See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/290581644

The Application and Value of AVO Analysis for De-risking Hydrocarbon

Paleocene Prospects in the UK North Sea

Article · May 2011

DOI: 10.3997/2214-4609.20149356

CITATIONS READS

2 449

4 authors, including:

Shuangquan Chen
China University of Petroleum - Beijing
54 PUBLICATIONS 370 CITATIONS

SEE PROFILE

All content following this page was uploaded by Shuangquan Chen on 01 March 2017.

The user has requested enhancement of the downloaded file.

The application and value of AVO analysis for de-risking hydrocarbon
Paleocene prospects in the UK North Sea
Nick Loizou1, Shuangquan Chen2,3 ,Don Cameron3 and Xiang-Yang Li3,2
1
Department of Energy & Climate Change, London, SW1A 2AW, UK
2
CNPC Keylab of Geophysical Prospecting, China University of Petroleum, Beijing, 102249, CHINA
3
British Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, UK

Summary

Perhaps over a billion barrels of oil remains trapped within undrilled Tertiary prospects in the North
Sea. This study’s primary aim has been to establish the suitability of AVO analysis for derisking
Paleocene prospects. For this, we have carried out a post-mortem analysis of seven wells drilled since
2005 in traps defined by seismic amplitude or possible AVO anomalies. Only two of these wells
encountered hydrocarbons. Conventional AVO attribute study has been carried out, together with
time-frequency spectral analysis. We conclude that the majority of failures relate to poor definition of
traps mapped on the basis of AVO artefacts and spurious amplitudes that could all have been avoided
by robust AVO analysis. Three key problems were recognized whilst carrying out the analysis: (1)
many 3D seismic datasets were acquired with less than 3 km cable length and are thus not ideal for
robust AVO studies; (2) previous data processing flow was not optimized for the purpose of AVO
analysis; and (3) the frequency content of the data in some cases was less than 30 hz, where spectral-
related anomalies could be critical in differentiating hydrocarbons from other fluid-related anomalies.

73rd EAGE Conference & Exhibition incorporating SPE EUROPEC 2011

Vienna, Austria, 23-26 May 2011
Introduction

Since 2000, twenty-five exploration wells in the Northern North Sea and the Moray Firth have been
positioned on Paleocene prospects, of which only five encountered notable hydrocarbons.
Approximately half of these wells were reported to have an amplitude or AVO (amplitude versus
offset) anomaly seen mainly on the full offset stack with very little evidence of amplitude increase on
the far offsets. With the majority of the wells located within mature play fairways, the expectancy of
overall success rate should have been much better than one in five. To find out whether AVO analysis
can provide a useful tool in helping to de-risk remaining prospects and hence improve future drilling
success rates, we selected seven of the twenty-five wells to perform an integrated AVO analysis of log
and seismic data. Of the seven wells, two encountered hydrocarbons, and the rest were dry holes.
Several techniques have been used to obtain AVO signatures, including conventional AVO attribute
study, together with more recent techniques such as time-frequency spectral analysis.

Previous AVO analysis of several wells located in the west of Shetland basins demonstrated that it is
paramount to understand the quality and suitability of the seismic data ahead of carrying out AVO or
other analysis (Loizou et al. 2008). One of the key aims of this new study has been to ascertain the
usefulness of AVO analysis, particularly in areas where oil has been the main objective.

Methodology

We performed a standard AVO analysis based on Shuey’s (1985) two-term approximation:

R(θ ) = A + B sin 2 (θ ) (1)
where, R is the reflectivity of the PP-wave, θ is the angle of incidence, A is the AVO intercept and B
is the AVO gradient. Rutherford and Williams (1989) suggested a classification scheme of AVO
responses for different types of gas sands based on the cross plot of intercept (A) versus gradient (B).
They defined three AVO classes according to where the top of the gas sands is located, in an A versus
B cross-plot.

For the AVO modelling, Gassmann’s (1951) equation was used to predict the dry-frame modulus
from the brine-saturated velocity, and then fluid substitution was carried out for other fluid types (i.e.
gas and oil), again using Gassmann’s equation.
(1 − κ dry / κ m ) 2
κ sat = κ dry + ; μsat = μdry (2)
φ / κ f + (1 − φ ) / κ m + κ dry / κ m2
Note that Gassmann’s equation predicts no change of shear modulus with fluids. However, shear-
wave velocity is expected to show a slight dependence on fluids because of the different densities of
the fluids. In this study, we discounted slight changes of shear-wave velocity with fluid.

