CHINESE JOURNAL OF GEOPHYSICS Vol.51, No.

3, 2008, pp: 639644

DISCRIMINATING GAS AND WATER USING MULTI-ANGLE
EXTENDED ELASTIC IMPEDANCE INVERSION
IN CARBONATE RESERVOIRS
PENG Zhen-Ming
1,2
, LI Ya-Lin
2
, WU Sheng-Hong
2
, HE Zhen-Hua
3
, ZHOU Yong-Jun
2
1 School of Opto-Electronic Information, University of Electronic Science and Technology of China,
Chengdu 610054, China
2 Sichuan Geophysical Company, CNPC Sichuan Petroleum, Chengdu 610213, China
3 State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Chengdu University of
Technology, Chengdu 610059, China
Abstract AVO response for carbonate reservoirs is weaker than the clastic reservoirs. So it is very hard to
discriminate uid in carbonate reservoirs using elastic impedance (EI), even under the optimal incident angle.
In this paper, a novel method is proposed for discriminating uid using multi-angle extended elastic impedance
(MEEI) inversion in carbonate reservoirs. We established the uid-bearing patterns and eective criterions for
discriminating gas and water by calculating MEEI of target layers using the known well data. MEEI inversion of
real seismic data shows that we can judge whether the reservoir is good or bad or is gas-bearing or water-bearing
by observing EEI change trend with various incident angles. The eciency of the proposed method is validated
by the practical application of real seismic data in Sichuan basin.
Key words Extended elastic impedance, Multi-angle analysis, Fluid criterions, Fluid discrimination, Carbonate
reservoirs
1 INTRODUCTION
The concept of elastic impedance (EI) was rst introduced by Connolly (1999)
[1]
, which is suitable for uid
discrimination and lithology prediction for various reservoirs because it includes more abundant information
of lithology and uid than acoustic impedance (AI). Recently, a great quantity of research work for EI has
been done by scientists in China and abroad. In order to overcome the drawback of incident angle restriction
and the rapid variation of inversion results with incident angle which gives rise to diculties for interpretation,
Whitcombe et al.
[2,3]
deduced the expression of the extended elastic impedance (EEI) by modifying the original
equation of the EI established by Connoly, which makes it more suitable for uid discrimination and lithology
prediction. They calculated the distribution map of lithology impedance and uid impedance using EEI in an
oil eld of North Sea and obtained the distinct imaging map of their channel system. Duaut et al.
[4]
introduced
the shear wave elastic impedance (SEI) to solve the problem for lithology prediction with converted waves.
Ezequiel
[5]
presented a formulation of P-to-S converted waves elastic impedance (PSEI) under arbitrary incidence
angles and used to distinguish zz water from commercial gas for predicting area distribution. From 2003, the
geophysical circles in China
[615]
began to pay attention to this subject, presented a series of concepts and
theories such as Ray-path elastic impedance (REI), Generalized elastic impedance(GEI), Zoeppritzt elastic
impedance(ZEI), etc., and a quantity of research and analysis works have been done in the areas of its practical
applications.
The authors consider that EEI is eective for lithology discrimination of clastic reservoirs by analyzing and
comparing the results of the above methods and computed the real seismic data. However, various case studies
show
[1417]
that AVO response for carbonate reservoirs is weaker than the clastic reservoirs. Diculties were
encountered when we discriminated uid for carbonate reservoirs under the optimal incident angle according to
the literature [3] because all EEI values indicate low value abnormity for gas-bearing or water-bearing reservoirs,
E-mail: zmpeng@uestc.edu.cn
640 Chinese J. Geophys. Vol.51, No.3
even under the optimal incident angle. We can eectively predict reservoirs, but it is beyond its capabilities to
use the above-mentioned method for evaluating the reservoir quality and discriminating gas- or water-content.
On the basis of the EEI theory, we computed the multi-angle extended elastic impedance (MEEI) of target
layers and observed its varying trend with various incident angles. In our experiments, we found that the MEEI
responses indicate the signicant dierence for gas-bearing or water-bearing reservoirs by combining logging
data (P- wave, S-wave and density) of drilled wells with the results of well log interpretation. It is proved
that using MEEI inversion to discriminate uid for carbonate reservoirs is eective and feasible and eective
criterion for discriminating gas and water was established in this paper. The eciency of the proposed methods
is validated by the application of real data in Sichuan basin.
2 THE DEFINITION OF ELASTIC IMPEDANCE AND EXTENDED ELASTIC IMPEDANCE
Elastic impedance (EI) is established on the basis of simplied equation of Aki-Richards
[18]
. If just consider
the incident angles less than 30

