Journal of Geophysics and Engineering

J. Geophys. Eng. 13 (2016) 220–233 doi:10.1088/1742-2132/13/3/220

Rock-physics and seismic-inversion based

reservoir characterization of the Haynesville
Shale
Meijuan Jiang1 and Kyle T Spikes
The University of Texas at Austin, Department of Geological Sciences, Austin, TX, USA

E-mail: jmjzsx@gmail.com and kyle.spikes@jsg.utexas.edu

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Received 6 June 2015, revised 16 January 2016
Accepted for publication 10 February 2016
Published 5 April 2016

Abstract
Seismic reservoir characterization of unconventional gas shales is challenging due to their
heterogeneity and anisotropy. Rock properties of unconventional gas shales such as porosity,
pore-shape distribution, and composition are important for interpreting seismic data amplitude
variations in order to locate optimal drilling locations. The presented seismic reservoir
characterization procedure applied a grid-search algorithm to estimate the composition, pore-
shape distribution, and porosity at the seismic scale from the seismically inverted impedances
and a rock-physics model, using the Haynesville Shale as a case study. All the proposed rock
properties affected the seismic velocities, and the combined effects of these rock properties on
the seismic amplitude were investigated simultaneously. The P- and S-impedances correlated
negatively with porosity, and the VP/VS correlated positively with clay fraction and negatively
with the pore-shape distribution and quartz fraction. The reliability of these estimated rock
properties at the seismic scale was verified through comparisons between two sets of elastic
properties: one coming from inverted impedances, which were obtained from simultaneous
inversion of prestack seismic data, and one derived from these estimated rock properties. The
differences between the two sets of elastic properties were less than a few percent, verifying
the feasibility of the presented seismic reservoir characterization.
Keywords: reservoir characterization, rock physics, shale gas, seismic impedance, modeling

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

Introduction permeability, pressure, temperature, and total organic carbon

In exploration and production, seismic amplitude data is used
to predict reservoir properties for areas away from wells. The
seismic responses depend on impedance contrasts and vary
spatially. In seismic reservoir characterization, understanding
which rock properties or combination of rock properties cause
these spatial variations is important. For complex unconven-
tional gas shales, rock properties such as porosity, pore-shape
distribution, and composition are important to identify optimal
drilling locations.
Shales can have large variations in petrophysical and
elastic properties as a function of depth, thickness, porosity,
(TOC) (Curtis 2002, Roth 2011). Because of the heteroge-
neity and anisotropy, seismic reservoir characterization for
shales is challenging. Zhu et al (2011) used rock-physics rela-
tionships and seismic modeling to understand the geophysical
responses of shale-gas plays at both well log and seismic
scales. They found that the mineral composition and inter-
action among minerals affected geophysical responses and/
or rock strength and fracability of the rock. Koesoemadinata
et al (2011) introduced a workflow that involved seismic
acquisition, processing, and prestack inversion to char-
acterize the Marcellus Shale and classify the lithofacies.
They identified four major lithofacies (silts and sand, shale,

1
Currently at Shell, New Orleans, LA, USA.

1742-2132/16/030220+14$33.00 220 © 2016 Sinopec Geophysical Research Institute  Printed in the UK

J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

(a) (b)

Figure 1.  (a) Haynesville Shale (gray shaded area) is at the boundary of Texas and Louisiana. Light orange areas are basins, and yellow
areas are structural highs. The extent of the Sabine Uplift is marked by the dashed area. This figure was modified from Hammes et al

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

(2011). (b) Stratigraphic column for the Haynesville Shale. The Haynesville Shale is a Jurassic-aged rock formation (about 150 Ma) below
the Cotton Valley Group and above the Smackover Formation.

(a)

(b)
Figure 2.  (a) A map view showing the negative loop with maximum amplitude in the Haynesville Shale. The white block shows the seismic
survey for this study, and the two vertical wells are marked by the black star (Well A) and white star (Well B). The white horizontal ticks show
the crossline from 2001 to 2367, and white vertical ticks show the inline from 1001 to 1341. The thick white line marks the seismic profile
(crossline 2184) shown in (b). The black line marks the seismic profile shown in the seismic profile visualization example. The colors in the
white square indicate the maximum negative seismic amplitudes near the base of the Haynesville Shale. (b) Poststack P-wave seismic data. Blue
colors represent positive reflections, and red colors represent negative reflections. The seismic profile is marked by the thick white line in (a), and
the profile is oriented roughly along the east–west direction. Well A is in the middle of the seismic profile, at inline 1166, with the top and bottom
of the Haynesville Shale marked by blue arrows. The green line shows the horizon near the bottom of the Haynesville Shale. The correlation
coefficient of the seismic well tie was 0.79. The artificial two-way reflection time is used to maintain confidentiality of the data location.

limestone, and dolomite) and found that the Marcellus for-
mation in their study area was highly heterogeneous elasti-
cally. Sena et al (2011) developed a workflow that integrated
analysis of azimuthally varying prestack seismic data with
well data to estimate geomechanical properties of shales and
to identify sweet spots. They estimated the ratio of differ­
ential horizontal stresses, the pressure to initiate fractures,
and closure pressure, which are critical parameters for the
fracture simulation process. Guo et al (2013) c­ onstructed a
rock-physics workflow in which the clay and kerogen particle
orientations were defined, and they applied this ­workflow to
the Barnett Shale and evaluated porosity, lithology and brit-
tleness index. They found that Poisson’s ratio and Young’s
modulus along the well path decreased as clay content
­

