Journal of Geophysics and Engineering

PAPER Related content

- Anisotropy parameters estimate and rock
An integrated petrophysical and rock physics physics analysis for the Barnett Shale
Zhiqi Guo, Xiang-Yang Li and Cai Liu
analysis to improve reservoir characterization of - Application of different classification
methods for litho-fluid facies prediction: a
Cretaceous sand intervals in Middle Indus Basin, case study from the offshore Nile Delta
Mattia Aleardi and Fabio Ciabarri
Pakistan - Estimation of anisotropy parameters for
shales based on an improved rock physics
model, part 2: case study
To cite this article: Tahir Azeem et al 2017 J. Geophys. Eng. 14 212 Feng Zhang

Recent citations
View the article online for updates and enhancements.
- Effect of pore structure on the dispersion
and attenuation of fluid-saturated tight
sandstone
Changsheng Duan et al

This content was downloaded from IP address 54.70.40.11 on 12/04/2018 at 14:37

 Journal of Geophysics and Engineering

J. Geophys. Eng. 14 (2017) 212–225 (14pp) doi:10.1088/1742-2140/14/2/212

An integrated petrophysical and rock

physics analysis to improve reservoir
characterization of Cretaceous sand
intervals in Middle Indus Basin, Pakistan
Tahir Azeem1,2, Wang Yan Chun2, MonaLisa1, Perveiz Khalid3,
Liu Xue Qing2, Muhammad Irfan Ehsan3, Muhammad Jawad Munawar4 and
Xie Wei2
1
Department of Earth Sciences, Quaid-I-Azam University, Islamabad, Pakistan
2
School of Geophysics and Information Technology, China University of Geosciences, Beijing 100083,
People’s Republic of China
3
Institute of Geology, University of the Punjab, 54590 Lahore, Pakistan
4
School of Geosciences, China University of Petroleum, Qingdao, People’s Republic of China

E-mail: [email protected]

Received 15 October 2015, revised 24 November 2016

Accepted for publication 30 November 2016
Published 20 January 2017

Abstract
The sand intervals of the Lower Goru Formation of the Cretaceous age, widely distributed in the
Middle and Lower Indus Basin of Pakistan, are proven reservoirs. However, in the Sawan gas
field of the Middle Indus Basin, these sandstone intervals are very deep and extremely
heterogeneous in character, which makes it difficult to discriminate lithologies and fluid
saturation. Based on petrophysical analysis and rock physics modeling, an integrated approach is
adopted to discriminate between lithologies and fluid saturation in the above-mentioned sand
intervals. The seismic velocities are modeled using the Xu–White clay–sand mixing rock physics
model. The calibrated rock physics model shows good consistency between measured and
modeled velocities. The correlation between measured and modeled P and S wave velocities is
92.76% and 84.99%, respectively. This calibrated model has been successfully used to estimate
other elastic parameters, even in those wells where both shear and sonic logs were missing.
These estimated elastic parameters were cross-plotted to discriminate between the lithology and
fluid content in the target zone. Cross plots clearly separate the shale, shaly sand, and gas-bearing
sand clusters, which was not possible through conventional petrophysical analysis. These data
clusters have been exported to the corresponding well for the purpose of interpolation between
wells and to analyze the lateral and vertical variations in lithology and fluid content in the
reservoir zone.
Keywords: petrophysics, rock physics modeling, elastic parameters, reservoir characterization,
cross well correlation, Sawan gas field

(Some figures may appear in colour only in the online journal)

1. Introduction and Hongwei 2007). In the literature, different techniques

have been proposed for fluid and lithology discrimination
Petrophysical analysis plays an important role in reservoir (Castagna and Swan 1997, Chi and Han 2009, He et al 2011,
characterization, especially in discriminating between the Hu et al 2011, Ahmed et al 2016). Generally, petrophysical
hydrocarbon and non-hydrocarbon bearing zones (Yuedong analysis is performed to transform the wireline log data into

1742-2132/17/020212+14$33.00 212 © 2017 Sinopec Geophysical Research Institute Printed in the UK

