Practical aspects of AVO modeling

YONGYI LI, Paradigm Geophysical, Calgary, Canada

JONATHAN DOWNTON, Veritas, Calgary, Canada,
YONG XU, Arcis Corporation, Calgary, Canada

A VO (amplitude variation with offset)

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

modeling contributes significantly in

seismic data acquisition design, and
prestack seismic data processing and
interpretation. It has become a common
practice in prestack seismic analysis. It
is also used for verifying and develop-
ing AVO theories. Conducting AVO
modeling is, by nature, an exercise of
multidiscipline integration. It thus
enhances reservoir characterization and
reduces the risk in hydrocarbon explo-
ration.
This paper presents practical
aspects of AVO modeling with the goal
of better understanding and more effec-
tively using AVO modeling in both
prestack data processing and interpre-
tation. Numerous examples demon-
strate the applications of AVO modeling.
We present approaches for generating,
processing, and analyzing AVO syn- Figure 1. Seismic rock properties derived from a set of dipole sonic logs from the Western Canadian
thetic data, and include the commonly Sedimentary Basin. Data points of the gas sand, oil-saturated sand, and overlying and underlying
used methodologies of single-interface shale are highlighted in red, green, black, and pink squares, respectively. Notice that rock properties
modeling, single-gather modeling, 2D in different domains have different sensitivity responding to fluid. The contrast between gas sand
and overlying shale indicates a class 1 AVO response. These crossplots can be used as templates to
stratigraphic modeling, and 2D full- interpret inverted elastic rock properties. The lines in black, blue, and red are the empirical relation-
wave elastic-equation modeling. We ships for shale, brine-saturated sand, and gas-charged clean sand.
illustrate the applications of these meth-
ods to P-wave data, converted-wave
data, and elastic rock property inver-
sion. In addition, we discuss some com-
monly encountered issues such as
tuning and the effects of noise contam-
ination in AVO processing and analy-
sis. Finally, we demonstrate the role of
AVO modeling in calibrating prestack
seismic processing and in assisting data
interpretation.

Fundamentals in AVO modeling.

Seismic rock properties are directly
responsible for seismic wave propaga-
tion and seismic responses. They may
be catalogued as basic rock seismic
properties (P- and S-wave velocities and
density, and VP/VS ratio and Poisson’s
ratio), impedances, modulus rock prop-
erties (bulk modulus K, shear modulus
µ, Lamé’s constant λ), and anisotropic
rock properties.
Often, rock properties and corre-
sponding AVO responses can be dis-
cerned from well-log data. To
demonstrate this, various rock proper-
ties were calculated using the log data
from the Western Canadian Sedimentary
Basin (WCSB) and are displayed in the Figure 2. Petrophysical analysis of a well from the Western Canadian Sedimentary Basin. Shale
plots in Figure 1, with the empirical rela- volume has direct effects on water saturation and porosity, which consequently results in changes
tionships for shale (solid black line), in elastic seismic rock properties and seismic responses.

MARCH 2007 THE LEADING EDGE 295

 used in reservoir characterization.
The petrophysical rock properties
that directly relate to seismic rock
properties are volume fractions of
mineralogies, porosity, and water sat-
uration (SW). Fluid type, gas/oil ratio
(GOR), oil and gas gravity (in API),
and brine salinity (in ppm) form
another set of important petrophysi-
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

cal properties. Fluid properties can

be calculated based on derived or in-
situ measured pressure (P) and tem-
perature (T). Predicting seismic rock
properties, especially shear-wave
velocity, is important in rock physics
analysis. The steps to accomplish this
usually include determining in-situ
conditions, calculating fluid proper-
ties, solid properties, and finally cal-
culating seismic rock properties. The
link between seismic rock properties
(bulk modulus and shear modulus)
and petrophysical rock properties
(porosity, fluid type, water saturation,
and mineral composition) can be seen
in these Gassmann equations:

(1)

(2)

where Kdry = effective bulk modulus

of the dry rock; Ksat = effective bulk
modulus of the rock with pore fluid;
K0 = bulk modulus of mineral mate-
rial making up the rock; Kfl = effec-
tive bulk modulus of the pore fluid;
φ = porosity; µdry = effective shear
modulus of the dry rock; and µsat =
effective shear modulus of the rock
with pore fluid. In Equation 1, K0, φ,
and Kfl are often calculated using
petrophysical logs. Notice that
Equation 2 indicates that fluid type
and saturation do not affect shear
modulus of rock.
brine-saturated sand (solid blue line), and gas-charged clean Numerous empirical relationships for seismic rock prop-
sand (solid red line) overlain. Figure 1 demonstrates that the erties, especially for shear sonic prediction, have been pub-
velocities and impedances do not provide sufficient dis- lished. They are useful when either petrophysical properties
crimination as to whether the reservoir is gas-charged. are not available or rock physics theories do not address the
However, the VP/VS ratio, Poisson’s ratio, λρ, and λµ ratio do. complexity of rock properties. Commonly used empirical
Further, as expected, the data points of the oil-saturated sand relationships are the P- and S-velocity linear relation for
(green squares) fall on the empirical line for the brine-satu- water-saturated clastics, also called the mudrock line
rated sand. In comparison with the gas sand, the lower imped- (Castagna et al., 1985); the velocity-porosity-clay relation
ance of the oil-saturated sand (red squares) indicates that it (Han, 1986); the critical porosity model for Kdry (Nur et al.,
has higher porosity. The lithology contrast leads one to expect 1995); the Greenberg-Castagna relation for mixed lithologies
a class 1 AVO anomaly at the top of the gas sand because the (Greenberg and Castagna, 1992); the velocity relation for
impedance of the gas sand is significantly higher than that clay-sand mixtures (Xu and White, 1996); and the empiri-
of the shale above it (black squares). cal velocity relations for carbonates (e.g., Li and Downton,
Petrophysical analysis is an important aspect in AVO 2000). Often used petrophysical empirical relations are the
modeling because the information can be used to assess density-velocity relation (Gardner et al., 1974); the sonic
reservoir conditions. The relationships between petrophys- and porosity relation (Wyllie et al., 1956); and the resistiv-
ical and seismic rock properties can also be established and ity-velocity relation (e.g., Faust, 1953). While the above