221
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

2360

2370

2380

Artificial Depth m
2390

2400

2410

2420

2430

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

60 160 3 4 1.5 2 2.5 3 2.4 2.6
GR Vp Vs Den
(km/s) (km/s) (g/cc)

(a) (b)
Figure 3.  (a) Well log curves from the study well. Left to right are logs of gamma ray, P-wave velocity, S-wave velocity, and density. The
shaded area is the Haynesville Shale. Depth is artificial due to the data confidentiality. (b) A zoomed-in view of the well logs within the
Haynesville Shale from Well A. At an artificial depth of about 2418 m, velocity and density decrease.

increased. In addition, they found that Poisson’s ratio was

a more reliable indicator than Young’s modulus to determine
clay content and shale texture (orientation of clay particles)
at high porosities. Loseth et al (2011) built a relationship
between TOC and acoustic impedance and used this relation-
ship to predict TOC volumes from inverted acoustic imped-
ance volumes. They showed a significant nonlinear reduction
in acoustic impedance with increasing kerogen content, and
they mapped the source rock presence, thickness, and basin-
wide variations in TOC.
When performing seismic reservoir characterization, the
anisotropy of shales is an important aspect to be considered.
Sayers (2013) examined how anisotropy affected the direction-
ally dependent Young’s moduli and Poisson’s ratios for trans-
versely isotropic shales. He indicated that partial alignment of
clay particles and compliant pores affects the Poisson’s ratios
and Young’s moduli along different directions. The presence Figure 4.  Workflow for the seismic reservoir characterization.
Seismic processing (super gathering, bandpass filtering and partial
of silt inclusions increases the Young’s moduli and decreases
stacking) was applied to the prestack angle gathers. The angle
the Poisson’s ratios, and the presence of kerogen inclusions stacks were inverted for P- and S-impedances, which were then
decreases the Young’s moduli and Poisson’s ratios. Bachrach input into the rock-physics modeling and grid searching process to
et al (2013) applied rock-physics modeling in deterministic estimate rock properties at the seismic scale. To verify the reliability
and stochastic manners to predict anisotropy and constrain the of the estimated rock properties, the elastic properties derived from
the estimated rock properties were compared to the ones inverted
ranges of anisotropic parameters of shales and sandy shales.
from the prestack seismic data. Both sets of impedances were at the
Their anisotropy predictions were consistent with effective seismic scale.
medium theory, and they matched that rock model to an aniso­
tropic velocity profile from checkshot information. ratios. A workflow that accounted for heterogeneity and aniso­
The study presented here applied a grid-search algorithm tropy of shales was introduced to estimate the above rock
to derive rock-property estimations at the seismic scale for properties of the Haynesville Shale (Jiang and Spikes 2013a).
the Haynesville Shale from seismically inverted imped- The estimated compositions at the well log scale were similar
ances and a rock-physics model. At the well log scale, Jiang to the log data/core measurements. A preliminary study on
and Spikes (2011) investigated the pore shape (pore aspect applying that workflow at the seismic scale was done (Jiang
ratio) and composition effects on the P-wave velocity for the and Spikes 2013b) for a small number of seismic traces. In
Haynesville Shale and showed that shales with stiff composi- this study, the continuous rock properties at the seismic scale
tion and larger pore aspect ratios have higher velocities than were ­ estimated. Although the seismic-inversion algorithm
the ones that have soft composition and smaller pore aspect was isotropic, the rock-physics modeling was anisotropic.

222
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 5.  Prestack seismic data. (a) The original angle gathers at the well location (Well A), ranging from about 4° to 50°. The red curve
shows P-impedance from Well A. From small to large angle ranges, the S/N decreases, and the traces at the largest angles (30°–50°) appear
very noisy. (b) The angle gathers after applying the super gather process. The S/N increased, particularly at the very large angles (between
about 30° to 50°). (c) The angle stacks at the well location. Angle stacks were generated from angle ranges 5°–15°, 15°–25°, and 25°–35°.
The middle values for each of the angle ranges (10°, 20°, 30°) were assigned to the angle stacks, respectively. Partial stacking significantly
increased the S/N. The shaded area marks the Haynesville Shale.

Figure 6.  Prestack seismic inversion analysis at the Well A location. The first four panels show P-impedance, S-impedance, density, and
VP/VS. The blue curves indicate smoothed well logs, black curves indicate initial models, and red curves indicate seismic inverted results.
The fifth panel shows wavelets extracted from the partial stacked angle gathers at small, middle, and large angle ranges, the synthetic
seismic traces, partial stacked angle gathers from data, and the difference between the two. The black box indicates the Haynesville Shale,
to which the rock-physics modeling and grid searching was applied.

Haynesville description and data set
Description
The Haynesville Shale of the upper Jurassic is an organic-
rich mudrock located in the East Texas basin of east Texas
and northwestern Louisiana (figure 1(a)). Stratigraphically,
it sits between the Cotton Valley Group and the Smackover
Limestone (figure 1(b)). The Haynesville Shale was depos-
ited in a restricted basin, surrounded by carbonate platform
and silisiclastic shelves (Hammes et al 2011). The shale and
carbonates were deposited during a worldwide transgressive
event, in a quiet-water environment, below the wave base
(Hammes et al 2011). High temperatures and overpressure
were indicated in the burial and thermal history (Ewing 2001,
Becker et al 2010, Nunn 2011). The overpressure results in
a high initial production rate, as well as a dramatic decline
rate (Hammes et al 2011). The Haynesville Shale is 3000
m to 4000 m in depth. Average permeability is  <0.001 mD,
and porosity is 3–14% (Wang and Hammes 2010). The main

223
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

composition includes quartz, calcite, clay, relatively small

amounts of pyrite, and organic matter (kerogen). Horne et al
(2012) showed that the Haynesville Shale exhibits vertical
transverse isotropy (VTI) at the well log scale. At the lab
scale, the pores and clay platelets are mostly aligned along the
horizontal bedding direction (Curtis et al 2010, Chalmers et al
2012), supporting that the Haynesville Shale exhibits effective
VTI anisotropy.