J. Geophys. Eng. 14 (2017) 212 T Azeem et al

reservoir properties such as volume of shale, porosity, per- combination of shear and acoustic velocities has been
meability, and water and hydrocarbon saturation. Proper extensively used as a seismic attribute in reservoir char-
analysis of these reservoir properties can significantly acterization (Avseth et al 2005, Chi and Han 2009). Seismic
enhance the ability to discriminate between the hydrocarbon velocities of rock provide information about minerals, pore
and non-hydrocarbon bearing zones (Ajisafe and Ako 2013). fluids, and their distribution within the rock skeleton (Mavko
However, petrophysical results can be affected by bad bore- et al 1998, Avseth 2000, Feng-Ying et al 2014). Like other
hole conditions, missing logs, temperature, pressure, and parameters such as porosity, permeability, fluid saturation,
salinity. In addition, based on single-well data, petrophysical fluid composition, formation temperature, pressure, and
models are generally established for the particular interval of mineralogy, clay content also influences the seismic proper-
interest. Therefore, petrophysical models are often consistent ties of a porous rock (Ahmed et al 2016, Wang et al 2015).
over the particular interval of the well and sometimes fail to The presence of clay decreases the seismic velocities
provide good results between the wells, even when the wells (Minear 1982, Han et al 1986, Marion et al 1992, Ahmed at
are very closely spaced (Bisht et al 2013). The integrated al 2016) and aspect ratio values (Sams and Andrea 2001,
workflows of petrophysics and rock physics are used to find a Sams and Focht 2013). Thus, it is important to account for the
consistent rock physics model for the entire area of interest. volume and distribution of clay when trying to estimate the
Once a consistent rock physics model has been established, it elastic velocities.
can be effectively used to synthesize the elastic logs, identify The Lower Goru Formation of the Cretaceous age is a
the inconsistencies in the well logs and quick well data ana- proven reservoir in the middle and lower Indus Basin of
lysis, and improve seismic to well tie, which improves the Pakistan. In the Sawan gas field of the middle Indus Basin,
reservoir characterization and minimizes the risk of uncer- the sand intervals of this formation are composed of quartz,
tainty (Bisht et al 2013, Sams 2014). These calibrated models feldspar, volcanic rock fragments, chlorite, clay, glauconite,
also have the ability to accurately predict the variations in and minor amounts of calcite. The core and wireline log data
lithology and fluid saturation (Odegaard and Avseth 2004, show that the mineral composition of these sands is very
Mavko et al 2009, Grana et al 2012). The most important much heterogeneous from one well to another, which makes it
aspect of these integrated models is to use the accurate and difficult to use a single petrophysical model for reservoir
consistent mineral and fluid properties, which leads to an characterization. The aim of this study is to use an integrated
accurate calibration of these rock physics models (Mavko approach based on petrophysical analysis and a rock physics
et al 2009). Accurate and consistent rock physics models not model to solve the rock heterogeneity effects (Fitch
only efficiently differentiate between the hydrocarbon- and et al 2015, Ahmed et al 2016, Khalid et al 2016) on seismic
non-hydrocarbon-bearing zones but also indicate the pro- properties such as velocities and elastic moduli of the sand
blems present in the well log data such as borehole washouts, intervals. The rock physics model proposed by Xu and White
data gaps, mud filtrate, and insufficient log suites (Avseth (1995) is calibrated using wireline logs for the study area.
et al 2001).
Integrated rock physics models provide more accurate
and reliable links between petrophysics and seismic and 2. Geological setting
reservoir properties (Avseth 2000, Xu and Payne 2009, He
et al 2011, Khalid et al 2014b). A significant number of rock The study area is located in the north–south trending prolific
physics models have been proposed by various workers middle Indus Basin, which is bounded by the Indian shield in
(Mavko et al 2009). Avseth et al (2005) has classified these the east; Kirther Ranges, Sulaiman Fold, and Thrust Belt in
models into different categories such as inclusion models the west; Sargodha High in the north; and Jacobabad–
(Kuster and Toksoz 1974, Cheng and Toksöz 1979, Berry- Khairpur High in the south (Kadri 1995). The location of the
man 1980, Liu and Sun 2015), contact models (Mindlin 1949, study area and Jacobabad High is shown in figure 1. In the
Walton 1987, Dvorkin et al 1994, Dvorkin and Nur 1996), northwest direction of the study area, the famous Khairpur
transformations (Gassmann 1951, Berryman and Mil- High is located that exhibits very high geothermal gradient up
ton 1991), bounds (Voigt 1910, Reuss 1929, Hill 1952, to 4.8 °C/100 m. Khairpur High played an important role in
Hashin and Shtrikman 1963), and computational models. the formation of structural traps in the Kadanwari, Sawan, and
Recently, Khalid et al (2014) proposed a modified rock surrounding areas (Ahmad and Chaudhry 2002, Berger
physics model based on thermodynamic properties of reser- et al 2009). Since the Indus Basin is rich in hydrocarbon and
voir fluids at in situ conditions. contains several complete petroleum systems, a large number
Inclusion models ponder the rock as an elastic block of of wells have been drilled in the middle and lower Indus
minerals containing the pore spaces. Therefore, results of Basin. The organic rich black shales of the Sembar Forma-
these models show better consistency with the measured well tion, early Cretaceous in age, are the proven primary source
log data. Xu and White (1995) proposed a clay–sand mixture rock in the basin. In the middle and lower Indus Basin, the
model which is based on the Kuster and Toksoz (1974) Sembar Formation, with variable thickness (0–260 m), has
inclusion model, supplemented by the Gassmann (1951) and been deposited in a marine environment (Iqbal and
effective medium theories (Zhang 2008). This model can Shah 1980). Sembar Formation mainly consists of type-III
account for the effect of clay content on the seismic velocities, kerogen organic matter, which is mainly favorable for gas
and is very useful for estimating shear wave velocity. A generation. This formation is deeper and thermally mature

213
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 1. Location of the study area and other prominent geological

features.

Figure 2. Generalized stratigraphic column showing sand and shale

towards the western part, while it is shallower and less mature distribution in the Lower Goru Formation (modified after Krois
towards the eastern part of the Indus Basin (Wandrey et al 1998. Reproduced with permission from Pakistan Association
et al 2004). Sembar Formation is overlain by the Goru For- of Petroleum Geologists Islamabad).
mation, which is divided into two parts. The upper part is
mainly composed of shale and is termed the Upper Goru
whereas the lower part is termed the Lower Goru 3. Methodology
(Kadri 1995).
The medium to coarse grained sandstone of the Lower The sand intervals of the Lower Goru Formation in the Sawan
Goru Formation is the main reservoir rock in the middle Indus gas field are very heterogeneous, are considered as hot sands,
Basin, which is deposited in a shallow marine environment. and reflect a high value of gamma ray and sonic logs, which
This reservoir formation is composed of sandstone, siltstone, makes it complicated to accurately estimate the reservoir
inter-bedded shale, and thin bedded limestone (Kazmi and zones through conventional petrophysical analysis. The
Jan 1997, Berger et al 2009). The lower sandy part of this reservoir exhibits high temperature (175 °C) and pressure
formation has been further divided (from bottom to top) into (37.14 MPa) due to high geothermal gradient and burial
four stratigraphic intervals, i.e., A, B, C, and D as shown in depth. Wireline logs (e.g., DT, GR, SP, RHOB, NPHI, LLD,
figure 2 (Krois et al 1998). The B, C, and D intervals act as LLS, and LLS) from four wells (Sawan-01, 02, 03 and 07)
potential gas reservoirs in the study area (Ahmad et al 2004, were used for this study. Well Sawan-07 has a full suite of
Munir et al 2011). Petrographic analysis reveals that the A logs and the borehole condition also seems to be good in the
and B intervals are quartz arenite, whereas the C interval is target zone. In other wells, shear log is missing and borehole
sublithic to lithic arenite, which includes a significant amount conditions in the target zone are also not as good as in the
(almost 13%) of partially altered basic volcanic rock frag- Sawan-07 well. Log data from the Sawan-07 well along with
ments (McPhee and Enzendorfer 2004, Berger et al 2009). mineral and fluid properties have been used to calibrate the
The upper part of the Lower Goru Formation acts as a Xu–White (1995) rock physics model. The workflow of the
regional seal which is mainly composed of transgressive, methodology is shown in figure 3. In the first step, the pet-
siderite cemented shales and siltstones. Chlorite acts as rophysical parameters of the reservoir intervals are estimated
cement and comprises almost 80% of the clay fraction; this using PowerLog software. Volume of shale, effective poros-
significant amount of the chlorite decreases the porosity of the ity, and water saturation parameters have been estimated by
rocks (McPhee and Enzendorfer 2004). using the available well log data set through petrophysical

