Geophysics

Mateeva_edit.qxp

10/4/07

11:24 am

Page 38

Accurate Estimation of Subsalt Velocities Using Virtual Checkshots

a report by

A l b e n a M a t e e v a , A n d r e y B a k u l i n , P a t s y J o r g e n s e n and J o r g e L o p e z
Shell International Exploration and Production, Inc.

Extracting velocity profiles from checkshots under complicated

Figure 1: Checkshot in Complex Medium

overburden is challenging because: (i) first arrival waveforms may

(a) Classical checkshot

be distorted and difficult to pick; and (ii) the fastest arrival may

(b) Virtual checkshot

not come along the shortest path between the source and
the receiver (see Figure 1a). Both of these problems could
be alleviated by placing the source in the well, so that the travelpath of the first arrival is short and close to a straight line (see
Figure 1b). In practice, it is difficult to place a physical source
downhole. However, it is easy to create a virtual source (VS) in
the well from walk-away (WAW) or 3-D VSP data. Measuring the
first arrival traveltime from a VS to a number of receivers below it,
we can construct a virtual checkshot that is insensitive to
overburden complexity.
An extension of this idea would be to create a virtual shear source in

Figure 2: Virtual Source Concept and Computation

the borehole and construct a shear virtual checkshot. Conventional

checkshots measure only P-wave velocity from first arrivals. Shear

Sk

waves, even if emitted from the surface source or generated by P-S

conversions in the overburden, arrive later and are difficult to

S k

unravel and pick. However, we could harvest P-S conversions in the

overburden to fuel a shear virtual source in the borehole. The first
arrival from such a source would be an S-wave (easy to pick) and
yield a shear-velocity profile.

Complex overburden

Virtual source
(at R )

Well
S k

Simpler middle
overburden

We test these ideas on a data set from a sub-salt prospect in the

deepwater Gulf of Mexico (GOM). We obtain profiles of P- and S-wave

Target

velocities that are in very good agreement with sonic logs under salt at

D (t) = N = 1 Sk (-t) S k (t)

more than 7km depth.

Figure 3: Walk-away VSP Acquisition

The VS method takes data recorded from surface sources into

Virtual Source
Air guns

a source in the borehole, at an existing receiver location. The

conversion is entirely data-driven we do not need to have a
velocity model of the medium to create VS data. All we have to do
to create a fit-for-purpose P VS at receiver and record it at

20,000

data that would have been recorded in the same receivers if we had

Depth (ft)
10,000

down-hole receivers (i.e. WAW or 3-D VSP) and converts them to

Receivers

receiver (see Figure 2):

Note: arrows from source to receiver represent data, not rays.

Take a trace from shot k (at the surface) to receiver and gate
the first arrival (the gate size and position do not matter much as

Sum cross-correlations over all surface shots.

long as it captures most of the down-going energy of the desired

mode in this case P-wave).

Input Data for Virtual Checkshots

Our example dataset is a conventional WAW VSP acquired in a

Cross-correlate this gate with the whole trace from shot k to

vertical well through a massive salt body in the deepwater GOM.

receiver (note that cross-correlation is equivalent to convolution

The survey was shot in four passes of the same shot line, with

with a time-reversed series).

four receiver tool settings, giving a total of 96 receiver depths

(4x24). About two-thirds of the receivers were in the salt, near

Repeat for all surface shots.

38

its base, and the rest were below the salt (see Figure 3). The

TOUCH BRIEFINGS 2007

Mateeva_edit.qxp

10/4/07

11:28 am

Page 39

receivers were interlaced to provide an effective spacing of 50ft.

However, we found some systematic time shifts between the
acquisition lines. For this reason, we performed our velocity
estimation using only receivers from the same tool setting (exactly
100ft apart).
P-wave Virtual Checkshot
To create a P-wave VS we used as input the vertical (Z) component of
the VSP data. A VS was created at every receiver depth. A common VS
gather with a P-wave Virtual Source located at the top-most receiver

...salt velocity estimates

can be useful in seismic
processing as sonic logs
are typically not
acquired in salt and
conbentional checkshots
can be very sensitive to
distortions from top of
salt geometry.
location is shown on the left in Figure 4. The first arrival on the VS data
is clear and easy to pick. The zone of interest in our example is below

www.dresser-rand.com

EVERY DATUM
COMPRESSOR
COMES WITH A
COMMITMENT.

the salt. The first arrivals on sub-salt receivers from virtual sources gave

Performance and Reliability. Our DATUM

a Vp profile that is in very good agreement with the smoothed sonic

compressors are inspiring a lot of loyalty. And

no wonderthey consistently demonstrate
superior performance.

log (see Figure 5).