Seismic data

The 3D seismic data area used in this work is about 4 km

by 4 km, and the survey is composed of 341 inlines and 367
crosslines, with a trace spacing of about 12 m (figure 2(a)).
Figure 2(b) shows an example of the poststack P-wave seismic

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

amplitude data along the profile (roughly east–west direction)
marked in figure  2(a). In this variable density plot, overlain
by wiggle traces, the blue colors (figure 2(b)) represent posi-
tive amplitudes, and red colors represent negative amplitudes.
The Haynesville Shale top and bottom are marked by the two
blue arrows at about 700 ms and 750 ms, respectively (the
two-way artificial time is used due to the confidentiality of
the true data). The Haynesville-to-Smackover interface causes
the large-amplitude reflection at ~750 ms. The Bossier-to-
Haynesville (~700 ms) reflection is not identified as easily due
to a relatively small contrast in impedance. From figure 2(b),
in the lower part of the Haynesville Shale (between 720 ms
and 740 ms), there is a relatively strong negative reflection
(red color). The color intensity changes from trace to trace,
indicating the amplitude of this negative reflection varies
spatially. Accordingly, the map view in figure  2(a) shows
the amplitude of this negative reflection. The amplitude was Figure 7.  The rock-physics modeling result. Crossplot of
extracted from the top to bottom of the Haynesville Shale. S-impedance versus P-impedance, colored by prior porosity
(a) and prior composition (b). When the composition case number
The spatial variation of the P-wave seismic data can be clearly increases, the quartz and calcite percentages decrease, and kerogen
seen from the different colors. Understanding which rock and clay percentages increase. Black points show the P- and
properties cause the spatial variations in the seismic amplitude S-impedances from the well log, gray points show the P- and
data is quite important for the exploration and production of S-impedances from the smoothed well log, and magenta points
the Haynesville Shale. show the P- and S-impedance from the seismic inversion at Well
A. The three sets of P- and S-impedances show similar trends,
which indicates that the models explain simultaneously the seismic
inverted P- and S-impedances.
Well log data
that the log amplitude variations were related to variations of
A vertical pilot hole through the entire Haynesville Shale for-
porosity, composition and pore aspect ratio (i.e. pore shape).
mation was drilled in the middle of the study area. The gamma
ray, sonic, caliper, density and resistivity logs were available
(Well A). The logs were edited using the caliper curve in an Method
attempt to eliminate questionable measurements corresp­
onding to local borehole conditions. However, uncertainty due The method for seismic reservoir characterization (figure 4)
to borehole conditions remains in some of the data. Curves included three major parts: rock-physics modeling, prestack
shown are gamma ray, velocity, both P-wave (VP) and S-wave seismic inversion, and value estimation from grid searching.
(VS) and density (figure 3(a)). Depth on the vertical axis is Through these procedures, the distributions of porosity, pore
artificial for confidentiality purposes. The gray area marks shape (pore aspect ratio), and composition at the seismic
the Haynesville Shale. The measurement variations within scale were jointly inverted from the measured logs and
the Haynesville Shale can be seen from the zoomed-in views inverted P- and S-impedances within the Haynesville Shale.
(figure 3(b)). At about 2418 m, there is a decrease in the VP, VS The rock-physics modeling included the self-­ consistent
and density values. These variations correspond to the nega- model and Chapman’s (2003) model. The self-consistent
tive reflection close to the base of the Haynesville Shale in the model (O’Connell and Budiansky 1974, Berryman 1980)
seismic data (figure 2(b)). Jiang and Spikes (2013b) showed initiated the numerical simulation with inclusions of grains

224
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 8.  The rock property estimations at the seismic scale. Each time sample had 100 estimations of the mineralogy as well as pore
shape and porosity. For the pore shape (pore aspect ratio) estimation (a) and porosity estimation (b), the background color represents
probability, where dark is high and light is low probability. Also shown is the mean of the estimates (blue) and the estimate associated with
the highest probability (magenta). In the porosity estimation plot (b), the well log data is in red, and the black curve is a smoothed version
of the red curve. (c) The composition estimation. Each time sample corresponds to different mineralogical percentages. (d) Verification of
the reliability of the estimated rock properties. Black solid and dashed curves are the P- and S-impedances from seismic inversion, and blue
solid and dashed curves are the P- and S-impedances derived from the mean values of the estimated rock properties.

and pores with a variety of aspect ratios and volumes.
Chapman’s (2003) model provided a vertical transversely
isotropic (VTI) medium by introducing aligned micro-
fractures. The rock-physics model was calibrated using
the measured well log data. Before inversion, the seismic
data was processed to increase the S/N. The inverted P- and
S-impedances were then input into grid searching steps
to estimate rock properties at the seismic scale. The grid-
search algorithm is described later in the paper. Finally, the
elastic properties derived from the estimated rock proper-
ties were compared to the ones inverted from the prestack
seismic data to verify the reliability of the estimated rock
properties.
Rock-physics modeling
In the rock-physics modeling, the effective shale was built
based on different mineral and pore phases. The mineral
phases included quartz, calcite, clay, kerogen and pyrite. The
pore phases included non-spherical/spherical pores, non-
spherical cracks and aligned fractures. The detailed descrip-
tion and calibration of the rock-physics model is described in
Jiang and Spikes (2013a).
The rock-physics model transformed prior distribu-
tions of mineralogy, porosity and pore shape to elastic
properties. The prior composition included quartz, calcite,
pyrite, kerogen and clay. Based on core analysis and XRD