214
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 3. Schematic diagram of the technique used for improved reservoir characterization.

analysis. In a second step, these estimated petrophysical Petrophysical parameters such as volume of shale, effective
parameters along with fluid properties, solid mineral matrix porosity, and water saturation were estimated as follows.
values, and pore aspect ratios were utilized to calibrate the
proposed model. Then, modeled and measured velocities 4.1.1. Volume of shale. Proper estimation of shale content
were cross-plotted to check the degree of matching. In next provides the basis for accurately deriving the other
step, this calibrated model was used to estimate elastic para- petrophysical parameters such as porosity and water
meters in the remaining available wells. These estimated saturation in the shaly formation. Different shale indicator
elastic parameters have been cross-plotted to mark the litho- methods are used in practice, which are based on the
fluid zones and cutoff values of each elastic and petrophysical estimation of gamma ray index (equation (1)) from the
parameter. Finally, correlation between wells was performed gamma ray log:
to check the lateral and vertical variations in the target zone. GRlog - GRmin
IGR = , (1 )
GRmax - GRmin
where IGR is gamma ray index. GRlog represents gamma ray
4. Results and discussion log value at a particular depth, while GRmin and GRmax are
minimum and maximum values of gamma ray log in a given
The methodology proposed in the previous section is applied depth interval. However, this linear method overestimates the
on the wireline logs of four wells: Sawan-01, -02, -03, and volume of shale in real formations (Poupon and
-07. In this section, the petrophysical study of reservoir Gaymard 1970), whereas non-linear methods give more
intervals is presented firstly. Then the results of rock physics accurate results (Larionov 1969, Stieber 1970, Clavier
modeling are thoroughly discussed. The calibrated rock et al 1971). Therefore, we have used these non-linear
physics model and petrophysical analysis are used for reser- methods to estimate the volume of shale (equations (2)–
voir characterization in the reservoir intervals encountered in (4)). The value of IGR has been substituted in equations (2)–
the four wells of the study area. (4) to estimate the volume of shale:
Vsh (Larionov _ old) = 0.33 (2(2 ´ IGR) - 1) (2 )
4.1. Petrophysical analysis
IGR
Vsh (Stieber) = (3 )
Petrophysical analysis fills a gap between core and seismic 3-2 ´ IGR
data and plays an important role in reservoir characterization.
Vsh (Clavier) = 1.7 3.38 - (IGR + 0.7)2 . (4 )
The estimation of various petrophysical parameters from
wireline logs with accuracy can significantly enhance the
ability to interpret the lithology and reservoir characterization Here, Vsh(Larionov_old), Vsh(Stieber), and Vsh(Clavier) represent
(Fitch et al 2015). The quality of log data has a strong volume of shale measured using Larionov old rock, Stieber,
influence on the accuracy of rock physics models (Avseth and Clavier methods, respectively. The estimated shale
et al 2001). Conventional well log curves such as density, volumes are shown in figure 4. The purpose of using different
gamma ray, caliper, resistivity, neutron porosity, and sonic methods was to choose the method which gives lowest
are available in the Sawan-07 well. Considering the avail- volume of shale in order to minimize the risk of errors due to
ability of well log curves and borehole conditions, the Sawan- the presence of hot sands (sands including some content of
07 well was chosen as a reference well for this study. radioactive material, usually potassium or thorium, and

215
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

(FNC)2 + (FDC)2
Fe = . (7 )
2

4.1.3. Water saturation. As the presence of shale also affects

the water saturation, we compensate for the effect of shale by
applying the Indonesian model (Poupon and Levaux 1971).
This model improves the results reliability in shaly
formations, as it is based on field observations
(Widarsono 2012, Alao et al 2013). The mathematical form
of this model is
⎡⎧ ⎛ 2 - Vsh ⎞1/2 -1 / 2
⎛ m ⎞1/2 ⎫2 ⎤
f
S w = ⎢⎨ Rt ⎥
⎪ V ⎪
⎜ sh ⎟ +⎜ e⎟ ⎬ , (8 )
⎢⎪ ⎝ Rsh ⎠ ⎝ R ⎠ ⎪ ⎥
Figure 4. Cross plot between gamma ray index and volume of shale ⎣ ⎩ w ⎭ ⎦
calculated by different methods.
where Sw represents water saturation; Rsh, Rw, and Rt are the
shale, water, and true resistivities, respectively; m is
cementation factor and its value is 2.15; Vsh is volume of
showing high values of gamma ray log) or interbedded shales shale; and fe is effective porosity. These estimated
(Hussein and Ahmed 2012). In practice, the neutron-density petrophysical parameters (volume of shale, porosity, and
method is used to minimize the effect of radioactive minerals. water saturation) are shown in figure 5.
However, in some cases, the neutron-density method does not
provide accurate results, especially when hot sand is saturated
4.2. Rock physics modeling/calibration of rock physics model
with gas or light hydrocarbon fluids (Hamada 1996, Adeoti
et al 2009). In the case of dirty sand, modern spectral gamma Rock physics modeling is a process of finding an appropriate
ray is a more appropriate technique for computing volume of model that shows good consistency with the available well
shale. Since spectral gamma ray was not available, non-linear log data (Walls et al 2004). The proposed Xu–White (1995)
methods proposed by different authors (Larionov 1969, clay–sand mixing model is based on the Kuster and Toksoz
Stieber 1970, Clavier et al 1971) were used to calculate the (1974) model supplemented by the Gassmann (1951) and
volume of shale of the target interval. Figure 4 clearly pore aspect ratio theories. This model has the ability to
indicates that the volume of shale calculated using the Stieber separate the sand- and clay-related pores by assigning them
(1970) method is lower compared to other methods. More- different aspect ratios. If αS and αc are the aspect ratios of
over, the Stieber method is also suitable for gas reservoirs sand- and clay-related pores, ΦS and Φc are porosities of sand
(Adeoti et al 2009). Therefore, volume of shale calculated by grains and clay content, respectively. Then these sand and
using the Stieber method was utilized for further analysis. clay grains can be mixed through clay content in order to
calculate the elastic properties of dry rock porous media, as
shown in equations (9)–(11):
4.1.2. Porosity. Gas saturation near the wellbore affects the Kd - Km 1 ⎛ K - Km ⎞
porosity logs. In the gas-bearing zone the density porosity log = ⎜ f ⎟ å Fl Tiijj (al ) , (9 )
3Kd + 4m m 3 ⎝ 3Km + 4m m ⎠ l = s, c
shows high values whereas the neutron porosity log shows
low values, so a combination of neutron and density md - m m
porosities is used to remove this effect. However, in shaly 6md (Km + 2m m) + m m (9Km + 8m m)
formation, it is necessary to remove the effect of shale as the mf - mm
presence of shale affects the porosity. Equations (5) and (6)
= å Fl F (al) ,
25m m (3Km + 4m m) l = s, c
(10)