For this particular dataset the conventional checkshot to sub-salt
receivers matched the smoothed log equally well because the base of
salt is quite flat, allowing all first arrivals to follow the same sub-salt
path (essentially vertical along the well). So, strictly speaking, the Pwave virtual checkshot was not a necessity at this location we
derived it to confirm that the VS would give an accurate Vp profile
below salt. The P-wave virtual checkshot would be of greater
importance in areas with more complicated overburden where
checkshot, log and seismic velocities often disagree (due to sloping or
rough salt boundaries).
S-wave Virtual Checkshot
Our example VSP was acquired with airguns in the water they

They are a leader in the industry in efficiency

and pressure rise per case. And the modular
design of the DATUM compressor allows easy
maintenance and increased availability
making it the most advanced turbocompressor
for the oil, gas, and process industries.
Along with great performance, youll get the
D-R heritage of continuous innovation and
worldwide, total life-cycle support. Not to
mention the fastest cycle times in the industry.
For more information contact D-R today.

do not excite shear waves directly. Nevertheless, a shear VS can

be created by harvesting P-S conversions at heterogeneities
above the receivers. The beauty of the method is that we do
not need to know where in the overburden the conversions
occurred. They may occur at many places at once and
be very complicated the VS method collapses them all to a
useful shear signal radiated from the shear VS as a simple
zero-phase wavelet.
Even though we need not know where the P-S conversions were

The Americas (713) 467-2221

European Served Areas 33-2-35-25-5225
Asia-Pacific 603-2093-6633
info@dresser-rand.com

Mateeva_edit.qxp

13/4/07

12:40 pm

Page 40

Accurate Estimation of Subsalt Velocities Using Virtual Checkshots

Figure 4: Common Virtual Shot Gather with a P (left) and an S (right)

Virtual Source

shear VS data are shown in Figure 4.

Picking the first arrival on sub-salt shear VS traces, we obtained the

24
48
Receivers in salt

72
96 1
24
48
Subsalt
Receivers in salt

72
96
Subsalt

sub-salt Vs profile shown in Figure 5. It matches the smoothed shear

sonic log very well.

`S

200

Salt Velocities

Time (ms)
400

focus was on the sediments below the salt rather than on

600

Since virtual checkshots measure interval velocities, our primary

average P- and S-wave velocities in salt. We opted for measuring

the relatively homogeneous salt. For the sake of completeness,

however, we also used the virtual checkshots to measure the

800

the average as opposed to a detailed velocity profile in salt,

mainly because it is a common practice to assign a single velocity
value to salt bodies for seismic processing. Also, since the salt
Figure 5: Virtual Checkshots (P and S) versus sonic logs (original and
smoothed to the resolution of the VSP)

is much faster than sediments, interval velocities (average over

100ft) would have much larger uncertainties than interval velocities
below salt.

6,300

A linear regression through the VS first arrivals in salt gave Vp =

14,660 330ft/s and Vs = 8,390 350ft/s. Smoothed logs over the

6,350

same depth interval (1,600ft) gave Vp=14,650 60ft/s, Vs = 8,340

35ft/s where the error bars reflect inhomogeneity rather than
measurement uncertainty. So, once again, virtual checkshot and well

6,400

velocities are in excellent agreement.

6,450

These salt velocity estimates can be useful in seismic processing as

sonic logs are typically not acquired in salt and conventional
checkshots can be very sensitive to distortions from top of

6,500
Depth (m)

salt geometry.
6,550

Conclusion
The
6,600

virtual

checkshot

technique

circumvents

fundamental

difficulties faced by conventional checkshots under complicated

overburden and allows us to estimate shear interval velocities in
addition to compressional velocities, even when the source at the

6,650

surface emits only P-waves. The method is capable of delivering very

accurate estimates of P- and S-wave velocities at a great depth in

6,700

and under salt. The more complicated the overburden, the better
and more valuable the virtual checkshot. This new technique is not

6,750

restricted to vertical boreholes it can be used to accurately

evaluate along-the-well velocities for deviated boreholes in the

1,000

1,500

2,000

2,500

3,000

3,500

presence of any heterogeneity and anisotropy.

V (m/s)

Virtual checkshots use only the first arrivals of VS data while the VS
Vp from virtual source
Vs from virtual source
Vp sonic
Vp estimates:
Mean (VS smoothed log) = 0%
Std (VS smoothed log) = 2%
Max (VS smoothed log) <5%

Vs sonic
Vp sonic (30m smoothing)
Vs sonic (30m smoothing)
Vs estimates:
Mean (VS smoothed log) = 1%
Std (VS smoothed log) = 5%
Max (VS smoothed log) = 10%

data contain a wealth of arrivals. The full VS wavefield has many

promising applications (e.g. in imaging) and, therefore, it may soon be
a standard practice to create P- and S-wave VS data from every new
WAW or 3-D VSP acquired. Virtual checkshots should become
inexpensive by-products.
To reap the benefits that the VS method (including Virtual checkshot)

generated, one may still wonder where the strongest ones come

can provide, we recommend designing VSP surveys with at least one

from. In this example, as perhaps in any other salt example, the

walk-away line and proper shot and receiver sampling.

strongest P-S conversion occurs at the top of salt. We verified this by

40

raytracing. To create the shear VS, we used the inline horizontal (X)

Acknowledgements

component of the VSP, muting it before the top-salt P-S arrival to

We thank TGS for the permission to show the surface seismic data in

remove substantial P remnants from the first arrival. The resulting

Figure 2 and Shell for allowing us to publish this paper.

EXPLORATION & PRODUCTION OIL & GAS REVIEW 2007 OTC EDITION