296 THE LEADING EDGE MARCH 2007

 water saturation (SW) were derived
from resistivity, gamma ray (GR),
density porosity, neutron porosity,
and local petrophysical parameters.
The observations from this petro-
physical evaluation include:

(1) water saturation is proportional to

Vsh
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

(2) gas saturation increases with

decreasing Vsh
(3) porosity increases with decreas-
ing Vsh

Figure 3, based on this petro-

physical analysis, shows how the vol-
ume of clay influences seismic rock
properties. It is notable that clean
sand is most sensitive to gas satura-
tion since the distance between the
clean sand data points and the empir-
ical lines is greatest. The sensitivity or
distance decreases with increasing
shale volume. Further, similar to what
is shown in Figure 1, the elastic rock
properties of the oil-saturated sand
again fall on the wet-sand empirical
lines. This example shows that petro-
physical analysis is useful for under-
standing the relationship between
reservoir quality and seismic rock
properties and, consequently, seismic
responses.

AVO equations and attributes. To

illustrate the relationship between
rock physical properties and seismic
reflections, we will use a two-layer
interface with incident P-waves,
reflected and transmitted P-waves,
and converted S-waves as shown in
Figure 4. The rock properties in the
layers are P- and S-wave velocities,
density, impedances, bulk modulus,
shear modulus, Lamé’s parameter,
and the attenuation coefficient Q. The
anisotropic properties are P- and S-
wave velocity anisotropic parameters
ε, λ, and δ (Thomsen, 1986). The rock
properties listed above are of great
interest in prestack seismic inversion.
In AVO modeling, they are often
obtained or derived from well-log
data.
Two main steps of AVO modeling
are to generate synthetic gathers,
CMP gathers or common-image gath-
ers (CIG), by using the exact solution
and then to extract AVO attributes by
using approximate equations. AVO
exact solutions are often generated
empirical relationships are useful, the empirical relationships by using the Zoeppritz equations with ray tracing and a
derived locally are strongly recommended. Further, cali- plane-wave assumption, or by using the full wave elastic
bration using available well logs may improve rock prop- equation with a finite-difference method.
erty predictions. Aki and Richards (1980) simplified the Zoeppritz equa-
Figure 2 shows a set of petrophysical log curves from a tions into a form that can be used to solve for meaningful
WCSB well. Shale volume (Vsh), effective porosity (φE), and reflectivities. The difference between the Zoeppritz exact

298 THE LEADING EDGE MARCH 2007

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 3. Influence of shale volume on elastic rock properties. Labels 1–3

represent the data points of the gas sand, oil sand, and oil-saturated shaly Figure 5. Crossplots of AVO attributes calculated from well logs. Dashed
sand. The black, blue, and red lines are the empirical relationships for wet lines indicate regional background trends. Arrows indicate the directions
sand, shale, and clean gas sand. The data points of the oil-saturated sand in which the data points shift when the rocks are saturated by hydrocar-
that fall on the empirical line for wet sand indicate the oil sand and the bon. Fluid factor can be obtained using the crossplot of RP and RS by
wet sand have similar rock properties. Based on the distance to the empiri- muting the data points at and near the regional trend. AVO classes can be
cal line for wet sand, the sensitivity of the sand responding to gas satura- identified in the crossplot of RP and gradient. The magnitude of density
tion decreases with increasing shale volume. reflectivity is much smaller than P-wave reflectivity.

AVO equation from usually five to no more than three. In

an HTI medium, Ruger’s two-term equation (for a single
set of fractures) with an elliptical assumption of amplitude
is often used and the attributes to be determined are frac-
ture orientation and fracture density.
Commonly used AVO attributes and their relations are
listed in Table 3. Some of these relations are exact and oth-
ers approximate. Using Table 3, an AVO attribute can be
derived from other AVO attributes by multiplying an oper-
ator vector to an attribute vector. For example, fluid factor
can be derived by multiplying operator vector (1, -b/γ) and
attribute vector (∆VP/VP, ∆VS/VS). This yields ∆F = (1, -b/γ)
(∆VP/VP, ∆VS/VS) - 1 = ∆VP/VP - (b/γ) ∆VS/VS, where γ is
Figure 4. Reflections and transmissions at an interface. The rock proper- VP/VS ratio. RP can be calculated from RS by the relationship
ties above and below the interface determine the energy splitting. RP = b/γ RS and gradient G can be calculated from RP and
RS by G = RP - 2RS. Notice that the fluid indicators, gradi-
solution and the Aki-Richards’ approximation is small when ent, and fluid factor increase with decreasing VP/VS ratio.
zero-offset reflectivity << 1. Rearranged forms of the Aki- Figure 5 shows the crossplots of AVO attributes calcu-
Richards equation are often used for describing and extract- lated from dipole sonic logs. The dashed lines highlight the
ing AVO attributes. Commonly used approximate P-wave regional background trends, and the arrows show the direc-
AVO equations are Shuey (1985) for P-impedance reflectiv- tions in which the data points will shift when brine in rock
ity Rp and gradient G; Smith and Gidlow (1987) for P- and is replaced by hydrocarbons. Figure 5a shows P-wave reflec-
S-velocity reflectivities RVp and RVs; Fatti et al. (1994) for P- tivity RP against S-wave reflectivity. Figure 5b, the crossplot
and S-impedance reflectivities RP and RS; and Hilterman and of P-wave reflectivity RP against gradient G, is the one in
Verm (1995) for P-impedance reflectivity RP and Poisson’s which AVO types for a gas-charged clastic reservoir can be
reflectivity Rσ. (Table 1) determined. In Figure 5d we can see that ∆µ/µ and ∆λ/λ have
AVO equations for converted S-wave (e.g., Jin et al., magnitudes approximately twice RP or RS.
2000), a vertical transverse isotropic (VTI) medium Figure 5c shows the plot of RP against density reflectiv-
(Thomsen, 1993), and a horizontal transverse anisotropic ity ∆ρ/ρ. This information is useful when density reflectiv-
(HTI) medium (Ruger, 1996) are listed in Table 2. Using con- ity is considered robust in discriminating lithologies or
verted-wave data, density reflectivity and shear-wave veloc- reservoirs. In general, the values of density reflectivities are
ity or impedance reflectivity can be calculated. An approach much smaller in comparison to RP. To demonstrate this, we
that solves P- and S-wave reflectivities and density reflec- derived the equation ∆FP = f (∆VP/VP) - ∆ρ/ρ based on
tivity simultaneously is called simultaneous P-P and P-S Gardner’s relation ρ = dVPf. Since f has a value of about 1⁄4,
inversion (e.g., Larson, 1999). Including both converted- density reflectivity is about one quarter of P-wave velocity
and P-P wave data in AVO inversion has the potential to reflectivity.
obtain a more stable estimation of density reflectivity. Few Figure 6 shows a set of synthetic gathers corresponding
efforts have been made to extract AVO attributes in VTI to class 1–4 AVO responses from the top of a reservoir. The
media because of the difficulties in simplifying anisotropic class 1 AVO response (higher-impedance gas sand) has a