were used to remove the effect of shale on neutron and Tijij (al )
density porosities (Schlumberger 1974, Doveton 1999): F (al ) = Tiijj (al ) - , (11)
3
FNC = FN - Vsh FNsh , ( 5) where Kd, Km, and Kf represent bulk modulus of dry rock
frame, solid matrix, and pore fluid respectively, and μd, μm,
FDC = FD - Vsh FDsh , (. 6 ) and μf represent the corresponding shear modulus. Φ is the
porosity, while Tiijj(αl) and Tijij(αl) are the scalar functions
Here, ΦNC and ΦDC are the corrected neutron and density of the aspect ratio, which have been calculated using the
porosities; ΦN and ΦD represent neutron and density porosity Eshelby (1957) approach.
log, respectively; while ΦNsh and ΦDsh represent neutron and The elastic properties of clay are not well established in
density porosities in the shaly area, respectively. Finally, we this area and vary dramatically for different clay types. At the
combined these corrected porosities in order to calculate the initial stage of this model, we used the typical solid mineral
effective porosity Φe: values for clay and sand as proposed by Han et al (1986) and

216
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 5. Display of log curves and estimated petrophysical parameters in the Lower Goru Formation encountered in the Sawan-07 well.

Table 1. Parameters used for rock physics modeling.

Parameters Symbols Numerical Value Units

10
Bulk modulus of sand Ks 3.7×10 Pa
Shear modulus of sand μs 4.4×1010 Pa
Density of sand ρs 2650 Kg m−3
Bulk modulus of clay Kc 2.14×1010 Pa
Shear modulus of clay μc 1.19×1010 Pa
Density of clay ρc 2590 Kg m−3
Reservoir temperature T 448 K
Reservoir pressure P 37.14 MPa
Salinity of brine Sb 20 000 ppm
Specific gravity of gas Ggas 0.642 SG
Bulk modulus of gas Kg 7.97×107 Pa
Density of gas ρg 171.23 Kg m−3
Bulk modulus of brine Kb 3.17×109 Pa
Density of brine ρb 928.18 Kg m−3
Aspect ratio of sand αS 0.12 Unitless
Figure 6. Analysis of gas saturation versus bulk modulus of fluid Aspect ratio of shale αC 0.035 Unitless
estimated using Brie’s fluid-mixing algorithm.

the Brie et al (1995) mixing algorithm as shown in

Carmichael (1989), respectively. The Wyllie et al (1956) time equation (12):
average equation was used as a mixing algorithm to get the
KBrie = (Kb - Kg)(Sw)e + Kg, (12)
solid matrix properties. Batzle and Wang (1992) proposed
relationships were used to compute in situ fluid temperature where Sw represents water saturation; Kb and Kg are the bulk
and pressure. These fluid properties have been mixed using modulus of brine and gas, respectively; KBrie represents the

217
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 7. Cross plot between measured and calculated velocities

using reference well data (Sawan-07): (a) P-wave velocity, (b)
S-wave velocity; color coding represents measured depth. Figure 8. Comparison between measured and modeled elastic
properties to check the effectiveness of the model. (a) Measured
acoustic impedance versus VP/VS ratio. (b) Modeled acoustic
impedance versus VP/VS ratio. The data points are color coded by
gamma ray log.
bulk modulus of fluid calculated using Brie’s approach; and e
is the exponent of fluid mixing whose value varies from 1 to
40. When e=1, the mixing is Voigt’s average and when
e=40, the mixing results are very near to wood’s average
(Brie et al 1995). Figure 6 clearly shows that the value of Density of saturated rock was calculated using
fluid modulus (KBrie) increases with decrease in gas saturation equation (15):
(Sg) and vice versa. This behavior indicates that the stiffness rsat = Fr f + (1 - F) rm. (15)
of effective modulus of gas-bearing sediments decreases with
the increase in gas saturation. Here, Ksat and μsat are the bulk and shear modulus of the
The elastic moduli of saturated rocks were calculated saturated rocks, respectively, whereas ρsat, ρf, and ρm repre-
using the Gassmann (1951) fluid substitution model, which sent the densities of the saturated rock, pore fluid, and solid
gives the relationship between bulk modulus of saturated matrix, respectively. Finally, we substitute Ksat, μsat, and ρsat
rock, dry rock modulus, pore fluid, and solid matrix values in equations (16) and (17) to obtain the elastic velo-
(equations (13) and (14)): cities

⎛ 1 - Kd ⎞2 4msat
⎜ ⎟ Ksat +
⎝ Km ⎠ VP = 3 (16)
Ksat - Kd = (13) rsat
F ( 1 - F) K
+ - d2
Kf Km Km
msat
VS = . (17)
msat = md . (14) rsat

218
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 9. Cross plots of acoustic impedance versus VP/VS ratio developed using modeled elastic parameters of (a) Sawan-01, (b) Sawan-02,
(c) Sawan-03, and (d) Sawan-07 wells at various gamma ray values.