, then
R() = A + B sin
2
, (1)
where A =
1
2

v
p
v
p
+

, B =
v
p
2v
p
4

v
s
v
p

v
s
v
s

v
s
v
p

, and v
p
, v
p
, v
s
, v
s
, , donate
the average value and dierence of compressional (P-wave) velocity, shear wave velocity and density in adjacent
formation, respectively.
Connolly
[1]
constructed an equation for calculating reectivity similar to acoustic impedance as follows
R() =
EI
i+1
EI
i
EI
i+1
+ EI
i

1
2
EI
EI

1
2
ln(EI), (2)
where EI is called as elastic impedance. Comparing Eq.(1) and (2) and by simple computation, EI expressions
can be written as follows
EI() = v
(1+sin
2
)
p
v
(8K sin
2
)
s

(14K sin
2
)
, (3)
where v
p
is P-wave velocity, v
s
is S-wave velocity, is density and K =

v
si
v
pi

2
+

v
si+1
v
pi+1

/2 is the average
value of the square of the ratio of P to S velocity. is the incident angle of P-wave. The main problems in
calculating EI using Eq.(3) is that its value dramatically decreases with incident angle, which probably lose
some information for analyzing elastic impedance at near and far angles, for example, the dierence of uid or
lithology change.
Therefore, in order to eliminate the eects, Whitcombe et al.
[3]
deduced a new formula of EI by normalizing
the original EI equation, as follows
EI() =
0

, (4)
where
0
,
0
,
0
are constant coecient, which donate average value of v
p
, v
s
, in reservoir intervals, respectively.
, , are v
pi
, v
si
,
i
and at dierent depth, respectively, and
a = 1 + sin
2
, b = 8K sin
2
, c = 1 4K sin
2
. (5)
Although Eq.(4) can remove the eect of dramatic EI decrease with incident angle and keep the same
orders of magnitude with post-stack acoustic impedance, the reectivity R() are probably larger than 1 con-
icting with the real seismic recorder. Thus, Whitcombe
[3]
further modied the EI equation, and replaced sin
2

by tan ( is theoretical incident angle being expanded, which changes between 90

and 90

), scaled the re-

ectivity R() by multiplication by cos . The so-called extended elastic impedance (EEI) can be formulated
as
EEI() =
0

, (6)
Peng Z M et al.: Discriminating Gas and Water Using Multi-Angle Extended Elastic Impedance Inversion 641
where
p = cos + sin, q = 8K sin, r = cos 4K sin. (7)
EEI value calculated by Eq.(6) not only eliminates the eect of EI dramatic decrease with incident angle,
but also controls the reectivity within [1,1] and is more suitable to real seismic recorder and benecial to
directly detect lithology and uid.
3 MEEI WELL MODEL AND FLUID CRITERION FOR CARBONATE RESERVOIRS
Fluid detection using EEI needs to build the uid discrimination model according to the prior information.
In general, the relationship between EI and water saturation at xed incident angle is adopted and then the
interpretation scheme is determined. However, these methods have two disadvantages: one is that it is dicult
to predene a reasonable incident angle; the other is that the multi-angle information of pre-stack seismic data
is lost under the xed incident angle. In the study we found that the EEI inversion results of target layers show
low value abnormity under the gas-bearing or water-bearing conditions in xed incident angle (for example,
16