MARCH 2007 THE LEADING EDGE 299

 false fluid-factor anomalies. For example, reflection from a
coal layer can produce a fluid-factor anomaly similar to that
from a gas-sand reservoir. A carbonate layer encased by
lower-impedance clastic rocks such as shale produces a peak
at the top of the carbonate layer followed by a trough at the
base (Figure 7). Ideally, a carbonate line should be used in
calculating fluid factor for a porous and hydrocarbon-filled
carbonate reservoir encased by tight carbonates. Further,
when multilayer interbedded carbonate and shale are pre-
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

sent, fluid-factor anomalies often become too complex to

interpret. In this circumstance, an understanding of local
geology becomes important to eliminate the false fluid-fac-
tor anomalies. In addition, a fluid-factor response may
appear as a single peak or a single trough when seismic rock
property contrast at the top or base of a reservoir is not strong
enough to produce a fluid-factor anomaly.
In the following section we will demonstrate attribute
inversion in AVO modeling and analysis by using a synthetic
CMP gather. Figure 8 shows the example in which the low-
impedance reservoir produces a class 3 AVO anomaly. We
compared the AVO attributes extracted from the CMP gather
to the attributes directly calculated from the well logs (and
consequently called “reference” AVO attributes). The dif-
Figure 6. Class 1–4 AVO responses to gas-charged clastic reservoirs.
ferences between the extracted attributes and the reference
attributes are mainly due to offset-dependent tuning, NMO
stretch, the angle range used in AVO extraction, and the
approximations introduced in the simplified Zoeppritz equa-
tions.
Figure 8a shows angle-limited stacks. The stack with the
angle range of 20–30° has a larger amplitude correspond-
ing to the top and base of the reservoir. It also shows that
the full-offset stack differs from the zero-offset stack (RP).
This explains why P-reflectivity extracted from CMP gath-
ers rather than from a migrated structure (full-offset) stack
should be used as input in impedance inversion. For this
typical example, P-impedance inversion using the full-off-
set stack would underestimate the impedance and give an
overly optimistic prediction of porosity.
P- and S-reflectivities extracted using Fatti’s equation are
shown in Figure 8b. The P-reflectivity has a trough corre-
sponding to the top of the reservoir since the reservoir has
low P-impedance. However, the S-reflectivity has a peak
since the reservoir has high S-impedance. The fluid factor,
Figure 7. Relationship between mudrock line, carbonate line, and fluid as expected, exhibits a trough and a peak corresponding to
factor in P- and S-wave velocity crossplot space. The drop in P-wave the top and base of the reservoir.
velocity due to gas saturation produces a fluid-factor trough that corre- The attributes extracted using Smith and Gidlow’s equa-
sponds to the top of the reservoir. In a clastic-carbonate environment, a tion are shown in Figure 8c. ∆VP/VP and ∆VS/VS look simi-
fluid-factor peak corresponds to the top of the carbonate. lar to the RP and RS extracted by Fatti’s equation, telling us
that the velocity-density relations may follow Gardner’s
peak at zero-offset that deceases with offset and may change relation.
polarity at far offset. A class 2 AVO response (a near-zero The gradient extracted using Shuey’s equation is shown
impedance contrast) has a weak peak that decreases with in Figure 8d. The negative gradient corresponds to the reflec-
offset with a polarity reversal at far offset, or a weak trough tion from the top of the reservoir, indicating an increase in
that increases in amplitude with offset. A class 3 AVO amplitude with incident angle (trough brightening). The
response (lower-impedance gas sands) has a trough at zero positive gradient corresponds to the base of the reservoir,
offset that brightens with offset. Finally, a class 4 AVO also indicating an increase in amplitude with incident angle
response (lower-impedance gas sands) has a trough that (peak brightening).
dims with offset. Figure 8e shows Poisson’s reflectivity extracted using
Fluid factor is a hydrocarbon interface indicator. It is the Hilterman and Verm’s equation. At reservoir level, the
weighted difference between ∆VP/VP and ∆VS/VS or between Poisson’s reflectivity appears similar to the gradient and
RP and RS (Table 3). Figure 7 explains how fluid factor and fluid factor; that is a trough followed by a peak at the top
P- and S-wave velocities are related. It shows that the data and base of the reservoir, respectively.
points from the gas sand, with lowered P-wave velocity due The last panel (Figure 8f) shows λ reflectivity, ∆λ/λ and
to gas saturation, deviate from the regional background µ reflectivity, ∆µ/µ extracted using Gray’s equation. These
trend. This results in a trough in fluid factor followed by a reflectivities were scaled by 0.5 because their magnitudes
peak corresponding to the top and base of the gas zone. are approximately twice P- or S-reflectivity. Note that ∆λ/λ
Lithologies that do not follow the regional trend produce has a strong trough and peak corresponding to the top and

300 THE LEADING EDGE MARCH 2007

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 8. Synthetic CMP gather and commonly used AVO attributes extracted from the gather: (a) angle stacks and full offset stack; (b) P- and S-
impedance reflectivity (RP and RS), and fluid factor ∆F; (c) P- and S-velocity reflectivity (RVp and RVs), and fluid factor ∆F; (d) gradient G; (e)
Poisson’s reflectivity, and (f) λ- and µ-reflectivity (Rλ and Rµ). The subscript r refers to the synthetic traces that were calculated directly from well logs.

