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

net/publication/266636031

AVO Interpretation of a Gas Bearing Sand Reservoir Encased in Its Source Rock

Article · June 2014

DOI: 10.3997/2214-4609.20141018

CITATIONS READS

0 171

5 authors, including:

Pedro Alvarez Francisco Bolivar

RSI - Rock Solid Images Tricon Geophysics
28 PUBLICATIONS   27 CITATIONS    16 PUBLICATIONS   27 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Rock property estimation from seismic data View project

Quantitative interpretation of prestack seismic and CSEM data View project

All content following this page was uploaded by Pedro Alvarez on 21 October 2016.

The user has requested enhancement of the downloaded file.

We D203 03
AVO Interpretation of a Gas Bearing Sand
Reservoir Encased in Its Source Rock
P. Alvarez* (Rock Solid Images), W. Marin (Rock Solid Images), F. Bolivar
(Rock Solid Images), M. Di Luca (Pacific Rubiales Energy) & T. Salinas
(Pacific Rubiales Energy)

SUMMARY
In this paper we show a case study where a gas bearing sand reservoir encased with shale of great
thickness and rich in organic matter, which are considered as source rock and potential reservoir (Barrero
et. al, 2007), occasioned the presence of unconventional background trend in the AVO intercept (A) and
gradient (B) crossplot. This condition, change the expected AVO response of the shales, altering the
direction of the typical background trend, in which coincide the AVO response of the shales and brine-
saturated sandstones. In consequence, to be able to identify AVO anomalies related to gas bearing sands in
this geological context, a detailed log conditioning process and rock physics analysis were performed to
have a robust set of well logs. Next, rock physics modeling, fluid substitution and thickness modeling was
performed in order to understand the difference between the AVO anomalies related to gas bearing sands
regarding to the ‘non-conventional’ trend produced by the organic shales. The results obtained along with
AVO attribute volumes calculated from real seismic data, after a rigorous conditioning process in pre-stack
domain, were used to identify potential AVO anomalies related to gas bearing sands.

76th EAGE Conference & Exhibition 2014

Amsterdam RAI, The Netherlands, 16-19 June 2014
Introduction

AVO interpretation is facilitated by cross-plotting of AVO intercept and gradient attributes. Under a
variety of reasonable geologic circumstances, brine-saturated sandstones and shales follow a well-
defined “background” trend in the intercept versus gradient cross-plot. “AVO anomalies” are properly
viewed as deviations from this background and may be related to hydrocarbons or lithological factors
(Castagna & Swan, 1997). We present a case study wherein an unusual geologic condition produces
the presence of an atypical background trend in the intercept versus gradient cross-plot space. This
condition is caused by the encasing of the reservoir between shales of great thickness which have
been identified as a source rock, as well as a potential reservoir (Barrero et. al, 2007). This
characteristic, change the expected AVO response of the shales, altering the direction of the typical
background trend, in which coincide the AVO response of the shales and brine-saturated sandstones.
To overcome this problem a methodology with three phases was developed. Firstly, AVO modeling
of different geological conditions was carried out based on rock physics algorithms. Secondly, AVO
attributes were calculated after rigourous conditioning of the pre-stack seimic data. Finally,
integration of the previous results served to identify those zones with higher probability of gas
content.

In situ AVO modelling

After a rigorous log conditioning and petrophysical analysis process, a synthetic CDP gather was
generated to evaluate the AVO response in the target zone using a ray tracing algorithm. The results,
shown in Figures 1 and 2, allowed us to identify AVO anomalies not only at the top and the base of
the gas sand reservoir as expected, but also in some shale intervals. The most significant AVO
responses identified in the synthetic gather can be described by the following three groups.

Lithology In-situ AVO response

X020

X289

X563

X841

X131

X424

X722

Figure 1: From left to right: Total porosity, water saturation, -lithology, upscaled logs of acoustic
impedance and Poisson’s ratio, and synthetic AVO response.

76th EAGE Conference & Exhibition 2014

Amsterdam RAI, The Netherlands, 16-19 June 2014
 Highlighted interfaces
Full range

Gradient

Intercept

Figure 2. Intercept vs gradient cross-plot showing the different AVO responses found in the synthetic
gather at in situ conditions. Black circles represent the highligted interfaces. Orange points show the
full range response

The first group, identified with the letters B and C, corresponds to the top and the base of the reservoir
respectively. A subtle class II AVO anomaly can be identified for this interval, which is located close
to the background trend in the intercept vs gradient crossplot (Figure 2). The second group, identified
with the letters D, E and F, is from a shale interval that contains several class III AVO responses
associated with a subtle drop in Poisson’s ratio. This repeat elastic response is likely produced by a
combination of gas and/or high organic content in the shale, as these intervals have been identified as
source rock as well as potential reservoir (Barreto et. al, 2007). The last group, identified with the
letter A, corresponds to a shale interval of dim reflectivity with no AVO anomalies, or Poisson’s ratio
variations.

Half space & fluid substitution modelling

Because the thickness of the reservoir interval (B-C) is below seismic resolution, AVO half space
modelling was carried out to eliminate this effect on the AVO response. Half space modelling, which
assumes that the layer involved in the AVO analysis has infinite thickness, was done by averaging the
elastic properties of a zone above the reservoir and a zone within the reservoir, and using these
average properties in the Zoeppritz (1919) equation (Hübert et. al., 2006). Figure 3 shows a detailed
view of the well-log information of the reservoir zone where the half space modelling was performed.
Inside the reservoir interval, two zones can be identified with different petrophysical and elastic
properties, named interval 1 and interval 2 (Figure 3). The AVO response between the upper shale
and each one of these intervals was modelled using the half space modelling technique. The results
are shown in the Figure 4a.