After building a rock physics model, the most important show very good consistency with the measured velocities
task is to adjust the input parameters for the specific reservoir. (figure 7).
Most of the default parameters failed to properly match the
modeled logs with the measured logs, as elastic properties of
4.3. Reservoir characterization
clay have not been clearly defined. Therefore, it is necessary
to adjust the elastic parameters (density, P and S wave Rock physics models are utilized as an important tool in
velocities) of clay to achieve an optimal fit with the measured reservoir characterization. They create the bridge between
logs within the target interval. For this purpose, we con- elastic properties (P-wave velocity, S-wave velocity, density,
structed a series of models to adjust the elastic values of clay impedance, and VP/VS ratio) and reservoir properties (por-
and aspect ratios. The typical values for sand and clay related osity, permeability, and saturation) (Avseth 2000, Chi and
Han 2009). The calibrated rock physics model has been uti-
pores are 0.12–0.15 and 0.02–0.0, respectively (Xu and
lized to estimate the elastic properties of reservoir interval
White 1995) which were tested in the calibration process. We
encountered in each well. The derived petrophysical para-
observed that if low values of elastic parameters were chosen,
meters from petrophysical analysis are cross-plotted against
then higher values of aspect ratios were required to achieve a the elastic properties computed from the rock physics model
good match between measured and modeled logs. Moreover, in the reservoir intervals. Cross-plotting of different para-
the value of aspect ratio decreases with the increase in clay meters is a powerful tool for visual analysis and helpful in
content and decrease in porosity. After trial and error, the final marking the data clusters in the target zones (White 1991).
values of aspect ratios and elastic parameters were selected Based on their response, these data clusters can be classified
(table 1). At these selected values, the modeled velocities into different lithologies/facies (Mavko et al 1998). Elastic

219
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 10. Cross plots of acoustic impedance versus VP/VS ratio developed using modeled elastic parameters of (a) Sawan-01, (b) Sawan-02,
(c) Sawan-03, and (d) Sawan-07 wells at different water saturation values. Data cluster with low acoustic impedance, VP/VS ratio, and water
saturation shows gas saturation.

attributes (VP/VS ratio) has the ability to discriminate between parameters, so it is more suitable and effective for differ-
different type of lithology and payable sand in the target zone entiating different type of facies as shown in figure 8. The
(Benzing et al 1983, Miller and Stewart 1990, Hughes facies against high GR values (shales) are clearly separated
et al 2008). However, a combination of P-impedance and VP/ from the facies of low GR values (sands). Keeping in mind
VS ratio can be utilized to efficiently predict the lithology and the effectiveness of the model, cross plots between VP/VS and
fluid saturation (Odegaard and Avseth 2004, Avseth and acoustic impedance (figure 9) have been developed in order to
Bachrach 2005, Chi and Han 2009). discriminate between the different type of facies in all wells.
To verify the effectiveness of the model, calculated and In each cross plot (figures 9(a)–(d)), different type of data
measured elastic impedance is cross-plotted against VP/VS clusters can be clearly identified in the graphs. The data
ratio for the entire reservoir interval (figure 8). The data points cluster present at the lower portion of the graph with low
are color coded using gamma ray log. Figures 8(a) and (b) are acoustic impedance, VP/VS ratio, and gamma ray values
totally different from each other. The cross plot between represents clean sand. However, the upper part of the cross
modeled parameters (figure 8(b)) clearly separates the dif- plots, where higher VP/VS ratio and gamma ray values are
ferent types of facies whereas cross plot between measured plotted in curve-like form, represent shale. The trends of these
(logs) parameters (figure 8(a)) fails to separate these facies. parameters are closely matched with the cross plots done by
Since the measured log data is affected by different para- Castagna et al (1993). Shaly sand exhibits different behavior
meters and environmental conditions, it is difficult to dis- than sand and shale, and it can be clearly identified at each
criminate fluid contents or lithology from log data. However, cross plot between these two sand and shale bodies. Shaly
in modeled data, we have more control over the input sand sediments exhibit gamma ray values between 80 to 120

220
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 11. Cross plots of acoustic impedance versus shear impedance calculated using modeled elastic properties of (a) Sawan-01, (b) Sawan-
02, (c) Sawan-03, and (d) Sawan-07 wells. Color coding represents water saturation. Data cluster with low water saturation, acoustic
impedance, and shear impedance represents gas-saturated zone.

Table 2. Quantitative values (in average) of elastic and petrophysical parameters for different identified rock types.

Rock Type
Parameters Gas-Bearing Sand Shaly Sand Shale
Gamma ray (API) <80 80–120 >120
Water saturation (V/V) <0.55 0.55–0.9 >0.9
Porosity (V/V) >0.11 0.05–0.11 <0.05
P-Impedance (Kg m−2*S) <1.02×107 1.02×107–1.2×107 >1.2×107
S-Impedance(Kg m−2*S) 6.2×106 6.2×106–7.2×106 >7.2×106
Vp/Vs ratio „1.65 >1.65–1.72 >1.72

API. On the other hand, if sand is saturated with gas, then the clearly seen that the gas-bearing sediments have VP/VS and
value of acoustic impedance, VP/VS ratio, and water satur- acoustic impedance values less than or equal to 1.65 and
ation parameters will be on the lower side (figure 10). The 1.02×107, respectively.
presence of gas strongly affects the elastic parameters and Rock physics attributes such as acoustic and shear
causes the acoustic impedance and VP/VS ratio values to impedances have been combined to discriminate fluid satur-
decrease (Zhao et al 2013). From figures 10(a)–(d), it can be ation. Water saturation is used to color code the data points

221
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Figure 12. Cross correlation of gamma ray and elastic properties of four offset wells representing the lateral and vertical variation in the
reservoir zone (units of acoustic and shear impedance are Kg/m2*S, Gamma ray gAPI, and depth in meters).