mask or induce an AVO anomaly. In AVO analysis, two

types of tuning must be considered: tuning due to thickness
of a reservoir and offset-dependent tuning due to amplitude
variation with offset. Both affect the true AVO expression
at an interface. In addition, offset-dependent tuning can be
further affected by the differential traveltime of adjacent
reflections. To demonstrate offset-dependent tuning, we
derived an equation to calculate the tuning effect on a pri-
mary reflection based on offset-dependent Ricker wavelets:

(3)

Figure 9. Examples of offset (angle) dependent tuning calculated using where A1 and A2 are the zero-offset reflectivities corre-
Equation 3 and Shuey's two-term AVO equation and Equation 3: (a) sponding to two adjacent interfaces. Indices p and i repre-
class 1 and (b) class 3 AVO responses. The lines in red and blue are AVO sent primary reflection and the interfering reflection,
responses without tuning. respectively. θp and θi are the incident angles at the upper
and lower interfaces; b is the thickness of the layer, and L0
base of the reservoir. is the wavelength of dominant frequency. The tuning effects
In practice, by examining the extracted AVO attributes of a class 1 AVO and a class 3 AVO for a reservoir with VP
as described above, one may determine which attribute(s) = 3674 m/s and VS = 2055 m/s were calculated and shown
to use in defining a reservoir. in Figure 9. The lines in red and blue represent the tuning-
free AVO responses from the top and base of the reservoir.
Tuning effects and NMO stretch. Tuning has a significant The AVO responses with ratio b/L0 equal to 1, 1⁄2, 1⁄4, and 1⁄8
impact on seismic resolution and AVO responses. It may are shown in Figure 9. A significant change in amplitude

MARCH 2007 THE LEADING EDGE 301

 polarity at far offsets. The AVO response becomes a trough
at zero offset and then dims with offset when the thickness
was reduced to 10 m.
Notice that reducing frequency will result in a similar
effect as tuning. Therefore, frequency content in a data set
is essential in achieving an accurate AVO analysis.
NMO-corrected gathers may suffer significant differen-
tial moveout, which distorts the waveform and bandwidth
especially at far offsets that correspond to large incident
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

angles. In AVO processing, far-offset traces are often muted

to limit the impact of NMO stretch. This imposes a constraint
in the AVO attribute extraction that needs far-offset infor-
mation, such as density reflectivity inversion using P-wave
data. Researchers have been attempting to develop meth-
ods to combat NMO stretch. Swan (1997) developed a
method by nulling NMO velocity error at a given two-way
traveltime on a CMP gather. In addition, residual moveout
due to anisotropy may also need to be corrected before AVO
Figure 10. Tuning effects on a class 3 AVO anomaly from a low-imped- attribute extraction.
ance unconsolidated sand reservoir: (left) synthetic gathers with varying
reservoir thickness, and (right) extracted amplitudes at the top of the
reservoir. The AVO character is preserved until the thickness reaches to 1 Effects of noise contamination. In AVO attribute extraction,
m. The CMP gathers were muted at 45°. CMP gathers or common image point gathers obtained
through amplitude preserving processing are modeled by
AVA equations to solve for AVO attributes. This is often con-
ducted using angles of incidence and linear fitting of AVO
amplitudes by L2 norm (the least squares) or L1 norm meth-
ods. The difference between L2 and L1 norms is that the for-
mer minimizes squared deviation while the latter minimizes
the absolute deviation of the data. L1 norm is robust in
eliminating coherence noise such as multiples. Ferré et al.
(1999) proposed an approach to optimize AVO attribute
extraction by using the least median of squares method
(Walden, 1991). This method is similar to L1 norm method
but with a progressive converging process to achieve an opti-
mized solution.
Signal-to-noise ratio, fold of a CMP gather, and offset
range all affect the results of an AVO analysis. Offset-depen-
dent attributes such as S-reflectivity and gradient are most
influenced by these factors. In investigating the effects of
noise, random and/or coherent noise can be added to syn-
thetic gathers. A CMP or CIP gather with random and coher-
ent noise can also be generated using elastic wave-equation
Figure 11. Tuning effect of a class 1 AVO anomaly from a gas reservoir modeling. Figure 12 shows an example in which random
with low porosity and high impedance: (left) synthetic gathers with vary- noise was added in a CMP gather with a signal-to-noise ratio
ing reservoir thickness, and (right) extracted amplitude at the top of the of 1:0.2. AVO extraction used Fatti’s equation. As expected,
reservoir. The overlying logs on CMP gathers are gamma ray (yellow) noise has little effect on the estimation of P-reflectivity. The
and P-impedance (red). The CMP gathers were muted at 45°. The AVO
character has a significant change when the thickness reduces to 10 m.
estimated S-reflectivity at reservoir level, however, experi-
ences significant distortion.

can be observed when b/L0 equals 1⁄4 and 1⁄8. This illustrates AVO modeling methodologies. The Zoeppritz equations or
that AVO analysis performed on reflectivities that suffer the Aki-Richards equation with ray tracing, and full elastic
from tuning will be inaccurate. wave equation with the finite-difference method (FDM) are
To further demonstrate the tuning effect, two synthetic commonly used to generate synthetic CMP or CIP gathers.
CMP gathers with a class 3 and class 1 AVO anomaly, respec- The former has the advantage of being fast and easy for ana-
tively, were generated from well-log data. The class 3 AVO lyzing primary reflections. Elastic wave-equation modeling
resulted from an unconsolidated sand reservoir that has an calculates particle displacements in the subsurface and
average P-wave velocity of 1900 m/s (Figure 10). The fre- accounts for direct waves, primary and multiple reflection
quency bandwidth used was 8–14 to 90–100 Hz. Due to the waves, converted waves, head waves, and diffractions. It
low impedance of the reservoir, the class 3 AVO anomaly overcomes the shortcomings of ray tracing which can break
can still be observed when the reservoir thickness is reduced down in many cases, such as at edges where the calculated
to 5 m. The amplitudes extracted from the top of the reser- amplitude is infinite or in shadow zones.
voir show a significant tuning effect at a thickness of 1 m. Different AVO modeling methodologies, based on the
The class 1 AVO example is from a tight gas-sand reser- Zoeppritz equations with ray tracing and elastic wave equa-
voir that has an average velocity of about 4600 m/s (Figure tions, have been developed to take into account issues asso-
11). The bandwidth used was 10–15 to 100–120 Hz. When ciated with data acquisition, processing, and interpretation.
the gas sand has a thickness of 18 m, the AVO response has The most commonly used methods are single-interface mod-
a peak at near offset that dims with offset and reverses eling, CMP gather modeling, 2D stratigraphic modeling,

