haynesville
formation. However a significant number of
samples from the Muskwa formation plot on
the biogenic trend, indicating samples from that
formation that contain a significant portion of
biogenic quartz.

Comparison of the Haynesville

formation and the Eagle Ford
formation
Most of the discussions in Part 2 of this article
have concentrated on the Haynesville formation.
Figure 7 uses a series of binary diagrams to
highlight that not all shales are the same.
Figures 7a and 7b show that the Haynesville
formation has generally higher Si2O/Al2O3 values
than the Eagle Ford formation, suggesting that
it is relatively quartz-rich. Figure 7c shows a
clear positive linear relationship between Zr
and SiO2, indicating that the quartz from both
the Haynesville formation and the Eagle Ford
formation is dominantly detrital in origin.
Figure 7b also demonstrates that overall the
Eagle Ford formation contains more CaO,
implying it is a more calcareous shale. XRD data
reveal that average calcite in the Eagle Ford
formation is 33% (unpublished data) and is 10%
in the Haynesville formation.

58

| PESA News Resources | April/May 2012

Figure 7e demonstrates that the Haynesville

formation has generally higher Fe2O3 values
than the Eagle Ford formation, which implies
it contains more chlorite. XRD data reveal that
average chlorite content of the Eagle Ford
formation is 4% (unpublished data) and is 7%
in the Haynesville formation, supporting the
assumption made from the elemental data.

pathways in horizontal multi-lateral wells;

3. Recognition of coarsening and fining
upward sequences which in some basins
(such as the Haynesville-Bossier) may be
interpreted in the context of transgressive
and regressive cycles within a sequence
stratigraphic model;
4. A means to quickly model bulk mineralogy;
5. A means to quickly model TOC contents;
6. Recognition of zones that contain biogenic
silica; and
7. A means to model the relative degree of
anoxia at the time of sediment deposition.
By providing all these lines of information from
a single, relatively inexpensive, rapidly acquired
dataset (including data capture at well-site),
the whole rock geochemical data are shown
to be a highly versatile and cost effective
tool for the exploration, development and
production of shale resource plays. Although
we also highlight differences between shale
formations here, with careful calibration
and data manipulation, the applications
demonstrated here can be readily exported to
ay shale resource play, of any age from around
the world.

Figure 7f shows that in the Haynesville formation

there is a broad linear relationship between
Al2O3 and V in samples where V is less than 120
ppm, which implies that the V is associated
with terrestrial input. However, the Haynesville
formation samples with V values over 120
ppm show no linear association with Al2O3,
implying that the high V values reflect authigenic
enrichment in anoxic
conditions. V values
Muskwa Formation
in the Eagle Ford
100
formation are generally
above 120 ppm and
Biogenic trend
do not display any
linear relationship with
80
Al2O3, which implies
that the V values in the
Eagle Ford are almost
solely associated with
60
authigenic enrichment
in anoxic conditions.

SiO2 (%)

Recognition of zones containing

biogenic silica

Figure 7d demonstrates that, largely, the

Haynesville formation has higher Na2O values
than the Eagle Ford formation, which implies
that it contains more plagioclase feldspar. XRD
data reveal that average plagioclase content of
the Eagle Ford formation is 1% (unpublished
data) and is 7% in the Haynesville formation,
supporting the assumption made from the
elemental data.

40

Conclusions
In Parts 1 and 2 of
this article, it has
been shown that the
50 element data set
acquired when carrying
out chemostratigraphy
on shales provides:
1. A means to
devise regional
stratigraphic
frameworks that can
be used for basin
modelling;
2. A means to
devise local,
high resolution
characterisations
that can be used to
determine well-bore

Terrestrial trend

Zr(ppm)

50

100

150

200

Haynesville Formation
SiO2 (%)

anoxia was most severe. It also suggests that

at least one regressive cycle is absent, or much
thinned in the east of the basin.

100
80
Terrestrial trend

60
40
20
0

Zr(ppm)

50

100

150

200

Figure 6. Zr vs. SiO2 binary plots for data from the Muskwa and Haynesville
formation. The positive linear tend line is termed the terrestrial trend and refers
to samples where SiO2 is derived from a terrestrial source. The negative linear
FIGURE
trend line is termed the biogenic
trend and6refers to samples where a significant
amount SiO2 is derived from biogenic sources.