(figure 11). The cross plots of these attributes clearly sepa- the logs of P-wave impedance, S-wave impedance, gamma
rates the data points into three types of clusters. The cluster ray, and VP/VS ratio are plotted together for all four wells. For
having low water saturation, acoustic impedance, and shear this purpose, interpolation between wells was performed.
impedance values represents gas-bearing sand. However, the Four wells (Sawan-01, -02, -03, and -07) have been used to
cluster having high water saturation (100% saturation), analyze the litho-fluid behavior of the reservoir. The upper
acoustic impedance, and shear impedance values represents thick shale of Lower Goru Formation is considered as a
shale bodies with no gas saturation. The separation between datum plane. A SW–NE trending cross section (figure 12)
these two clusters (sand and shale) represents water saturation through four wells shows the correlation between calculated
(figure 11). The water-bearing sediments exhibit higher lithologies. Figure 12 shows that the upper part of the Lower
values of acoustic impedance and water saturation as com- Goru Formation is composed of thick shale intervals with
pared to gas-bearing sediments. These water-bearing sedi- interbedded thin intervals of shaly sand, and the lower part of
ments can also be clearly identified on the cross plots having the formation is mainly composed of reservoir-quality sand
elastic and water saturation values between gas-bearing sand with a shaly sand interval at the bottom. The analysis indi-
and shale bodies. On the basis of petrophysical and cross-plot cates that this sand interval is a gas-bearing reservoir in the
analysis, the average quantitative (cutoff) values of elastic and study area. The thickness of shale increases from the SW to
petrophysical properties for different rock types have been NE direction, whereas thickness of reservoir-quality sand
defined (table 2). The cross-plotting results show that our decreases in this direction. In this area, Khairpur High plays
calibrated model effectively predicts the lithology and fluid an important role in controlling the reservoir quality and
content in the Sawan gas field, Middle Indus Basin, Pakistan. stratigraphic traps (Ahmad and Chaudhry 2002). The uplift-
In order to develop a good understanding of the geo- ing of Khairpur High positioned the Lower Goru reservoir-
metry, depositional setting, and continuity of different strati- quality sand of proximal depositional system in structurally
graphic units encountered in different wells in the same area, deep areas. While non-reservoir-quality shale and shaly sands
well correlation is performed. Therefore, to analyze the lateral are positioned updip to form traps (Berger et al 2009), Sawan-
and vertical variations in the marked stratigraphic units 02 well shows reservoir compartmentalization and has rela-
encountered in the four wells (Sawan-01, -02, -03, and -07), tively poor reservoir character; this is due to the relatively

222
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

proximal location of the Sawan-02 well to Khairpur High Ahmad N and Chaudhry S 2002 Kadanwari gas field, Pakistan: a
compared to other wells. disappointment turns into an attractive development
opportunity Petrol. Geosci. 8 307–16
Ahmed N, Fink P, Sturrock S, Mahmood T and Ibrahim M 2004
Sequence stratigraphy as predictive tool in Lower Goru
Fairway, Lower and Middle Indus Platform, Pakistan Pakistan
5. Conclusions Association of Petroleum Geoscientist (PAPG), Annual
Technical Conference (ATC) 85–104
In this study, an integrated petrophysical and rock physics Ahmed N, Khalid P and Anwar A W 2016 Rock physics modeling
modeling approach is adopted to understand the reservoir to assess the impact of spatial distribution pattern of pore fluid
characterization of the Lower Goru Formation in the Sawan and clay contents on acoustic signatures of partially-saturated
reservoirs Acta Geod. Geophys. 51 1–13
Gas Field of the Middle Indus Basin, Pakistan. The developed Ajisafe Y C and Ako B D 2013 3D seismic attributes for reservoir
methodology helped us to build up an accurate and consistent characterization of ‘Y’ field Niger Delta, Nigeria IOSR
rock physics model. The calibrated model shows good con- J. Appli. Geol. Geophys. 1 23–31
sistency between measured and modeled velocities. The Alao P A, Ata A I and Nwoke C E 2013 Subsurface and
correlation between measured and modeled P and S wave petrophysical studies of shaly-sand reservoir targets in Apete
field, Niger Delta ISRN Geophys. 2013 102450
velocities is 92.76% and 84.99%, respectively (figure 7). The Avseth P 2000 Combining rock physics and sedimentology for
good consistency of the model laid a significant foundation seismic reservoir characterization of North Sea turbidite
for improved reservoir characterization in the study area. The systems PhD Thesis Stanford University, CA
calibrated model has also proven helpful in accurately esti- Avseth P and Bachrach R 2005 Seismic properties of unconsolidated
mating the elastic parameters (density, P and S wave velo- sands: tangential stiffness, Vp/Vs ratios and diagenesis 75th
Annual Meeting SEG Expanded Abstracts pp 1473–6
cities) even in those wells (Sawan-02) where both sonic and Avseth P, Mukerji T and Mavko G 2005 Quantitative Seismic
shear logs were missing in the target zones. Cross plots of Interpretation–Applying Rock Physics Tools to Reduce
these calculated parameters clearly delineate the lithology and Interpretation Risk (Cambridge: Cambridge University Press)
fluid content. On the basis of these cross plots, the quantita- Avseth P, Mukerji T, Jorstad A, Mavko G and Veggeland T 2001
tive values (average cutoff values) of elastic and petrophysical Seismic reservoir mapping from 3D AVO in a North
Sea turbidite system Geophys. Soc. Expl. Geophys. 66
parameters have been defined in order to discriminate 1157–76
between the gas-bearing sand, shale, and shaly sand zones. It Batzle M L and Wang Z 1992 Seismic properties of fluids
is found that the VP/VS ratio is more sensitive to gas-bearing Geophysics 57 1396–408
sand followed by acoustic impedance. Lateral and vertical Benzing W M, Byerly P E and Hopkins J R 1983 Shear and
variation in the reservoir is analyzed through cross-correlation compressional wave data interpretation—Midland basin, Texas
53rd SEG Annual Meeting, Las Vegas Expanded Abstracts
methodology. The correlation shows that thickness of shale pp 358–9
increases whereas quality of sand decreases from the SW to Berger A, Gier S and Krois P 2009 Porosity-preserving chlorite
NE direction. The proposed model allows for accurate dis- cements in shallow–marine volcanic clastic sandstones:
crimination between different types of facies and provides evidence of the Sawan gas field, Pakistan AAPG Bulletin 93
quick results. It can also be effectively utilized in seismic 595–615
Berryman J G 1980 Long-wavelength propagation in composite
inversion to improve seismic reservoir characterization. elastic media J. Acoust. Soc. Am. 68 1809–31
Beryman J G and Milton G W 1991 Exact results for generalized
Gassmann’s equations in composite porous media with two
constituents Geophysics 56 1950–60
Acknowledgments Bisht B S, Sas S K, Chaudhuri P K, Singh R B N and Singh S K
2013 Integration of petrophysics & rock-physics modeling in
The authors would like to express their appreciation to the single workflow reduces uncertainty in seismic reservoir
characterization: a case study Geohorizons 44–7
Directorate General of Petroleum Concession (DGPC),
Brie A, Pampuri F, Marsala A F and Meazza O 1995 Shear sonic
Pakistan for the release of data used in this study. We extend interpretation in gas bearing sands SPE Annual Technical
our appreciation to the China Scholarship Council for pro- Conference and Exhibition SE-30595-MS
viding us a chance to enhance our research abilities. We are Carmichael R S 1989 Practical Handbook of Physical Properties of
also thankful to Mr Shafiq-ur-Rehman (meteorologist at Rocks and Minerals (Boca Raton, FL: CRC Press)
Castagna J P and Swan H W 1997 Principles of AVO crossplotting
Pakistan Meteorological Department, Islamabad, Pakistan)
Lead. Edge 17 337–42
and our labmates working in the Geodetection Lab of China Castagna J P, Batzle M L and Kan T K 1993 Rock physics—the link
University of Geosciences, Beijing for their valuable sug- between rock properties and AVO response Offset-Dependent
gestions and fruitful discussions. Reflectivity–Theory and Practice of AVO Analysis (Tulsa, OK:
Society of Exploration Geophysicists) vol 8, pp 135–71
Cheng C H and Toksöz M N 1979 Inversion of seismic velocities for
the pore aspect ratio spectrum of a rock J. Geophys. Res. 84
References 7533–43
Chi X and Han D 2009 Lithology and fluid differentiation using rock
physics templates Lead. Edge 28 60–5
Adeoti L, Ayolabi E A and James P L 2009 An integrated approach Clavier C, Hoyle W and Meunier D 1971 Quantitative interpretation
to volume of shale analysis: Niger Delta example, offshore of thermal neutron decay time logs: I. Fundamentals and
field World Appl. Sci. J. 7 448–52 techniques J. Petrol. Technol. 23 743–55