302 THE LEADING EDGE MARCH 2007

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 12. Effects of random noise (S/N = 1:0.2) on AVO extraction for P- and S-reflectivities. The noise has little effect on P-reflectivity, but the effects
are apparent on the S-reflectivity at the reservoir level.

and 2D full wave elastic modeling. 3D elastic wave-equa-

tion modeling needs significant computer power and cur-
rently can only be performed by some major oil companies
and a few research institutions.
Single interface modeling. Single-interface modeling is free
from tuning. It is often used to show theoretical AVO
responses. To demonstrate this method, two examples were
generated. Figure 13a shows the class 3 AVO responses from
an interface of shale overlying porous gas sand, where two-
and three-term Shuey’s equations are compared to evalu-
ate the error introduced by the truncation of the third term.
Also, the maximum and minimum amplitude locations and
inflection point location of the three-term equation were cal- Figure 13. Examples of single-interface AVO modeling: (a) isotropic case
culated using the first and second derivatives. It can be seen using Shuey’s two-term and three-term equations, where blue and green
that the error due to the truncation of the third term increases dashed lines are the first and second derivatives of Shuey’s three-term
at large incident angles. A second example shows the ampli- equation; and (b) horizontal transversal isotropic case, where colors repre-
tude variation with both offset and azimuth for an HTI sent the amplitude that increases in the direction perpendicular to the
medium or for the rock with vertical fractures (Figure 13b). fractures and decreases in the direction parallel to the fractures.
In the fracture plane, the rock is considered isotropic and
the AVO response is the same as that from isotropic medium. eling are often conducted at the same time. Since the modes
In the plane perpendicular to the fracture plane, the AVO of an elastic wave cannot be isolated from each other in elas-
has a different response. By fitting an amplitude surface to tic wave-equation modeling, one may use Kennett’s method
the data using Ruger’s equation (1996), the orientation of (1979) to calculate a single mode of elastic waves such as
the fractures can be solved, and “fracture density” can be multiples or converted waves. Other approaches to con-
calculated by using the gradients parallel and perpendicu- duct wave-equation modeling can be found in Carcione et
lar to the fractures. al. (2002).
Single-gather modeling. For single CMP gather modeling, Figure 14 shows the CMP gathers generated by the
the Zoeppritz equations with ray tracing and/or the full Zoeppritz equation with ray tracing and elastic wave-equa-
wave elastic equation with finite-difference method are often tion modeling with finite difference method. It can be seen
used. While Zoeppritz equation modeling calculates primary that the Zoeppritz modeling produces clean primary reflec-
reflections, elastic modeling accounts for transmission losses, tions (Figure 14a), and the elastic wave-equation modeling
multiples, and converted waves. Therefore, the Zoeppritz generates more waveforms including interbed multiples
equation with ray tracing and elastic wave equation mod- and converted-wave reflections (Figure 14b). The elastic

304 THE LEADING EDGE MARCH 2007

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 14. Single CMP gather AVO modeling for a well in the Western Figure 16. 2D stratigraphic AVO modeling for a play with several low-
Canadian Sedimentary Basin: (a) Zoeppritz equation with ray tracing, impedance gas zones in the Gulf of Mexico: (a) P-reflectivity brightens at
and (b) elastic wave-equation modeling with finite-difference method. The the gas well location but does not provide definite information for defining
former generates primary-only reflections, and the latter generates pri- the reservoirs; and (b) fluid-factor stack shows anomalies with a trough
mary reflections, multiples, and other modes of reflections such as con- and a peak corresponding to the top and base of each reservoir. The overly-
verted waves. The elastic modeling shows the distortions in some of the ing logs P-impedance on P-reflectivity, and VP/VS ratio on fluid-factor
primary reflections. The interbedded multiples may manifest as real reflec- stack.
tions.

Figure 17. Inverted elastic rock property sections from a 2D stratigraphic

AVO modeling for a gas-charged reservoir: (a) P-impedance, (b) S-imped-
ance, (c) λρ, and (d) λρ−µρ. P- and S-impedances do not provide defin-
ing information on the reservoir. λρ and λρ−µρ, however, show a clear
Figure 15. P-wave and converted-wave AVO modeling using Zoeppritz anomaly.
equations with ray tracing; the two-way traveltime of the converted-wave
gather was corrected to P-wave time. The attributes shown are the full- the frequency bandwidths of 10–14 to 110–120 Hz and 10–14
offset stacks of the P-P and P-S gathers, inverted P- and S-reflectivities to 55–60 Hz were used in the P-P and P-S AVO modeling,
(Rppp and Rpps) from the P-P gather, and inverted S-reflectivity and
density reflectivity (Rpss and Rpsρ from the P-S gather.
respectively.
Two-dimensional stratigraphic modeling. 2D stratigraphic
modeling takes into account lateral variations which can be
modeling may be used to test strategies in noise attenua- due to structure, reservoir thickness, porosity, lithology,
tion. In practice, acoustic wave-equation modeling may be fluid type, and fluid saturation. The velocity and density
incorporated with elastic wave-equation modeling to iden- models may come from seismic interpretation or constructed
tify converted-wave energy. using well logs as control points. Using the velocity and den-
Finally, we present an example of converted-wave AVO sity models as input, CMP gathers of a 2D line can be gen-
modeling (Figure 15) which is useful in analyzing multi- erated and then processed.
component data. In practice, three-component data are sep- Figure 16 shows 2D stratigraphic modeling based on the
arated to P-wave (P-P) data and converted-wave (P-S) data wells from a play in the Gulf of Mexico, where the gas zones
for AVO processing. Often, converted-wave data have lower have low P- and S-impedances. It can be seen that P-reflec-
frequency bandwidth. The P-P data can be incorporated tivity section does not provide definitive information on the
with P-S data in AVO attribute inversion. The CMP gathers gas-saturated reservoirs, but the fluid factor does.
in Figure 15 were generated using P-P and P-S Zoeppritz A second example is from a WCSB gas play with well
equations with ray tracing, where the traveltime of the con- control at four locations (Figure 17). The target zone varies
verted-wave gather was corrected to P-wave traveltime. from silt sand to gas sand, and then to brine-saturated sand.
They were processed, and P- and S-reflectivities were In this study, we inverted P- and S-reflectivities from the
inverted from P-P gather using Fatti’s equation, and S-wave synthetic CMP gathers, and then performed P- and S-imped-
and density reflectivities were inverted from P-S gather ance inversion and calculated λρ and µρ. Different sensitivity
using a converted-wave equation (Table 2). In this example, of the inverted elastic rock properties in response to the gas