shale exploration

Jenkyns, H.C. 2010. Geochemistry of oceanic

anoxic events. Geochemistry Geophysics
Geosystems, v. 7, p.1-30.
Negri, A., Ferretti, A., Wagner, T. and Meyers,
P.A. 2009. Organic-carbon-rich sediments
through the Phanerozoic; processes,
progress, and perspectives. Palaeogeography,
Palaeoclimatology, Palaeoecology, v. 273, p.
302-328.
Paktunc, A.D. 2001. MODAN; a computer
program for estimating mineral quantities
based on bulk composition; Windows version.
Computers & Geosciences, v. 27, p. 883-886.
Pearce, T.J., Wray, D.S., Ratcliffe, K.T., Wright, D.K.
and Moscarello, A. 2005. Chemostratigraphy
of the Upper Carboniferous Schooner
formation, southern North Sea. Carboniferous
hydrocarbon geology: the southern North Sea
and surrounding onshore areas. In: Collinson,
J.D., Evans, D.J., Holliday, D.W. and Jones N.S.
(eds) Carboniferous hydrocarbon geology: the

| PESA News Resources | April/May 2012

Turgeon, S. and Brumsack. H.J. 2006. Anoxic

vs. dysoxic events reflected in sediment
geochemistry during the Cenomanian
Turonian Boundary Event (Cretaceous) in the
UmbriaMarche Basin of central Italy. Chemical
Geology v. 234 p. 321-339.

Schlanger, S.O. and Jenkyns, H.C. 1976.

Cretaceous oceanic anoxic events: causes and

(b)

20

SiO2/Al2O3

Al2O3 (%)

(a)
25

Eagle Ford Fm
low SiO2/Al2O3

15

Haynesville Fm
low CaO

4
Haynesville Fm
high SiO2/Al2O3

10

10

20

30

40

50

60

70

Eagle Ford Fm
high CaO

80

10

20

30

40

50

CaO (%)

SiO2 (%)

(d)

Zr (ppm)

(c)
250
Haynesville and Eagle Ford fms
terrigenous SiO2

200

25

15

100

10

50

10

20

30

40

50

60

70

eagle Ford Fm
low Na2O

20

150

Haynesville Fm
high Na2O

0
0.0

80

0.5

1.0

1.5

SiO2 (%)

(f)

20

15

20

10

Haynesville Fm
high Fe2O3

120 (ppm) V

25

15

Haynesville Fm
lowFe2O3

10

2.0

Na2O (%)

(e)
25

10

Fe2O3 (%)

Linear
relationship

100

No Linear
relationship

200

Haynesville Formation
Eagle Ford Formation

300

400

500

V (ppm)

Figure 7. A-F: Cross plots of selected elements displayed for datasets from Haynesville (white squares) and Eagle
Ford (black squares) formations.

FIGURE 7
60

Tribovillard, N., Bout-Roumazeilles, V., Algeo,

T., Lyons, T.W.; Sionneau, T., Montero-Serrano,
J.C., Riboulleau, A. and Baudin, F. 2008.
Paleodepositional conditions in the Orca Basin
as inferred from organic matter and trace metal
contents. Marine Geology, v. 254, p. 62-72.

Rosen, O.M., Abbyasov, A.A.; Tipper, J.C. 2004.

MINLITH; an experience-based algorithm
for estimating the likely mineralogical
compositions of sedimentary rocks from bulk
chemical analyses. Computers & Geosciences,
v. 30, p. 647-661.

Al2O3 (%)

Embry, A.F. 2010. Correlation siliciclastic

successions with sequence stratigraphy. . In:
Ratcliffe, K.T. and Zaitlin B.A. (eds) Application
of Modern Stratigraphic Techniques: Theory
and Case Histories. SEPM Special Publication
No. 94 p. 5-33.

Tribovillard, N., Algeo, T., Lyons, T.W. and

Riboulleau, A. 2006. Trace metals as paleoredox
and paleoproductivity proxies; an update 2006.
Chemical Geology, v. 232, p. 12-32.

Ratcliffe, K.T., Wright, A.M., Montgomery, P.,

Palfrey, A., Vonk, A., Vermeulen, J. and Barrett,
M. 2010. Application of chemostratigraphy to
the Mungaroo formation, the Gorgon Field,
offshore Northwest Australia. APPEA Journal
2010 50th Anniversary Issue p. 371 385.

References
Donovan, A., 2010. The sequence stratigraphy
family tree: understanding the portfolio of
sequence methodologies. In: Ratcliffe, K.T.
and Zaitlin B.A. (eds) Application of Modern
Stratigraphic Techniques: Theory and Case
Histories. SEPM Special Publication No. 94 p.
5-33.

consequences. Geol. Mijnb., v. 55, p. 179-194.

Al2O3 (%)

The authors would like to extend our

gratitude to David Wray and Lorna Dyer of
The University of Greenwich at Medway in the
UK for preparing and analysing the samples
on the ICP OES and MS. We would like to
thank BP for its support and input throughout
this study and for allowing data from wells
Watson-4, Glaspie Ocie GU-10, George T.W.
GUA-8H and CGU 13-17 to be presented. We
would also like recognise PetroHawk and St
Marys for allowing their cores (from wells Elm
Grove Plantation-63 and Johnson Trust 1-2H,
respectively) to be sampled for these analyses
and for permission to utilise their cores and to
present these results. Finally, we are grateful
to Chemostrat for allowing us the time and
providing the support needed to prepare the
article.

southern North Sea and surrounding onshore

areas. Yorkshire Geological Society, Occasional
Publications, Series, 7, p. 14764.

Al2O3 (%)

Acknowledgements