1

Seismic anisotropy measurement from both seismic and VSP

data around the Pierce diapirs, Central North Sea.
2

Keith Hawkins , Tom Armstrong , Steve Gouldesborough , Steve Merlin and Nicky Hutchinson
1
Enterprise Oil plc, St. Magnus House, Guild Street, Aberdeen AB11 6NJ, UK
2
Veritas DGC Ltd., Crompton Way, Manor Royal Estate, Crawley, West Sussex, RH10 2QR, UK

Summary
Poor seismic data quality and significant well misties in excess of 100ms adjacent to salt diapirs on 3D
post-stack depth migration (post-SDM) data in the Pierce Field led to seismic re-processing with
anisotropic pre-stack depth migration (pre-SDM). Here we describe how we initially determined the
presence of seismic anisotropy within the overburden using legacy VSP data and how this was
subsequently confirmed during the prestack depth migration process.
Introduction and Nature of the Problem
Pierce Field lies in Blocks 23/22a and 23/27 of the UK central North Sea, and comprises two
accumulations, Pierce North and Pierce South, associated with twin salt diapirs. The salt diapirs have
created steeply-dipping Palaeocene reservoirs truncated by the salt. Development drilling in 1998-99
relied on a reservoir depth model based on the interpretation of 3D seismic data acquired in 1992 (dual
source, triple streamer) and processed to 3D post-SDM in 1993.
Seismic reflectors commonly overlap with conflicting dips in the target zone and data quality is
severely degraded adjacent to the salt diapir. Steeply dipping reflectors show spatial aliasing.
Significant well misties between 100ms and 260ms for target horizons exist adjacent to the salt diapirs.
A contributing factor to these misties is believed to be mis-positioning of steeply-dipping seismic
events during post-SDM using an isotropic velocity model. The magnitude of vertical and lateral
positioning errors due to ignoring overburden anisotropy can be significant (errors of order 300m are
reported, eg. Hawkins et al. 2001; Vestrum & Lawton, 1999; Vestrum, Lawton and Schmid, 1999).
In view of these imaging and positioning problems associated with the 1993 post-SDM data, a preSDM project was initiated. Was it worth including seismic anisotropy?
Overburden Velocities
Velocity issues included seismic anisotropy, shallow velocity anomalies (Quaternary canyons, shallow
gas, and gas escape features?), fast-velocities adjacent to diapirs due to structural uplift of sediments,
and the presence and location of the salt. Given the potential impact of seismic anisotropy for both
lateral and vertical positioning of seismic reflectors, made more severe by the steep sedimentary dips
exceeding 50 close to the salt diapirs, we needed to determine whether or not this was likely to be a
significant factor in the Pierce overburden.
Banik (1984) measured seismic anisotropy in the North Sea by comparing well-log sonic data and the
interval velocity profile obtained from the surface seismic data (and also from a comparison of the
seismically predicted depth and the well-log depth). He attributed the seismic anisotropy primarily to
the presence of shales. However, the reliability of seismically-derived conventional stacking
velocities over Pierce in the presence of both shallow velocity anomalies (Al-Chalabi, 1979;
Armstrong et al., 2001) and steeply-dipping beds was a concern. Therefore, we investigated the
suitability of existing legacy VSP data to estimate the degree of seismic anisotropy in the
overburden.