MARCH 2007 THE LEADING EDGE 305

 Figure 18. Elastic wave-
equation modeling with
finite-difference method
for a structurally com-
plex play in Mackenzie
Delta, Canada: (a) input
P-wave velocity model,
(b) a shot gather, (c)
structure stack, and (d)
prestack depth-migrated
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

section. The location of

the shot gather is indi-
cated in (c) and (d).

Figure 19. Common

image gathers and fluid-
factor section corre-
sponding to Figure 18:
(a) common image
gather in offset domain,
(b) common image
gather in incident angle
domain, and (c) fluid-
factor section. Fluid-
factor section was
calculated based on the
P- and S-reflectivities
extracted from the com-
mon image gathers.

reservoir is evident: The P- and S-impedances are unable to complex area in the Mackenzie Delta, Canada. The purpose
delineate the reservoir, but the λρ and λρ−µρ attributes was to investigate whether the reservoir can be character-
show a clear gas anomaly. ized by AVO attributes extracted from prestack depth migra-
Two-dimensional elastic wave-equation modeling. In struc- tion. In this study, the velocity models and density model
turally complex areas, seismic imaging alone can not com- were constructed from well logs. Figure 18a shows that an
pletely describe a reservoir. Therefore, there is great interest unconformity is between 1750 m and 2700 m, and the gas
in using 2D or 3D elastic wave-equation modeling to gen- reservoir is under the unconformity and sealed by a fault.
erate synthetic data for AVO analysis. In comparison to ray- Figures 18b–d display a shot gather, the structure stack, and
tracing methods, elastic wave-equation modeling generates the prestack depth-migrated section. The results from the
more realistic synthetic data. Recent advances in comput- AVO analysis are shown in Figure 19. Figures 19a and 19b
ing power make single-shot gather modeling trivial and 2D illustrate the conversion of a common image gather from
modeling practical. AVO domain to AVA domain. Figure 19c shows the fluid-
Figure 18 shows the application of 2D full wave-elastic factor section based on P- and S-reflectivity extracted from
modeling with finite-difference method to a structurally the common image point gathers, in which the reservoir was

306 THE LEADING EDGE MARCH 2007

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 20. Seismic CMP gather ties to a synthetic CMP gather: (a) com- Figure 21. Effect of frequency bandwidth in AVO analysis: (a) synthetic
parison of CMP gathers, and (b) comparison of class 1 AVO responses gather with the frequency bandwidth of the 2D data ties to a CMP gather
from the top of the reservoir. The overlying logs are P-impedance (yellow) from the 3D data, and (b) synthetic gather with the frequency bandwidth
and gamma ray (black). of the 3D data ties to a CMP gather from the 3D data. The overlying logs
are P-impedance (red) and gamma ray (black).

Figure 22. Comparison

of inverted elastic rock
properties and well-log
data in a carbonate
play: (a) and (b)
inverted elastic rock
properties from seismic,
where the gas-charged
dolomite porosity zone
is highlighted in black
squares; and (c) and (d)
elastic rock properties
from well logs, where a
gas-charged dolomite
porosity zone is high-
lighted in red squares,
and a wet zone is high-
lighted in green squares.
The overlain empirical
lines are for wet sand
(blue), shale (black),
clean gas sand (red),
limestone (black dashed
line), and dolomite
(green dashed line). The
inverted elastic rock
properties in (a) and (b)
indicate that the reser-
voir is gas-charged with
porosity similar to the
porous dolomite high-
lighted in green squares
in (c) and (d).

clearly defined. This example demonstrates that, for a struc- AVO responses in CMP gathers or by analyzing inverted
ture play, an a-priori study using elastic wave-equation AVO attributes. An understanding of local petrophysics,
modeling and prestack depth migration helps determine rock physics, and geology provides constraints for QC in
whether AVO analysis is feasible. AVO processing and interpretation. Ideally, AVO process-
ing and interpretation should be conducted in parallel with
AVO modeling applications. AVO modeling contributes sig- AVO modeling. As a result, uncertainty and risk can be
nificantly to prestack data processing and interpretation. reduced.
Inappropriate processing may be detected by examining Data processing. Calibration on CMP gathers and AVO
MARCH 2007 THE LEADING EDGE 307
 Figure 23. Elastic wave-
equation modeling using
finite-difference method
for a dolomite reservoir in
the Wabamun Formation
in the Western Canadian
Sedimentary Basin: (a)
with reservoir, (b) with-
out reservoir, and (c) the
difference between (a) and
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

(b). The overlying logs

are gamma ray (black), P-
impedance (red), and S-
impedance (yellow). In
this play, multiple conta-
mination at reservoir
level is a crucial issue in
determining whether a
bright spot in amplitude
indicates porosity or the
result of an interbedded
multiple. The elastic
modeling shows signifi-
cant difference in ampli-
tudes for the cases with
and without the reservoir.