225
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 9.  The prestack seismic inverted P-impedance (a), S-impedance (b), and VP/VS (c) within the Haynesville Shale. The solid curve
shows the inverted results at Well A, and the dotted curve shows the corresponding smoothed log at Well A. The cold colors indicate lower
values, and hot colors indicate higher values. Each of the inverted elastic properties varies both vertically and laterally. They all show a
decrease at the bottom of the Haynesville Shale between about 720 ms and 740 ms. The horizontal dashed curve shows the horizon for the
map views in figures 12–15. The horizon approximately shows the low P-impedance near the bottom of the Haynesville Shale.

measurements, pyrite comprised only a small portion of the
composition, and it was fixed at 2% in the rock-physics mod-
eling. For the other compositions (quartz, calcite, kerogen
and clay), 400 scenarios (case 1 to case 400) were intro-
duced, and each scenario had a composition assemblage
with certain percentages of each mineral. The 400 scenarios
provided a wide enough composition range while main-
taining low computational cost. In the 400 scenarios, the
quartz and calcite content were negatively correlated with
the kerogen and clay content. From case 1 to case 400, the
quartz and calcite percentages decreased linearly, and the
kerogen and clay percentages increased linearly. Therefore,
the higher the case number, the more compliant the rock
is; the lower the case number, the stiffer the rock is. The
prior porosity was from 0 to slightly more than 0.4, and the
prior pore aspect ratio was from 0 to 0.12. The ranges of
prior porosity and pore aspect ratio were large enough to
contain most combinations expected in this interval. Jiang
and Spikes (2013a) showed that at the well log scale, the
porosity and composition estimations matched the observa-
tions quite well. For example, for a well (outside the seismic
survey) that is about 3 km away from the vertical pilot hole,
the average modeled porosity was 4.6% compared to density
porosity of 4.2%. Meanwhile, the average modeled quartz,
calcite, kerogen and clay were 31.3%, 23.3%, 4.7% and
38.7% compared to XRD results of 31.6%, 25.8%, 5.3%
and 35.3%, respectively.
Prestack seismic inversion
Simultaneous inversion (Hampson et al 2005) of prestack seismic
data was applied to obtain P-impedance, S-impedance and den-
sity. Three assumptions are central to this algorithm (Hampson
et al 2005): a linearized approximation for reflectivity, the angle-
dependent reflectivity can be calculated by Aki–Richards equa-
tions  (Aki and Richards 2002), and the natural logarithms of
P-impedance, S-impedance and density are linearly related.
The angle gathers at the well location (figure 5(a)) have
incidence angles from about 4° to 50°. Overall, these angle
gathers have a relatively low S/N. In particular, the traces
at large incidence angles (between about 30° to 50°) were
noisier than the traces at smaller angles. A seismic-well-tie
was done based on poststack seismic data that has much
higher S/N than the prestack seismic data. Supergather anal-
ysis and partial stacking were applied to increase the S/N.
The number of common depth point (CDP) gathers added
in the super gathers was 25, which increased the S/N (part­
icularly for the large angle gathers between about 30° to
50°), and preserved the main features shown in the original
angle gathers (figure 5(b)). Partial stacking further enhanced
the S/N (figure 5(c)). Three angle ranges were selected:
5°–15° (near), 15°–25° (middle), and 25°–35° (far). Angles
from 35°–50° were excluded due to relatively low S/N and
nonexistence below about 700 ms. For each of the three angle
ranges, the middle angle value was assigned to the angle

226
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 10.  (a) Upper panel: porosity estimation. The hot colors indicate large values, and cold colors indicate smaller values. The black
solid curve shows the inverted P-impedance. The porosity estimation is negatively correlated with P-impedance and S-impedance. Lower
panel: relative standard deviation of the porosity estimation, showing the estimation uncertainty. The black dashed line indicates Well A.
(b) Upper panel: Pore-shape distribution estimation. The hot colors indicate large values, and cold colors indicate smaller values. The
black solid curve shows the inverted VP/VS. The pore-shape distribution is correlated negatively with VP/VS. Lower panel: relative standard
deviation of the pore shape (pore aspect ratio) estimation, showing the estimation uncertainty. The black dashed line indicates Well A.

stack, e.g. the angle stack from angle range 5°–15° had an
angle value of 10°. In the simultaneous inversion, these three
angle stacks (10°, 20°, and 30°) were treated as prestack angle
gathers.
Grid search
A grid searching method was used to estimate the rock prop-
erties and the associated uncertainties at each angle gather
location. In the grid searching, P- and S-impedances from the
rock-physics models were compared with the ones inverted
from the seismic data. The modeled impedances that satisfied
the criteria (L1 norm)
Obj_P = IP model − IP inverted ⩽ 0.2
(1)
Obj_S = IS model − IS inverted ⩽ 0.1
provided the estimated rock properties. The value 0.2 and
0.1 for P- and S-impedances were chosen to make their
errors 2–2.5%. This limit was chosen to account for physi-
cally reasonable possibilities and to ensure the computational
efficiency.
For each trace at each time sample, the grid searching
provided a large number of estimated rock properties based
on the objective function (equation (1)). The mean value and
standard deviation were calculated for each rock property at
each sample. The standard deviation shows the uncertainty of
the estimations laterally and vertically.
Comparison of the estimated elastic properties
The spatially continuous rock properties were simultaneously
estimated by inputting the prestack seismic inverted elastic
properties into the rock-physics modeling and grid searching.
For each of the rock properties, both the estimation and uncer-
tainty (standard deviation) were calculated. The reliability of
these rock properties was verified further through the com-
parisons between the elastic properties derived from them and
the ones inverted from the prestack seismic data. Based on
the estimated rock properties at the seismic scale, the rock-
physics model generated a set of elastic properties. Those
elastic properties should be close to the ones from the prestack
seismic inversion if the estimated rock properties are reliable.
Results
Prestack seismic inversion
Analysis of the prestack inversion results at the well loca-
tion is shown in figure 6. In the first four panels, the well log
P- and S-impedances, density, and VP/VS are plotted in blue,