Frequency-dependent AVO analysis is based on the theory of Chapman et al. (2006), who argued that
in contrast with water saturation, the presence of hydrocarbons induces frequency-dependent AVO for
class I and class III types of reservoirs. This implies that it is possible to use frequency-dependent
AVO analysis to distinguish fluid types, which is often implemented using spectral decomposition
techniques to perform time-frequency analysis. Here, for time-frequency analysis, we chose the
Wigner-Ville distribution (WVD) method (Cohen, 1995). The WVD method can reach a higher time
and frequency resolution than other time-frequency decomposition techniques, and it is also a very
good method for time-frequency analysis of non-stationary signals. If we define the seismic signal as
x(t), the Wigner-Ville distribution of it will be:
∞
WVD(t , f ) = ∫ x(t + τ / 2) x (t − τ / 2)e− j 2π f τ dτ (3)
−∞
where τ is the time delay variable, and x (t ) is the analytical signal associated with the real signal x(t).

73rd EAGE Conference & Exhibition incorporating SPE EUROPEC 2011

Vienna, Austria, 23-26 May 2011
Datasets and analysis procedures

The seven wells analysed include two with hydrocarbons (wells 16/23-7 and 15/19b-10). Wells 9/5b-
5A, 9/10b-5A, 9/27a-4, 15/13b-8 and 15/30a-14 are all dry holes. All wells in the study were drilled
within the last five years, and encountered sandstone reservoirs with high porosity ranging from 24%
to 29%. The depths to the objective reservoirs ranged from 1570 m to 2350 m TVDSS. The datasets
include all wireline logs, original 2D and 3D seismic data, and processed prestack seismic gathers,
together with supporting well testing and stratigraphic information.

The following analysis was performed for each of the wells: (1) AVO modelling to determine the
AVO class at the top of the reservoir interval, and the effects of fluid substitution on the AVO
signatures using Gassmann’s equation; (2) conventional processing including velocity analysis, NMO
correction, and stacking to generate full-, near- and far-offset stacked sections and to compare the
amplitude variations; (3) standard AVO analysis including AVO cross-plots and standard AVO
attribute sections; and (4) iso-frequency analysis by spectral decomposition of the full-offset stack
sections to evaluate frequency-related amplitude variations.

Results

Snake prospect well 9/5b-5A focused on the Paleocene Heimdal sandstone as the key reservoir target
(Figure 1a). The gathers show an obvious decrease in amplitude from approximately 2100 ms, and
residual statics can still be observed in the far offsets (Figure 1b). The AVO modelling suggests a
Class I, and the cross plot confirms this with no noticeable anomaly (Figures 1c and 1d). There were
no anomalies shown in the attributes sections (Figures 1e and 1f), and consequently no AVO support
for this well.

Well 9/10b-5A was located on a

Paleocene Heimdal sandstones
prospect referred to as Red Deer.
The 3D seismic data is poor to
average quality, and the gathers
are heavily muted, whilst there is
very little difference in the near
and far offsets to suggest the
presence of hydrocarbons.
Extracting a 48 ms window over
the reservoir revealed no AVO
anomalies; the results are very
similar to the 9/5b-5A well.

Well 9/27a-4 was positioned on a

Paleocene Forties Sandstone
prospect referred to as Hendrix.
Generally, the 3D seismic data
quality is viewed as below
average quality. Top reservoir is
approximately 1.745 seconds Figure 1 Well 9/5b-5A - well data and AVO characteristics. (a)
with the gathers showing an Well log curves; (b) the migrated prestack gathers; (c) AVO
overall decrease in amplitude intercept and gradient (A-B) cross-plot around the target; (d)
with offset. Furthermore, the cross-plot result with the stacked seismic section; AVO attribute
cross-plot for this well shows a sections (e) A*B and (f) polarity of A.
classic wet rock trend, and no AVO anomaly.

73rd EAGE Conference & Exhibition incorporating SPE EUROPEC 2011

Vienna, Austria, 23-26 May 2011
Well 15/13b-8 was located on a
prospect referred to as Guinea.
Its main target was Paleocene,
Balmoral Sandstone; AVO
modelling indicates a Class III
response, but no anomaly can be
observed.

For the above four wells, no

amplitude anomalies were
observed in the iso-frequency
sections obtained through
spectral decomposition.