), and it is hard to discriminate the gas and water in reservoir. Some literatures also introduce the approach
to partition gas and water boundary by EI inversion for the limited partial stack seismic section at near and
far incident angle and the cross-plot analyses are done for inversion results and data samples are articially
classied for gas or water. However, this method is probably aected by various articial factors, because these
data are not completely divided, and cause the large interpretation error.
This paper presents a new model for discriminating uid using multi-angle extended elastic impedance,
(MEEI), in which the change trend of EEI with incident angles is observed continuously by calculating a series
of EEI within the ecient range of incident angles for reservoir zone. Then the corresponding uid criterions
are determined by prior information of uid-bearing (for example, log interpretation results). The analyses of
logging data (P-wave velocity, S-wave velocity and density) from several wells show that the proposed method can
construct an ecient uid-bearing pattern for carbonate reservoirs. Fig. 1 shows EEI varying with the incident
angle (theoretical angle ) of gas-bearing and water-bearing reservoirs of well X6. This well was interpreted as
gas-bearing layer at 43844386 m and 44714480 m depth and as water-bearing layer at 44884511 m. We
can see from the gure that EEI value increases with the incident angle decreasing for water-bearing reservoirs
and EEI value decreases with the incident angle increasing for water-bearing reservoirs, while it changes slowly
for gas-water-bearing. To further verify the regularity and obtain a consistent conclusion, the same analyses
are also done for the other wells in this area. Fig. 2 shows that EEI values for gas-bearing and water-bearing
layers of well X4 and X5 are consistent with Fig. 1. The EEI change trend of gas-bearing layer in well X1 is
also basically consistent.
The calculation of real data from several wells shows that MEEI analysis is eective for discriminating gas
and water in carbonate reservoirs and its response is obvious. Comparing with acoustic impedance, EEI also
642 Chinese J. Geophys. Vol.51, No.3
shows low value for gas-bearing or water-bearing reservoirs. However, MEEI for the dierent uid in reservoir
interval show the dierent responses as follows: (1) EEI increases with the increase of incident angle for water-
bearing case; (2) EEI decreases with increasing incident angle for gas-bearing case; (3) EEI changes slowly with
the increase of incident angle for gas-water-bearing case.
In practice application, gas-bearing cases of a few wells show inconsistent result with the above uid
criterion, in which EEI increases with incident angle increasing. It is caused by absence of S-wave velocity in
this well and using Castagna equation to t it. Because the relationship between P-wave and S-wave velocity
is not linear, it causes the uncertainty of calculating results. Generally, according to Castagna equation, P-
wave and S-wave velocity at most layers are correlative. However, practice shows that those zones with poor
correlation between v
p
and v
s
are just the places for favorable reservoir bodies. Therefore, the results and their
accuracy will be aected due to the absence of S-wave velocity in uid discrimination using MEEI inversion.
For the sake of analyzing problem, the label of the abscissa is the theoretical angles (not the actual incident
angle) in Fig. 1 and Fig. 2, which change within 0

90

. The relationship between the actual incident angle

and the theoretical angles is monotonic increasing or decreasing, and we aim at observing the EEI change trend
with incident angle for gar-bearing or water-bearing reservoirs rather than numerical value of actual incident
angles. It will not aect the result of analyzing problem, although the theoretical angles are not converted to
the actual incident angles.
4 CASE STUDY AND RESULTS ANALYSIS
MEEI analyses are conducted for logging data and the gas and water criterions are established by forward
calculation based on the above approaches, then the real seismic data can be inversed for angle trace gathers
with pre-stack NMO and pre-stack time migration. In the study area, MEEI inversion was tested for two 2-D
seismic line across gas well and water well, respectively.
Figure 3 and 4 are the results of MEEI inversion and uid detection for real 2-D line across typical gas
well and water well, respectively. Fig. 3a, 3b, and 3c are the EEI inversion sections at gas-bearing reservoir
intervals of inline 450 (across well X4) with various actual incident angles (5

, 13

, and 21

), where navy blue

color donates the low EEI value. We can see from the regions marked as red ellipses in the gures that the
color of inversion results within gas-bearing reservoir interval deepens gradually with incident angles increasing,
indicating that EEI values decrease with incident angles increasing. The EEI inversion results of reservoir
intervals of inline 557 (across well X6, interpreted as water-bearing layer in logging ) under various actual
incident angles (5