227
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 11.  (a) Upper panel: clay fraction estimation. The hot colors indicate large values, and cold colors indicate smaller values. The
black solid curve shows the inverted VP/VS. The clay fraction is correlated positively with VP/VS. Lower panel: relative standard deviation
of the clay fraction estimation, showing the estimation uncertainty. The black dashed line indicates Well A. (b) Upper panel: quartz fraction
estimation. The hot colors indicate large values, and cold colors indicate smaller values. The black solid curve shows the inverted VP/VS.
The quartz fraction is correlated negatively with VP/VS. Lower panel: relative standard deviation of the quartz fraction estimation, showing
the estimation uncertainty. The black dashed line indicates Well A.

the initial models are plotted in black, and the inverted results
are plotted in red. To utilize the poststack data that had higher
S/N and generate a proper initial model, the poststack inverted
P-impedance was chosen as the initial P-impedance model. The
initial density and VP/VS models were derived from log data. The
Haynesville Shale is marked by the black box. Due to the large
and abrupt impedance contrast at the base of the Haynesville
(about 40% increase in about 10 m), it is difficult to obtain an
accurate inversion result at the bottom of the Haynesville Shale.
The density match between inversion and well log data is not as
good as impedances because (1) the density log itself fluctuates
due to the rugose borehole environ­ment; (2) the seismic traces
at large angles have low S/N. The inverted VP/VS has some misfit
with the borehole measurements but fits the log measurements
within the Haynesville Shale fairly closely. The synthetic angle
stacks (red seismograms) are similar to the seismic data (black
seismograms). The correlation coefficient is about 0.90, with
relatively small error (red seismograms on right). The similarity
between the synthetic seismic traces and observed seismic data
indicates that the inversion results are acceptable.
Calibration of the procedure at the well location
The seismic inverted P- and S-impedances were then input
into the rock-physics modeling. The modeling result (figure 7)
shows that the rock-physics models explained the seismic
inverted P- and S-impedances simultaneously. In figure 7, the
crossplot of S-impedance versus P-impedance is colored by
the prior porosity (figure 7(a)) and prior composition (figure
7(b)). The black points are from the well logs, gray points are
from the smoothed well logs, and magenta points are from
the seismic inversion at the well location. The three sets of
P- and S-impedances show similar trends. The model trends
explained all the data points quite well.
Using the seismic reservoir characterization procedure, the
rock properties at the well location were estimated (figure 8).
At each time sample, we obtained multiple estimates of pore
aspect ratio, porosity and composition assemblages based on
the objective function in the grid searching, and then randomly
selected 100 estimates. The probability of each estimate was
obtained (background color in figures 8(a) and (b)). The mean
value of the estimates (blue curves in figures 8(a) and (b)) and
the value with the highest probability of the estimates (magenta
curves in figures  8(a) and (b)) are similar, while the mean
value curve is smoother. There is an increase of both porosity
and pore aspect ratio below 720 and 740 ms. The estimated
porosity at the top part of the Haynesville Shale is similar to
the density porosity log (red curve, figure 8(b)), particularly to
the smoothed density porosity (black curve, figure 8(b)). The
mismatch at the bottom part of the Haynesville Shale could

228
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 12.  A horizon visualization showing the prestack inverted P-impedance (a), S-impedance (b) and VP/VS (c) close to the base of the
Haynesville Shale (black dashed curve in figure 9). In each panel, the hot colors represent large amplitudes, and cold colors represent small
amplitudes. Both the inverted P- and S-impedances show low amplitudes in the north area and large amplitudes in the southeast area. The
VP/VS shows spatial variation, but the patterns are not necessarily the same as shown from P- or S-impedance maps. The black line marks
the seismic profile shown in figures 9–11.