attributes by AVO modeling is useful in optimizing AVO pro- Calibration can be applied to other AVO attributes such
cessing. It is often implemented by using well logs, synthetic as S-reflectivity, gradient, and fluid factor. It can also be applied
gathers, walkway VSP data, and known relationships between by plotting P-reflectivity against S-reflectivity, and P-imped-
AVO attributes or between seismic rock properties. Calibration ance against S-impedance. Figure 22 shows an example using
can be performed at a CMP location, or globally, on a data set well logs to calibrate inverted elastic rock properties in a gas-
to answer such questions as these: charged dolomite play. In Figures 22a–b, the inverted elastic
rock properties from the zone of interest are highlighted in
• Have the CMP gathers been properly processed with an black squares. The data points shift toward low λρ values and
amplitude-preserving processing workflow? low λ/µ ratios, indicating effects of gas. To calibrate this, dipole
• Are phase, tuning, signal-to-noise ratio, and frequency sonic logs from the same area were plotted (Figures 22c–d).
bandwidth influencing the AVO solutions? In this well, the data from the gas-charged dolomite (red
squares) have lower porosity and the data from a brine-satu-
Tying synthetic CMP gather(s) to recorded seismic data rated porous dolomite (green squares) have higher porosity.
often gives a quick insight into data quality and to type of The comparison between seismic data and well-log data con-
AVO response. One may perturb the well logs to represent firms that the reservoir is gas-charged because it has low val-
possible reservoir conditions to examine the variation of seis- ues of λρ and λ/µ. Also, it indicates that the reservoir has
mic responses. For example, gas substitution may be per- porosity similar to the data points highlighted by the green
formed on a wet well to examine the gas effects. Other squares in Figures 22c–d.
parameters often perturbed in AVO modeling are porosity, Global calibration using AVO modeling is an approach to
reservoir thickness, lithology, and frequency and angle to be conduct QC for a prestack seismic data set. For example, AVO
used for AVO attribute extraction. within a time window or from a specific reflection can be com-
Figure 20a shows an example in which a synthetic CMP pared and calibrated with the results from AVO modeling. This
gather is tied to a recorded CMP gather from a 2D survey. By type of exercise can tell whether the data were processed by
comparing the AVO responses from the top of the gas reser- an amplitude-preserving workflow.
voir, a class 1 AVO anomaly with polarity reversal at the far Phase analysis is an important aspect in AVO processing
offsets is confirmed. This is further validated quantitatively QC. Non-zero-phase data result in inaccurate AVO attributes
by the amplitudes extracted from both the seismic and the syn- and inverted elastic rock properties. Phase analysis is usually
thetic gathers (Figure 20b). In this study, 3D data were initially accomplished by crosscorrelation using zero-offset synthetics
considered for AVO processing. Further study, however, and migrated stack. It produces wavelets accompanied by
revealed that the 3D data have a frequency content below the phase information. Well ties using CMP gathers may remove
resolution for delineating the reservoir. Figure 21a shows that some ambiguity because of additional information from AVO
a synthetic CMP gather generated using the bandwidth from responses. For example, a typical AVO response or an AVO
the 2D data ties to the 3D data. It is evident that the frequency anomaly can be criteria to tie well to seismic. Also, improved
content in the 3D data is significantly lower than that in the wavelet(s) may be obtained by using extracted P-wave reflec-
2D data. In Figure 21b, a good well-tie was reached when the tivity (RP) instead of migrated stack because the former are
frequency bandwidth of the 3D data was used in the synthetic considered zero-offset data. Furthermore, AVO attributes such
CMP gather. as fluid factor may be useful in phase analysis. For example,

MARCH 2007 THE LEADING EDGE 309

Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 24. Interpretation of a gas-charged dolomite reservoir with an average porosity of 14% in the Western Canadian Sedimentary Basin: (a) syn-
thetic CMP gather, and (b) stack section and CMP gathers at three wells. The gas-charged dolomite porosity produces a class 3 AVO anomaly (peak) at
the base of the reservoir. This was confirmed by AVO modeling.

in a clastic environment, a peak followed by a tough in a for field data. The synthetic gathers, AVO attributes and
fluid-factor section may indicate that the seismic data have inverted elastic rock properties can thus be directly compared
reversed polarity. to their counterparts from seismic data.
Data interpretation. AVO modeling has an important role As a special case study, Figure 23 shows an example that
in assisting interpretation on CMP gathers, AVO attributes, uses elastic wave-equation modeling to investigate the inter-
and inverted elastic rock properties. It helps validate AVO ference between primary reflections, multiples, and converted
responses and links seismic expression to known reservoir con- energy at the Wabamun dolomite porosity in the WCSB. This
ditions. It is able to integrate petrophysics, rock physics, and is a classic case in which multiples were often misinterpreted
geology with seismic data. As a result, confidence in inter- as porosity on migrated stack. We were interested in know-
pretation is increased, and the risk in drilling is reduced. ing how multiples and converted energy interfere with the
Tuning effects may invoke or mask an AVO anomaly as primary reflections at the reservoir level. AVO modeling was
described earlier in this paper. AVO modeling using varied conducted with and without porosity. Figure 23 shows that
frequency bandwidth or reservoir thickness may yield the there is a significant difference in seismic responses between
answers. Complex lithologies may produce AVO anomalies these two cases. Further, the information from this study
that can be investigated by AVO modeling as well. For exam- helped optimize noise attenuation and amplitude recovery.
ple, AVO responses from tight sand may manifest as class 1 The final example is from a study on a carbonate reser-
AVO anomalies and show brightening of amplitude in a voir in the WCSB (Figure 24). AVO modeling shows that the
migrated stack. This type of AVO anomaly may be excluded gas-charged dolomite reservoir that has an average porosity
by high P- and/or S- impedances. Coal, carbonate, and the of about 14% produces a class 3 AVO anomaly. This agrees
lithologies that do not follow the mudrock trend may produce with the AVO response observed in the CMP gather at the pro-
fluid-factor anomalies. It is possible to exclude these anom- ducing well. In contrast, a completely different seismic
alies through integrating geologic information. Further, high response was observed at the tight well locations. In this case,
clay content in rock results in low gas saturation, and this type the information obtained from AVO modeling validated the
of partial gas saturation may be distinguishable because the seismic responses.
increase in clay content results in an increased VP/VS ratio, and
consequently changes the AVO responses. Conclusions. We have demonstrated the importance and
In data interpretation, synthetic CMP or CIP gathers should practical aspects of AVO modeling in prestack seismic pro-
be processed through the same data processing sequence used cessing and interpretation. We have also presented and dis-