, 15

and 25

) are shown in Fig. 4a, 4b, and 4c. Similarly, we can see from the regions marked
as purple ellipses in the gures that the color of inversion results within water-bearing reservoir interval lightens
gradually with incident angles increasing, indicating that EEI values increase with incident angles increasing.
The inversion results of real seismic data are all consistent with uid discrimination criterion established
in this paper. Moreover, model-based inversion of MEEI was realized for the 3-D working area of 100 m
2
.
Consistency verication has been done by using 3 wells of the 6 wells drilled in the area as constraining conditions
and taking the others as the unknown wells, the resultant consistency ratio is 100%. The test results show that
the proposed method based on MEEI for discriminating uid is ecient and feasible.
5 CONCLUSIONS
The uid discrimination criterions are established by analyzing logging data at reservoir intervals and
forward calculation in this paper. The test results for real 3-D seismic data show that it is an eective approach
to discriminating gas and water in carbonate reservoirs using MEEI inversion and is better than EEI inversion
under the xed incident angle or EEI cross-plot analysis between near and far angles. Although the present
method is used to detect uid in carbonate reservoirs, it also has referential signicance to lithology identication
in sand-shale formation or clastic reservoirs.
Peng Z M et al.: Discriminating Gas and Water Using Multi-Angle Extended Elastic Impedance Inversion 643
The actual incident angles is (a) 5

, (b) 13

, (c) 21

. The actual incident angles is (a) 5

, (b) 15

, (c) 25

.
ACKNOWLEDGMENTS
Authors would like to express their gratitude to Sichuan Geophysical Company of CNPC Sichuan Petroleum
for permitting to publish this project. This work has been supported by the Aeronautic Science Foundation of
China (20060112116), Research Fund for the Doctoral Program of Higher Education of China (20070614016)
and National Natural Science Foundation of China (60572092).
REFERENCES
[1] Connolly P. Elastic impedance. The Leading Edge., 1999, 18(4): 438452
[2] Whitcombe D N. Elastic impedance normalization. Geophysics, 2002, 67(1): 6062
[3] Whitcombe D N. Extended elastic impedance for uid and lithology prediction. Geophysics, 2002, 67(1): 6367
[4] Kenneth D, Martin L, Hege R. Shear-wave elastic impedance. The Leading Edge, 2000, 19(11): 12221229
[5] Ezequiel F G, Tapan M, Gray M, et al. Near and far oset P-to-S elastic impedance for discriminating zz water
from commercial gas. The Leading Edge, 2003, 22(10): 10121015
644 Chinese J. Geophys. Vol.51, No.3
[6] Ni Y. A new method for calculation of elastic wave impedance. Oil Geophysical Prospecting (in Chinese), 2003, 38(2):
147150
[7] Ma J F. Forward modeling and inversion method of generalized elastic impedance in seismic exploration. Chinese J.
Geophys. (in Chinese), 2003, 46(1): 118124
[8] Ma J F, Morozov I B. Ray-path elastic impedance. CSEG National Convention, 1012 May 2004
[9] Ma J F, Morozov I B. Fluid detection study from Zoeppritz elastic impedance, 75th Annual Internat. Mtg., Soc.
Expl. Geophys., Expanded Abstracts, 2006
[10] Gan L D, Zhao B L, Du W H, et al. The potential analysis of elastic impedance in the lithology and uid prediction.
Geophysical Prospecting for Petroleum (in Chinese), 2005, 44(5): 504508
[11] Wang B L, Yin X Y, Zhang F C. Elastic impedance inversion and its application. Progress in Geophysics (in
Chinese), 2005, 20(1): 8992
[12] Cao M Q, Wang J S, Shao L H. Prestack elastic wave impedance inversion technique and application. Oil Geophysical
Prospecting (in Chinese), 2006, 41(3): 323326
[13] Yuan S J, Yu C Q. Elastlc impedance and seismic inversion in anisotropic media. Progress in Geophysics (in Chinese),
2006, 21(2): 520523
[14] Peng Z M, Li Y L, Ye L, et al. Discriminating gas and water in carbonate reservoirs using multi-angle extended
elastic impedance inversion. SPG/SEG Kunming 2006 International Geophysical Conference, Expanded Abstracts,
168171, Oct. 2006
[15] Peng Z M, Li Y L, Liang B, et al. Application prestack elastic impedance inversion in gas and water recognition of
the reservoir. Natural Gas Industry (in Chinese), 2007, 38(4): 4345, 52
[16] Li Y, Goodway B, Downton J. Recent advances in application of AVO to carbonate reservoirs. Cseg Recorder, 2003.
3540
[17] Wang D W, Liu Z, Chen X H, et al. Feasibility analysis on division of ow units using time-lapse seismic data.
Chinese J. Geophys. (in Chinese), 2007, 50(2): 592597
[18] Aki K, Richards P G. Quantitative Seismology: Theory and Methods. Volume 1, W. H. Freeman and Company,
1980