be related to the challenge to obtain accurate seismic inver-
sion results due to the very large and abrupt impedance con-
trast (about 40% increase in about 10 m) from the Haynesville
Shale to the Smackover Limestone. The average value of the
estimated porosity indicated by the blue curve is 5.4% within
the Haynesville Shale, and the average density porosity from
the well log is 6.1%. The composition assemblage (figure
8(c)) is from the mean value of the estimated composition at
each time. Each time corresponds to various percentages of
the minerals. Calculations of the vertically averaged miner-
alogy is 30.8% for quartz, 22.8% for calcite, 2% for pyrite,
4.8% for kerogen and 39.6% for clay. These numbers are close
to the ones from core measurements from a well that is about
3 km away, which are 31.6%, 25.8%, 2%, 5.3%, and 35.3%,
respectively.
Additional steps were applied to verify the estimated rock
properties. Based on the estimated rock properties, the cali-
brated rock-physics model provided a set of elastic properties
(P- and S-impedances). Those derived elastic properties were
then compared with the ones from the prestack seismic inver-
sion. The black set of impedances (figure 8(d)) represents the
ones from seismic inversion, and the blue set of impedances
(figure 8(d)) represents the ones derived from the estimated
rock properties. These two sets of P- and S-impedances are
very similar, indicating that the estimated rock properties are
reliable.
Seismic profile visualization
The seismic reservoir characterization procedure is a trace-by-
trace operation, and it provides both seismic profile visualiza-
tion and horizon-based visualization. Figure 9 shows a seismic
profile representation of the inverted P-impedance (figure
9(a)), S-impedance (figure 9(b)) and VP/VS (figure 9(c)).
The location of the seismic profile is marked by the black
line in figure  2(a), with Well A (inline 1166) in the middle
of the seismic profile. The hot colors indicate large values,
and cold colors indicate small values. Spatial variability is
clearly present in each section. Vertically, each inverted para­
meter shows a decrease at the base of the Haynesville Shale
between about 720 ms and 740 ms. Laterally, the color inten-
sity changes among different inlines.
The estimated porosity (figure 10(a)) varies vertically and
laterally. Vertically, the porosity is larger between about 720 ms
and 740 ms relative to the other time ranges. This relatively
larger porosity is consistent with the low P- and S-impedances

229
J. Geophys. Eng. 13 (2016) 220 M Jiang and K T Spikes

Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021

Figure 13.  The estimated rock properties. In each panel, the hot colors indicate large values, and cold colors indicate smaller values.
Those rock properties are all correlated to the elastic properties shown in figure 12. (a) Porosity estimation, which is generally negatively
correlated with P-impedance and S-impedance. (b) Pore-shape distribution estimation, which is negatively correlated with VP/VS. (c) Quartz
fraction estimation, which is negatively correlated with VP/VS. (d) Clay fraction estimation, which is positively correlated with VP/VS. The
black line marks the seismic profile shown in figures 9–11.

observed from the inversion result at the same time range.
The porosity is smaller on the top and very bottom of the
Haynesville Shale, which is consistent with the large inverted
P-impedances at the top and very bottom of the Haynesville
Shale. Laterally, the color intensity of the porosity estima-
tion varies at different locations. For example, between about
720 ms and 740 ms, the porosity has large values between
inline 1130 and 1140, at inline 1150, 1170, and 1200, and
small values at 1140 and 1180. In fact, the porosity affects
both P- and S-impedances. Generally, large porosity corre-
sponds to low P- and S-impedances and vice versa. The rela-
tive standard deviation (figure 10(a)) is 20% to 25%.
The pore-shape distribution is represented by the pore
aspect ratio. Similar to porosity, the pore-shape distribution
also varies vertically and laterally (figure 10(b)). These varia-
tions correlated negatively with the VP/VS (figure 9(c)): higher
pore aspect ratios correspond to lower VP/VS, and vice versa.
In particular, between about 720 ms and 740 ms, the hot spots
in the pore-shape distribution between inline 1130 and 1140,
at inline 1150, 1170, and 1200 correspond to the cold spots
in the VP/VS (figure 9(c)), respectively. Similar to the porosity
estimation, the relative standard deviations (figure 10(b)) of
the pore-shape distribution is about 20%.
The estimates of composition include fractions of clay, quartz,
calcite, and kerogen. Clay fraction estimation (figure 11(a))
and quartz fraction estimation (figure 11(b)) are shown as
examples. Similar to the pore-shape distribution, the clay
and quartz fraction distributions appear to be correlated with
VP/VS. More clay corresponds to larger VP/VS, and less clay
corresponds to smaller VP/VS. More quartz corresponds to
smaller VP/VS, and less quartz corresponds to larger VP/VS.
The relationship between clay and VP/VS is consistent with the
fact that clay has higher VP/VS than the other minerals in the
Haynesville Shale. Similarly, the relationship between quartz
and VP/VS is consistent with the fact that quartz has lower
VP/VS than the other minerals in the Haynesville Shale. The
relative standard deviations for the composition estimation are
very small (figures 11(a) and (b)).
Horizon-based visualization
A horizon close to the base of the Haynesville Shale (black
dashed curve in figure  9) is shown in terms of prestack
seismic inverted P-impedance (figure 12(a)), S-impedance
(figure 12(b)) and VP/VS (figure 12(c)). Well A is in the middle
of the view (inline 1166, crossline 2184). In each panel, the