Seismic anisotropy inferred by VSPs

There are hundreds of VSP surveys in the North Sea. Most of these are little more than over-sampled
check shot surveys. However, the imaging problems associated with some reservoirs has led to more
exotic VSPs being acquired such as walkaway and fixed offset VSPs. The inclusion of
anisotropy in seismic imaging has led to these exotic VSPs having new added value. Here we bring to
life some of this legacy VSP data to help investigate overburden anisotropy and to confirm our
anisotropy measurements from conventional streamer data.
A number of fixed offset VSPs were acquired in some of the Pierce area exploration wells. Figure 1
shows the shooting geometry for three VSP experiments carried out in well 23/27-9. The salt
proximity survey was carried out in order to define the salt-sediment interface of the nearby diapir. All
raypaths encounter the salt and this invalidates their use in quantifying the seismic anisotropy of the
adjacent overburden. Ray paths for the rig source and NE fixed-offset source surveys traverse the
overburden in different directions. In the presence of seismic anisotropy, the velocity along these slant
paths will be different. Ignoring refraction through the overburden, the difference in velocity along
straight slant paths could give an indication of the strength of seismic anisotropy (this also assumes
that the overburden traversed by raypaths for each survey does not contain lateral velocity variations or
discrete velocity anomalies). This analysis would be valid only if the raypaths did not include travel
through the salt. Ray tracing of the rig source survey carried out by SSL (the VSP contractor) showed
that there were no direct arrivals for the deepest levels due to refraction through the salt.
Figure 2 shows the average velocity of the NE fixed-offset versus rig source surveys for straight slant
paths into each downhole geophone location. Note that, except for the shallowest depths, the average
velocities calculated for the rig source survey are greater than those calculated for the NE fixed-offset
survey. For the shallow recordings above 3500 feet, the angles of incidence to the layering are largest
for the NE fixed-offset survey. However, in the presence of a 45 geological dip encountered beneath
3500 feet, the angle of incidence varies between only 0 and 11 for the NE fixed-offset survey and
between 30 and 44 for the rig source survey. Assuming that the symmetry axis of any seismic
anisotropy is perpendicular to bedding (tilted transverse isotropy or TTI), we would predict that, for
depths beneath 3500 feet, the slant path average velocity of the rig source survey to be greater than for
the NE fixed-offset survey at each recording level. This is confirmed in Figure 2.
The potential exists to compute interval anisotropy parameters and (Thomsen, 1986) directly from
the VSP data. However, with the interval velocity along the slow axis being a third unknown, we have
more unknowns than observations. Besides which, the incidence angles from the two VSPs were not
ideal for all layers. Even so, the VSP can provide valuable data with which to QC anisotropy
parameters that have been derived from conventional seismic data.
Anisotropy measurements from seismic
Although there are limitations with the straight ray assumptions of slant path velocities, there is a
strong indication from the VSP that anisotropy exists in the overburden. It was therefore decided to
apply anisotropic pre-SDM to the 3D survey over Pierce. The initial velocity model was defined by 12
seismic horizons with well-consistent vertical interval velocities. These velocities were an average of
sonic and checkshot implied velocities from all suitable wells over Pierce. Figure 3 shows isotropic
pre-SDM image gathers for the shallow overburden around the off-structure well 23/27-6 produced
using this velocity model. The fact that the gathers are not flat confirms that significant anisotropy
really does exist. Analysis of these gathers produced estimates of the anisotropic parameters, and
(Thomsen, 1986) for each layer and these were applied as laterally invariant values across the field
assuming a TTI axis of symmetry. The interval velocities were subsequently updated using a
tomographic solution during a number of iterations and the final anisotropic pre-SDM volume was
produced.
As a QC of the final velocity and anisotropy model, the observed travel times of the rig and NE fixed
offset VSPs of the 23/27-9 well were compared to equivalent ray-traced travel times for each
downhole receiver location. Figure 4 shows the difference in two traveltimes (VSP minus raytrace) as