223
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Doveton H 1999 Basic Oil and Gas Log Analysis Kansas Geological Larionov V V 1969 Radiometry of Boreholes (in Russian) (Moscow:
Survey USA 8–16 Nedra)
Dvorkin J and Nur A 1996 Elasticity of high-porosity sandstones, Liu Z S and Sun S Z 2015 The differential Kuster–Toksöz rock
theory for two North Sea data sets Geophysics 61 559–64 physics model for predicting s-wave velocity J. Geophys. Eng.
Dvorkin J, Nur A and Yin H 1994 Effective properties of cemented 12 839
granular material Mech. Master 18 351–66 Marion D, Nur A, Yin H and Han D 1992 Compressional velocity
Eshelby J D 1957 The determination of the elastic field of an and porosity in sand–clay mixtures Geophys. 57 554–63
ellipsoidal inclusion and related problems Proc. R. Soc. Lond. Mavko G, Mukerji T and Dvorkin J 1998 The Rock Physics
Ser. A 241 376–96 Handbook; Tools for Seismic Analysis in Porous Media 1st
Feng-Ying Y, Xing-Yao Y and Bo L 2014 S-wave velocity self- edn (New York: Cambridge University Press)
adaptive prediction based on a variable dry rock frame Mavko G, Mukerji T and Dvorkin J 2009 The Rock Physics
equivalent model J. Geophys. Eng. 11 045015 Handbook; Tools for Seismic Analysis of Porous Media 2nd
Fitch P J, Lovell M A, Davies S J, Pritchard T and Harvey P K 2015 edn (New York: Cambridge University Press)
An integrated and quantitative approach to petrophysical McPhee C A and Enzendorfer C K 2004 Sand management solutions
heterogeneity Mar. Pet. Geol. 63 82–96 for high-rate gas wells, Sawan field, Pakistan SPE
Gassmann F 1951 Uber die elastizitat poroser medien Vier. der International Symposium and Exhibition on Formation
Natur Gesellschaft 96 1–23 Damage Control 94–103
Grana D, Pirrone M and Mukerji M 2012 Quantitative log Mindlin R D 1949 Compliance of elastic bodies in contact J Appl.
interpretation and uncertainty propagation of petrophysical Mech. 16 259–68
properties and facies classification from rock-physics modeling Minear M J 1982 Clay models and acoustic velocities 57th Annual
and formation evaluation analysis Geophysics 77 45–63 Meeting American Institute of Mining and Metallurgical
Hamada G M 1996 An integrated approach to determine shale Engineers (New Orleans) Paper 11031-MS SPE
volume and hydrocarbon potential in shaly sand Presented at Munir K, Iqbal M A, Farid A and Shabih S M 2011 Mapping the
SCA Int. Symp. 2093–107 productive sands of Lower Goru Formation by using seismic
Hashin Z and Shtrikman S 1963 A variational approach to the elastic stratigraphy and rock physical studies in Sawan area, southern
behavior of multiphase materials J. Mech. Phys. Solids 11 Pakistan: a case study J. Petrol. Explor. Prod. Technol. 1
127–40 33–42
He F B, You J and Chen K Y 2011 Gas sand distribution prediction Odegaard E and Avseth P 2004 Well log and seismic data analysis
by elastic inversion based on rock physics modelling and using rock physics templates First Break 23 37–43
analysis Appl. Geophys. 8 197–205 Poupon A and Gaymard R 1970 The evaluation of clay content from
Hill R 1952 The elastic behaviour of a crystalline aggregate Proc. logs SPWLA 11th Annual Logging Symp. Conf. Paper 1970-G
Phys. Soc. London 65 349–54 Poupon A and Levaux J 1971 Evaluation of water saturation in shaly
Hu R Y, Holden T and Broussard M 2011 Petrophysics and rock formations Society of Professional Well Log Analysts 12th
physics modeling to improve seismic reservoir characterization Annual Logging Symp. Transactions paper SPWLA-1971-
—case study of lower hackberry sandstone AAPG Annual vXIIn4a1
Convention and Exhibition (Houston, Texas, USA) Article # Reuss A 1929 Berechnung der fliessgrense von mischkristallen auf
40774 grund der plastizitatbedingung fur einkristalle Zeitschrift fur
Hughes P, Eykenhof V R and Mesdag P 2008 Estimation of Ange-Wandte Mathematikaus Mechnik 9 49–58