310 THE LEADING EDGE MARCH 2007

 cussed the AVO modeling methodologies—single-interface damental elastic constants” by Gray et al. (SEG 1999 Expanded
modeling, single-gather modeling, 2D stratigraphic model- Abstracts). “Shear-wave velocity estimation in porous rocks:
ing, and 2D full wave elastic modeling. Except for the AVO Theoretical formulation, preliminary verification and applica-
modeling using Zoeppritz equation with ray tracing, full tions” by Greenberg and Castagna (Geophysical Prospecting, 1992).
wave elastic-wave equation modeling helps AVO analysis Effect of Porosity and Clay Content on Acoustic Properties of
in structurally complex areas, and converted-wave AVO Sandstones and Unconsolidated Sediments by Han (PhD dis-
modeling provides useful information on analyzing con- sertation, Stanford University, 1986). “Lithology color-coded
verted-wave data. We have demonstrated that it is essential seismic sections: The calibration of AVO cross-plotting to rock
to examine seismic rock properties and their contrasts in order properties” by Hilterman and Verm (TLE, 1995). “Shear-wave
Downloaded 06/06/13 to 139.133.11.3. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

to understand the sensitivity of seismic rock properties velocity and density estimation from PS-wave AVO analysis:
responding to fluid. The information from analyzing petro- Application to an OBS dataset from the North Sea” by Jin et al.
physical and seismic rock properties is useful in predicting (GEOPHYSICS, 2000). “Theoretical reflection seismograms for elas-
AVO responses. In AVO modeling, effects of reservoir thick- tic media by Kennett” (Geophysical Prospecting, 1979). AVO
ness, sensitivity of rock properties responding to fluid and Inversion by Simultaneous PP and PS Inversion by Larson
lithology are important and need to be investigated. Special (Master’s thesis, University of Calgary, 1999). “Application of
attention needs to be paid to seismic resolution as it alters amplitude versus offset in carbonate reservoirs: Re-examining
AVO responses. Offset-dependent tuning, noise contamina- the potential” by Li and Downton (SEG 2000 Expanded Abstracts).
tion, NMO stretch, limitations in seismic data acquisition, “An empirical method for estimation of anisotropic parameters
and structural effects need to be studied. Furthermore, AVO in clastic rocks” by Li (TLE, 2006). “Amplitude and AVO
modeling is an exercise in multidisciplinary integration of responses of a single thin bed” by Liu and Schmitt (GEOPHYSICS,
petrophysics, rock physics, seismic, geology, and petroleum 2003). “Critical porosity: The key to relating physical properties
engineering. It provides important information on reservoir to porosity in rocks” by Nur et al. (SEG 1995 Expanded Abstracts).
characterization and risk reduction in hydrocarbon explo- Reflection Coefficients and Azimuthal AVO Analysis in
ration. Finally, AVO modeling of an anisotropic medium, Anisotropic Media by Ruger (PhD dissertation, Colorado School
especially an HTI medium, faces challenges due to the dif- of Mines, 1996). “Amplitude-versus-offset variations in gas
ficulties in obtaining accurate anisotropic input. sands” by Rutherford and Williams (GEOPHYSICS, 1989). “A sim-
plification of the Zoeppritz equations” by Shuey (GEOPHYSICS,
Suggested reading. Quantitative Seismology by Aki and Richards 1985). “Weighted stacking for rock property estimation and
(Freeman, 1980). “Thin-bed AVO effects” by Bakke and Ursin detection of gas” by Smith and Gidlow (Geophysical Prospecting,
(Geophysical Prospecting, 1998). “Relationships between com- 1987). “Removal of offset-dependent tuning in AVO analysis”
pressional-wave and shear-wave velocities in clastic silicate by Swan (SEG 1997 Expanded Abstracts). “Velocity from ampli-
rocks” by Castagna et al. (GEOPHYSICS, 1985). “Framework for tude variations with offset” by Swan (GEOPHYSICS, 2001). “Weak
AVO gradient and intercept interpretation” by Castagna et al. elastic anisotropy” by Thomsen (GEOPHYSICS, 1986). “Weak
(GEOPHYSICS, 1998). “Seismic modeling” by Carcione et al. anisotropic reflections” by Thomsen (in Offset Dependent
(GEOPHYSICS, 2002). “Thin bed AVO: The effect of NMO and NMO Reflectivity—Theory and Practice of AVO Analysis, Investigations
correction on reflectivity sequences” by Castoro et al. (SEG 1998 in Geophysics Vol. 8, SEG, 1993). “Making AVO sections more
Expanded Abstracts). “Amplitude responses of thin beds: robust” by Walden (Geophysical Prospecting, 1991). “Elastic wave
Sinusoidal approximation versus Ricker approximation” by velocities in heterogeneous and porous media” by Wyllie et al.
Chung and Lawton (GEOPHYSICS, 1995). “AVO detectability (GEOPHYSICS, 1956). “A physical model for shear-wave velocity
against tuning and stretching artifacts” by Dong (GEOPHYSICS, prediction” by Xu and White (Geophysical Prospecting, 1996). TLE
1999). “High-resolution AVO analysis before NMO” by Downton
and Lines (SEG 2003 Expanded Abstracts). “Detection of gas in Acknowledgments: Jonathan Downton and Yong Xu were employed by
sandstone reservoirs using AVO analysis: A 3D seismic case his- Paradigm when this study was done. The authors thank Larry Lines for
tory using the Geostack technique” by Fatti et al. (GEOPHYSICS, reviewing that improved the quality of this paper. They also acknowledge
1994). “A velocity function, including lithologic variation” by Keith Young, Michael West, Jan Dewar, Paul Hewitt, Bob Somerville,
Faust (GEOPHYSICS, 1953). “Improving AVO attributes with robust Huimin Guan, and Luiz Loures for enlightening discussion and assis-
regression and quality control” by Ferré et al. (SEG 1999 Expanded tance.
Abstracts). “Formation velocity and density—The diagnostic
basis for stratigraphic traps” by Gardner et al. (GEOPHYSICS, Corresponding author: yli@paradigmgeo.com
1974). “Bridging the gap: Using AVO to detect changes in fun-

MARCH 2007 THE LEADING EDGE 311