a function of down-hole depth for the NE offset position. The rays for the NE offset VSP travel closer
to the normal to reservoir beds and are more relevant to their positioning. As a comparison, the time
differences are also shown for the case where the axis of anisotropic symmetry is assumed to be
vertical (VTI). It can be seen from this analysis that, for the TTI case, the ray tracing travel time
disparity increases from 0% at 4000ft to around 0.5% (too fast) approaching reservoir depth (7000ft).
In the VTI case the error at reservoir is around 2% too fast. There is evidence that, for depths greater
than 7000 ft, the VSP arrivals become affected by refraction within the salt and the comparison is not
as conclusive as for depths less than 7000 ft.
Seismic imaging results
Figure 5 shows the imaging results in depth achieved from anisotropic pre-SDM across the 23/27-9
well assuming both VTI & TTI symmetry. It can be seen that the TTI image has a better tie with the
formation tops of the well. Furthermore, when measured along the well track that is approximately
normal to the reservoir bedding and roughly parallel to the NE offset VSP raypaths, the Top Sele
(yellow marker) is about 0.5% deeper, relative to the well pick. This is in agreement with the predicted
percentage travel time error that is shown in figure 4. The equivalent mistie on the VTI product is
about 2% too deep, also in agreement with the prediction of figure 4.
In such a steeply dipping environment, a mistie in depth of 0.5% is considered quite acceptable. The
23/27-9 well is to the NE of Pierce South. Two other wells - 23/27-3 & 8 - SW of the same diapir have
been tied to the TTI seismic image with negligible observed error at Top Sele.
Conclusions
Simple analysis of the existing VSP data in well 23/27-9 indicated a strong likelihood that the
Pierce overburden shows significant seismic anisotropy.
The presence of seismic anisotropy was confirmed and measured by analysis of isotropic pre-SDM
image gathers produced with a well consistent velocity.
VSP arrival times were compared to final anisotropic ray-tracing travel times for a TTI axis of
symmetry. The two travel times were found to match within acceptable tolerances.
The accuracy of the final depth migrated volume in relation to formation tops in the 23/27-9 well
was found to be consistent with the VSP travel time analysis.
A TTI symmetry assumption proved essential for accurate imaging.
Acknowledgements
We wish to express our thanks to: Pierce partners for granting permission to publish this paper - Agip
(UK) Ltd., Enterprise Oil plc and MOC Exploration UK Limited.
The views expressed are purely those of the authors and are not necessarily those of the field operator.
References
Al-Chalabi, M. 1979. Velocity determination from seismic reflection data. In: Developments in Geophysical
Exploration Methods, Vol.1 (ed. A.A Fitch), pp. 1-68. Applied Science Publishers.
Armstrong, T., McAteer, J. and Connolly, P. 2001. Removal of overburden velocity anomaly effects for depth
conversion. Geophysical Prospecting, 49, 79-99.
Banik, N.C. 1984. Velocity anisotropy of shales and depth estimation in the North Sea basin. Geophysics, 49,
1411-1419.
Hawkins, K., Leggott, R., Williams, G. and Kat, H. 2001. Addressing anisotropy in 3D pre-stack depth
migration: A case study from the Southern North Sea. The Leading Edge, 20(5), 528-543.
Thomsen, L. 1986. Weak elastic anisotropy. Geophysics, 51, 1954-1966.
Vestrum, R. and Lawton, D.1999. Anisotropic depth migration: Reducing lateral-position uncertainty of
subsurface structures in thrust-belt environments. 69th SEG meeting, Houston, Expanded Abstracts, 1107-1109.
Vestrum, R.W., Lawton, D.C. and Schmid, R. 1999. Imaging structures below dipping TI media. Geophysics,
64(4), 1239-1246.

Fig1: VSP shooting geometry for well 23/27-9 with thumbnail

Seismic cross-section along the well bore.

Fig3: Isotropic PSDM image gathers produced using a

well-consistent Velocity model.

VSP minus Raytracing

time

Fig2: Average velocity of NE fixed offset survey v rig source survey for
Straight ray paths into each downhole geophone location. nb the black
line indicates where the two average velocities are equal to the right, rig
source Vavg is faster.

Fig4: Comparison of the difference of VSP arrival times

and ray-traced traveltimes assuming both TTI and VTI.

Fig5: A comparison of well-bore and seismic ties at 23/27-9. The misties are consistent with those observed in fig 4.