Hydrocarbons in-Place by Simultaneous (AVO) Inversion, Sams M 2014 Constraining petrophysics with rock physics EAGE/
Constrained by Iteratively Derived Low Frequency Models FESM Joint Regional Conf. Petrophysics Meets Geoscience
(Stavanger, Norway: Fugro-Jason) (Kuala Lumpur, Malaysia)
Hussein R A and Ahmed M E B 2012 Petrophysical evaluation of Sams M S and Andrea M 2001 The effect of clay distribution on the
shaly sand reservoirs in Palouge-Fal oilfield, Melut Basin, elastic properties of sandstones Geophys. Prospect. 49 128–50
South East of Sudan. Engineering and computer sciences Sams M S and Focht T 2013 An effective inclusion-based rock
(ECS) J. Sci. Technol. 13 2 physics model for a sand–shale sequence First Break 31 61–71
Iqbal M W A and Shah S M I 1980 A A Guide to the Stratigraphy of Schlumberger Limited 1974 Log Interpretation: Applications 2
Pakistan 53 (Geological Survey of Pakistan) Stieber S J 1970 Pulsed Neutron Capture Log Evaluation—
Kadri I B 1995 Petroleum Geology of Pakistan 1st edn (Karachi, Louisiana Gulf Coast Society of Petroleum Engineers (https://
Pakistan: Pakistan Petroleum Ltd) doi.org/10.2118/2961-MS)
Kazmi A H and Jan M Q 1997 Geology and Tectonics of Pakistan Voigt W 1910 Lehrbuch der Kristallphysik (Leipzig: Teubner)
1st edn (Karachi, Pakistan: Graphic Publishers) Walls J, Dvorkin J and Carr M 2004 Well logs and rock physics in
Khalid P, Ahmed N, Naeem M and Khan K A 2014 A modeling seismic reservoir characterization Offshore Technology Conf.
study of AVO-derived attributes to differentiate reservoir Offshore Technology Conf.
facies from non-reservoirs facies and fluid discrimination in Walton K 1987 The effective elastic moduli of a random packing of
Penobscot area, Nova Scotia Geosci. J. 19 471–80 spheres J. Mech. Phys. Solids 35 213–26
Khalid P, Broseta D, Nichita D V and Blanco J 2014(b) A modified Wandrey C J, Law B E and Shah H A 2004 Sembar Goru/Ghazij
rock physics model for analysis of seismic signatures of low composite total petroleum system, Indus and Sulaiman-Kirthar
gas-saturated rocks Arab. J. Geosci. 7 3281–95 geologic provinces, Pakistan and India (No. 2208-C)
Khalid P, Ahmed N, Mahmood A and Saleem M A 2016 An Wang Q, Wang Y, Guo S G, Xing S T and Liu Z W 2015 The effect
integrated seismic interpretation and rock physics attribute of shale properties on the anisotropic brittleness criterion index
analysis for pore fluid discrimination Arab. J. Sci. Eng. 41 from laboratory study J. Geophys. Eng. 12 866
191–200 White R E 1991 Properties of instantaneous seismic attributes Lead.
Krois P, Mahmood T and Milan G 1998 Miano Field, Pakistan, a Edge 10 26–32
case history of model driven exploration Proc. Pakistan Widarsono B 2012 Choice of water saturation model in log analysis
Petroleum Convention, Pakistan Assoc. Petroleum Geologists, and its implication to water saturation estimates–a further
Islamabad pp 111–31 investigation Sci. Contrib. Oil Gas 35 3
Kuster G T and Toksoz M N 1974 Velocity and attenuation of Wyllie M R J, Gregory A R and Gardner L W 1956 Elastic wave
seismic waves in two-phase media: 1. Theoretical formulation velocities in heterogeneous and porous media Geophysics 21
Geophysics 39 587–606 41–70

224
J. Geophys. Eng. 14 (2017) 212 T Azeem et al

Xu S and Payne M A 2009 Modelling elastic properties in carbonate Zhang Z 2008 A rock physics model for hydrates bearing sediments
rocks Lead. Edge 28 66–74 of near surface SEG Las Vegas Annual Meeting pp 1799–803
Xu S and White R E 1995 A new velocity model for clay-sand Zhao L, Geng J, Han D H, Cheng J and Guo T 2013 Rock physics based
mixtures Geophys. Prospect. 43 91–118 probabilistic lithology and fluid prediction in a heterogeneous
Yuedong Q and Hongwei A 2007 Study of petrophysical parameter carbonate reservoir Annual SEG Meeting Houston (USA)
sensitivity from well log data Appl. Geophys. 4 282–7 pp 2387–91 (http://dx.doi.org/10.1190/segam2013-0486.1)

225