Well 15/19b-10 was positioned

on the Stag Prospect and
encountered 27 ft of oil in
Paleocene, Balmoral Sandstone
(Figure 2a). Although the oil
bearing zone is relatively thin, Figure 2 Well 15/19b-10. The log data and AVO characteristic.
the gathers do indicate a distinct (a)~(f) are the same as in Figure 1.
and noticeable increase in
amplitude with offset (Figure 2b). The crossplot shows a Class III AVO anomaly coinciding with the
27 ft oil column. Furthermore, there is a clear decrease of amplitude with frequency at the target
position in the iso-frequency sections from 10 hz to 40 hz (Figures 3a~3d).

Positioned on the Dunkeld prospect, well 15/30a-14 encountered thick, high porosity Paleocene
Beauly sands with no hydrocarbons (Figure 4a). Modelling shows that the far offsets are over-scaled
by a factor of 1.5, which gives rise to a false Class III AVO response (Figures 4b and 4c). Moreover,
there is no obvious anomaly in the iso-frequency sections (Figure 5).

Figure 3 The iso-frequency sections for well Figure 5 The iso-frequency sections for well
15/19b-10 (a) 10Hz; (b) 20Hz; (c) 30Hz; (d) 15/30a-14 (a) 10Hz; (b) 20Hz; (c) 30Hz; (d) 40Hz.
40Hz.

Positioned on a prospect referred to as Kinnoull, well 16/23s-7 encountered 52 m of oil-bearing

Paleocene Balmoral Sandstone with 26% porosity. Unfortunately, the 3D seismic data was acquired
with a cable length of less than 3 km, resulting in very limited far offset data, particularly as the
objective was at 2.35 km depth. Examination of the gathers shows nothing striking, with the signal
noise ratio relatively low, and the target event unidentifiable on the gathers. Not surprisingly, AVO
modelling revealed a Class I with no obvious anomalies in the AVO attributes section. Furthermore,
no frequency-related variations can be observed. It is clear the well is positioned on a structural
closure rather than a distinct AVO anomaly. Although it is a success in terms of hydrocarbons, it is

73rd EAGE Conference & Exhibition incorporating SPE EUROPEC 2011

Vienna, Austria, 23-26 May 2011
 the only well that cannot be explained by the AVO analysis, mainly because of the somewhat
inadequate, poor quality of the only 3D seismic dataset available.

Conclusions

An integrated AVO analysis was performed using wireline logs and seismic data from recently drilled
exploration wells located in the Northern North Sea and Moray Firth. Careful AVO modelling was
carried out to determine the AVO
classes, fluid and lithology
characteristics. With the exception
of the 15/30a-14 well, none of the
dry wells showed strong amplitude
anomalies on the far offset. No
obvious AVO anomalies or
apparent frequency-related effects
were observed. For the 15/30a-14
well, despite a strong Class III
AVO anomaly which is a
processing artefact, the spectral
decomposition reveals no clear
frequency-related effects. For the
hydrocarbon-bearing reservoir
encountered in the 15/19b-10 well,
the analysis confirms the presence
of a Class III AVO anomaly, which Figure 4 Well 15/30a-14. The well data and AVO
is further supported by frequency- characteristic. (a)~(f) are the same as in Figure 1.
dependent AVO analysis. Well 16/23-7 gave anomalous and inconclusive results, mainly due to the
poor seismic data quality. To conclude, the results show that the appropriate use of AVO analysis
should provide a valuable tool for de-risking Paleocene exploration prospects in the UK North Sea.
However, it is paramount and fundamental to understand the quality and suitability of the seismic data
for such analysis.

Acknowledgements

The authors thank DECC for permission to show the data. This work is supported by the Edinburgh
Anisotropy Project (EAP) of the British Geological Survey, and is published with the approval of the
Executive Director of BGS and the EAP sponsors.

References

Chapman, M., Liu, E. and Li, X-Y, The influence of fluid-sensitive dispersion and attenuation on AVO
analysis. Geophys. J. Int. 2006, 167, 89-105.
Cohen, L., 1995. Time–Frequency Analysis. Prentice Hall Inc., New York, USA.
Gassmann, F., 1951, Uber die Elastizitāt Porōer Medien, Vier. der Natur. Gesellschaft in Zürich, 96, 1-23.
Loizou, N., Liu, E. and Chapman, M., 2008. AVO analyses and spectral decomposition of seismic data
from four wells west of Shetland, UK. Petroleum Geoscience. 14, 355-368.
Rutherford, S. R. and Williams, R. H., 1989, Amplitude-versus-offset variations in gas sands: Geophysics,
54, 680-688.
Shuey, R. T., 1985, A simplification of the Zoeppritz equations: Geophysics, 50, 609–614.

73rd EAGE Conference & Exhibition incorporating SPE EUROPEC 2011

Vienna, Austria, 23-26 May 2011

View publication stats