230
J. Geophys. Eng. 13 (2016) 220
Figure 14.  A horizon visualization showing the P-impedance (a), S-impedance (b) and VP/VS (c) derived from the estimated rock properties
from figure 13 close to the base of the Haynesville Shale (black dashed curve in figure 9). In each panel, the hot colors represent large
values, and cold colors represent small values. These elastic properties are very similar to the ones inverted from seismic data (figure 12).
The black line marks the seismic profile shown in figures 9–11.
hot colors represent large values, and cold colors represent
small values. P- and S-impedances show similar spatial varia-
tions. In the north area, P- and S-impedances are relatively
small, and in the southeast area, P- and S-impedances are rela-
tively large. The VP/VS also shows spatial variation, whereas
the variation pattern is different from P- and S-impedances. In
general, the VP/VS in the north and west areas are higher than
the ones in the south and east areas.
Figure 13 shows the estimates of porosity, pore shape
(pore aspect ratio) distribution, quartz fractions and clay frac-
tions. These rock properties are all correlated with the P- and
S-impedances and VP/VS. The porosity estimation (figure 13(a))
is relatively higher at the north area and lower at the south-
east areas. These spatial variation patterns in porosity are
negatively correlated with the ones in P- and S-impedances.
Similarly, the pore shape (pore aspect ratio) distribution and
quartz fraction is negatively correlated with VP/VS, and the
clay fraction is positively correlated with the VP/VS.
The elastic properties were then computed from the esti-
mated rock properties (figure 14) using the rock physics
model. Those elastic properties are very similar to the
inverted elastic properties in figure  12. Residuals between
the rock-property-derived elastic properties and the seismic-
inverted elastic properties are all close to zero (figure 15). The
231
M Jiang and K T Spikes
Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021
similarities between the two sets of elastic properties indicate
that the estimated rock properties are reliable.
Discussion
The rock properties were estimated at the seismic scale, and
they were correlated with the seismic impedance and VP/VS
variations. Some correlations are explicit, such as the negative
correlation between porosity and P- and S-impedances, the
negative correlation between quartz fraction and VP/VS, the
positive correlation between clay fraction and VP/VS, the posi-
tive correlation between pore shape and quartz fraction, and
the negative correlation between pore shape and clay fraction.
Some correlations are implicit, e.g. the relationship between
composition and impedances, and the relationship between
the pore shape and impedances. The rock property estima-
tion result shows a low clay fraction and high quartz frac-
tion between about 720 ms and 740 ms (figure 11), whereas
P- and S-impedances are relatively low in the same time range
(figure 9). In horizon visualization, the rock property esti-
mation result shows low clay fraction and high quartz frac-
tion toward the east of Well A (figure 13), whereas P- and
S-impedances are low in the same areas (figure 12). These
J. Geophys. Eng. 13 (2016) 220
Figure 15.  A horizon visualization showing the residual of P-impedance (a), S-impedance (b) and VP/VS (c) close to the base of the
Haynesville Shale (black dashed curve in figure 9). Those residuals are the difference between the rock-property derived elastic properties
and the seismic-inverted elastic properties. These maps show that the residuals are very close to zero (gray color).
relationships seem contrary to the normal case in which the
stiffer quartz matrix has higher impedance and the softer clay
matrix has lower impedance. From microstructure images
(Chalmers et al 2012), the majority of the porosity in the
Haynesville Shale sits in between stiffer mineral grains (i.e.
quartz and calcite) instead of clay platelets. Accordingly,
the estimated porosity is higher where the quartz fraction is
higher (and clay fraction is lower), and vice versa. The higher
porosity generates the lower P- and S-impedances. Similarly,
the above relationship between composition and porosity
helps to interpret the relationship between pore-shape distri-
bution and impedances. The explicit and implicit relationships
between rock and elastic properties imply the need to investi-
gate simultaneously the rock properties when interpreting the
seismic attribute variations. These interpretations also rely on
the prior distributions or the correlations between them and
the weight that has been put on P- and S-impedances during
the rock-physics modeling.
In the seismic reservoir characterization workflow pre-
sented here, there are three major parts: rock-physics mod-
eling, seismic inversion, and value estimation from grid
searching. Each of these affects the accuracy of characteri-
zation. The rock-physics model accounted for the anisotropy.
The grid searching results depended on the objective function
(equation (1)). On the one hand, if the maximum values for
232
M Jiang and K T Spikes
Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021
Obj_P and Obj_S were too small, then the constraint would
be too harsh. On the other hand, if the maximum values for
Obj_P and Obj_S were too large, then there would be almost
no constraint, and the estimated rock properties would be
quite random. The objective function used in this study was
the L1 norm, and other types of objective functions might pro-
vide different estimations of the rock properties.
The accuracy of the seismic inversion is very important. If
the elastic properties were not inverted correctly, then the esti-
mated rock properties would also be inaccurate, and it would
be difficult to correlate the rock properties with the seismic
attributes. For our case, the impedance contrast at the bottom is
very large and abrupt (about 40% increase in about 10 m), due
to the formation change from shale to limestone. Therefore, it
is difficult to obtain an accurate inversion result at the bottom
of the Haynesville Shale. In addition, the seismic-inversion
algorithm is isotropic, whereas the rock-physics modeling
is ­anisotropic. On the one hand, it is ­difficult to perform
­anisotropic seismic inversion with a lack of anisotropy meas-
urements at both the log scale and laboratory scale. On the other
hand, theoretically, the anisotropic features would mainly be
related to the far offset signals in the seismic data, whereas in