B
Interval 1
30’
Interval 2
C

Figure 3: Well-log information of the reservoir zone. Black line: conditioned well log, blue line:
upscaled log, straight line: average properties used in the half space modelling for each interval

76th EAGE Conference & Exhibition 2014

Amsterdam RAI, The Netherlands, 16-19 June 2014
 Figure 4: Intercept vs gradient cross-plot showing: a) AVO responses between the upper shale and
interval 1 (magenta), interval 2 (cyan), average properties of interval 1 and 2 fluid subtituted by brine
(blue). The black circles represent the in situ response. b) Zones associated with gas sand (red),
organic shale (green) and shale and wet sand (blue). Gas sand with a thickness below seismic
resolution or low gas saturation is expected to fall in the yellow or blue band.

In the case of interval 1 (magenta square), the AVO response obtained was a class III anomaly which
shows a good separation from the background trend, including the interfaces D, E &F. For Interval 2
(cyan square), which has lower porosity and higher water saturation than Interval 1, a subtle class I
anomaly was identified. Fluid subtituion, via Gassmann’s Equation (1955), was performed in order to
identify the direction of the backgroung trend (related to brine saturated rocks) in the gradient versus
intercept domain, necessary information for interpretation of AVO anomalies (black line in Figure
4b). This calibration to wet conditions was accomplished by modeling the AVO response between
the upper shale and a sand rich layer fluid substituted 100% by brine (blue square). From these
analyses it is possible to see that factors such as porosity and thickness (if the reservoir is under or
close to seismic resolution), in addition to the fluid effect, might influence the AVO response.
Nevertheless, it also shows that a porous and gas- bearing seismically resolved sand, can be separated
from the background trend, including the organic shale response.

AVO Inversion and interpretation

Prior to AVO inversion, a seismic conditioning process was applied to the pre-stack data, following
the approach of Singleton (2009). The key processes involved were coherent and random noise
suppression, residual move-out analysis and spectral balancing between near and far traces. Next,
amplitude values of the pre-stack seismic data were scaled to the amplitude values of the synthetic
gathers based on the well tie. Those well ties made it possible to estimate the phase rotation to be
applied to the pre-stack data in order to get it close to zero degrees. Then, an AVO inversion using a
3-term Aki & Richard equation was performed. Gradient and intercept volumes were cross-plotted
and compared with the dynamic range of the attributes estimated from the synthetic gathers. AVO
attribute interpretation was performed based on the in situ and modelled AVO responses, using the
color-coded bands shown in Figure 4b. The areas with top and base anomalies were isolated. A
detailed gather-to-gather revision was completed for each anomaly in order to discard any false
positives due to remaining noise or any kind of artifact in the seismic data. Results from the detection
process were compared to a fluid factor volume calculated using the equation defined by Smith and
Gidlow (2003):

Where A and B are the intercept and gradient respectively, and Mis the fluid angle, which is shown in
Figure 4b. The result obtained for one of five analysed intervals is shown in Figure 5. An example of
one AVO anomaly with high probability of being related to gas-bearing sands is identified with the
number 1. CDP gather and AVO response related to the anomaly were analysed for seismic data

76th EAGE Conference & Exhibition 2014

Amsterdam RAI, The Netherlands, 16-19 June 2014
quality to gauge the fidelity of the AVO response. Additionally, as can be seen in the fluid factor
section, the anomaly is located just below the structural crest of a normal-faulted four-way closure,
which is consistent with the geological information for the area and adds confidence to the results.

Figure 5. a) Map view of the top (red) and base (blue) AVO anomalies likely related to gas bearing
sands. b) Crossplot of intercept vs gradient together with the polygons used to identify the AVO
anomalies. c-d) CDP gather and AVO response related with the anomaly identified with the number
1. e) Cross-section of the fluid factor along the anomaly identified with the number 1.

Conclusions

An AVO analysis and interpretation case study has been shown, where initial well-log modeling of
the AVO response shows certain geological conditions which make interpretation difficult. These
encountered conditions are reflected in (1) gas-bearing sands with thickness below seismic resolution
and (2) intervals of shale that show multiple AVO class III anomalies. The application of half space
modelling and fluid substitution techniques allowed us to understand that gas bearing sand with
thickness greater than the seismic resolution can be separated from the organic shale interval in the
intercept vs. gradient cross-plot, as well as, to identify the direction of the background trend, which in
these geological conditions cannot be estimated using traditional approaches. The understanding of
the aforementioned factors was the key to perform a quantitative interpretation of AVO attributes that
allowed us to isolate AVO responses, which can be related to the presence of gas bearing sand with
thickness great enough to be identified by the seismic data.

References

Barrero, D., Pardo, A., Vargas, C., Martinez, J. [2007]. Colombian Sedimentary Basin: Nomenclature,
Boundaries, Petroleum Geology, a New Proposal. ANH report.
Castagna, J. and Swan, H. [1997]. Principles of AVO crossploting. The Leading Edge. April 1997.
Hübert, L., Müller, K., Selnes, A. [2006]. Improving AVO modeling using geological knowledge, 4
examples from the Norwegian Continental Shelf. 76th SEG Technical Program Expanded
Abstracts, 2006.
Singleton, S. [2009]. The effects of seismic data conditioning on pre-stack simultaneous impedance
inversion. The Leading Edge, 28(7), 772–781.
Smith, C. and Gidlow, M. [2003]. The Fluid Factor Angle and the Crossplot Angle. 73th SEG
Technical Program Expanded Abstracts.

76th EAGE Conference & Exhibition 2014

Amsterdam RAI, The Netherlands, 16-19 June 2014

View publication stats