our case, the data quality is poor in the far offsets (figure 5).
Therefore, isotropic seismic inversion was performed in this
study. Anisotropic seismic inversion could be performed if
J. Geophys. Eng. 13 (2016) 220
there were anisotropy measurements, and the seismic data
quality at far offsets was better.
Conclusion
In this study, a seismic reservoir characterization proce-
dure was developed. The rock properties were estimated at
the seismic scale through rock-physics modeling, prestack
seismic inversion, and grid searching. Different distribu-
tions of the rock properties helped to interpret the seismic
amplitude variations. All these rock properties affected the
seismic velocities, and their effects on the seismic ampl­
itude were investigated simultaneously. Generally, the P- and
S-impedances are negatively related to porosity, and VP/VS is
positively related to clay fraction distribution, and negatively
related to pore shape (pore aspect ratio) and quartz fraction
distributions. Based on the estimated rock properties, it is sug-
gested to do hydraulic fracturing in areas with higher porosity,
larger pore aspect ratio, more quartz, and less clay.
The presented seismic reservoir characterization procedure
for the Haynesville Shale considered the complex mineral and
pore phases, as well as anisotropy. Each step (rock-physics
modeling, prestack seismic inversion and value estimation
from grid searching) in the workflow affected the accuracy of
the rock property estimations, and the reliability of the estima-
tions were verified. The characterization helped to interpret
the correlations between the seismic amplitude variations and
the rock properties. The same seismic reservoir characteriza-
tion procedure could be applicable to other unconventional
reservoirs using appropriate calibrations.
Acknowledgments
Acknowledgment is made to the donors of the American
Chemical Society Petroleum Research Fund for support (or
partial support) of this research. The Exploration and Devel-
opment Geophysics Education and Research (EDGER) Forum
at The University of Texas at Austin also supported this work.
BP provided seismic, well log and core data for this study, for
which we are grateful. Thanks also go to Dr Ursula Hammes
at the Bureau of Economic Geology (BEG) for providing core
measurements.
References
Aki K and Richards P G 2002 Quantitative Seismology 2nd edn
(San Francisco, CA: Freeman)
Bachrach R, Osypov K, Nichols D, Yang Y, Liu Y and
Woodward M 2013 Applications of deterministic and
stochastic rock physics modelling to anisotropic velocity model
building Geophys. Prospect. 61 404–15
Becker S P, Eichhubl P, Lauback S E, Reed R M, Lander R H
and Bodnar R J 2010 A 48 m.y. history of fracture opening,
temperature, and fluid pressure: cretaceous travis peak
formation, East Texas Basin Geol. Soc. Am. Bull. 122 1081–93
Berryman J 1980 Long-wavelength propagation in composite elastic
media I. Spherical inclusions J. Acoust. Soc. Am. 68 1809–19
233
M Jiang and K T Spikes
Chalmers G R, Bustin R M and Power I M 2012 Characterization
of gas shale pore systems by porosimetry, pycnometry, surface
area, and field emission scanning electron microscopy/
transmission electron microscopy image analyses: examples
from the Barnett, Woodford, Haynesville, Marcellus, and Doig
units AAPG Bull. 96 1099–119
Chapman M 2003 Frequency-dependent anisotropy due to meso-
scale fractures in the presence of equant porosity Geophys.
Prospect. 51 369–79
Curtis J 2002 Fractured shale-gas systems AAPG Bull. 86 1921–38
Curtis M E, Ambrose R J, Sondergeld C H and Rai C S 2010
Structural characterization of gas shales on the micro- and
nano-scales Proc. of the Canadian Unconventional Resources
and International Petroleum Conference, ‘Society of Petroleum
Engineers’ SPE Paper 137693
Ewing T E 2001 Review of Late Jurassic depositional and potential
hydrocarbon plays, northern Gulf of Mexico Basin Gulf Coast
Assoc. Geol. Soc. Trans. 51 85–96
Downloaded from https://academic.oup.com/jge/article/13/3/220/5113370 by guest on 17 October 2021
Guo Z, Li X-Y, Liu C, Feng X and Shen Y 2013 A shale rock
physics model for analysis of brittleness index, mineralogy and
porosity in the Barnett Shale J. Geophys. Eng. 10 025006
Hammes U, Hamlin H S and Thomas E E 2011 Geologic analysis
of the Upper Jurassic Haynesville Shale in East Texas and West
Louisiana AAPG Bull. 95 1643–66
Hampson D P, Russell B H and Bankhead B 2005 Simultaneous
inversion of pre-stack seismic data 75th SEG Meetings SEG
Expanded Abstracts pp 1633–6
Horne S, Walsh J and Miller D 2012 Elastic anisotropy in the
Haynesville Shale from dipole sonic data First Break
30 37–41
Jiang M and Spikes K T 2011 Pore-shape and composition effect
on rock-physics modeling in the Haynesville Shale 81st SEG
Meetings Expanded Abstracts vol 30 p 2079
Jiang M and Spikes K T 2013a Estimation of reservoir properties of
the Haynesville Shale by using rock physics modeling and grid
searching Geophys. J. Int. 195 315–29
Jiang M and Spikes K T 2013b Correlation between rock properties
and spatial variations in seismic attributes for unconventional
gas shales—a case study on the Haynesville Shale 83rd SEG
Meetings Expanded Abstracts
Koesoemadinata A, El-Kaseeh G, Banik N, Dai J, Egan M,
Gonzalez A and Tamulonis K 2011 Seismic reservoir
characterization in Marcellus shale 81st SEG Meetings
Expanded Abstract
Loseth H, Wensaas L, Gading M, Duffaut K and Springer M 2011
Can hydrocarbon source rocks be identified on seismic data?
Geology 39 1167–70
Nunn J 2011 Burial and thermal history of the Haynesville Shale:
implications for gas generation, overpressure, and natural
hydrofracture (abs.) AAPG 2011 Meeting Abstracts Houston
90124
O’Connell R and Budiansky B 1974 Seismic velocities in dry and
saturated crack solids J. Geophys. Res. 79 5412–26
Roth M 2011 North American shale gas reservoirs—similar, yet
so different AAPG Convention (Calgary, Canada, 12–15
September)
Sayers C M 2013 The effect of anisotropy on the Young’s
moduli and Poisson’s ratios of shales Geophys. Prospect.
61 416–26
Sena A, Castillo G, Chesser K, Voisey S, Estrada J, Carcuz J,
Carmona E and Hodgkins P 2011 Seismic reservoir
characterization in resource shale plays: stress analysis and
sweet spot discrimination Leading Edge 30 758–64
Wang F P and Hammes U 2010 Effects of reservoir factors on
Haynesville fluid flow and production World Oil 231 D3–6
Zhu Y, Liu E, Martinez A, Payne M A and Harris C E 2011
Understanding geophysical responses of shale-gas plays
Leading Edge 30 332–8