In this microteaching session, I would like to present a paper written by Brian H.

Russel with
the title of Prestack seismic amplitude analysis: An integrated overview

SLIDE 1
➢ This paper discuss about an overview of the techniques that are in use for prestack
seismic amplitude analysis
➢ These techniques can be classified as being based on the computation and analysis of
either some type of seismic reflection coefficient series or seismic impedance
➢ Those techniques that are based on the seismic reflection coefficient series, or
seismic reflectivity for short, are called amplitude variation with offset methods, and
those that are based on the seismic impedance are referred to as prestack amplitude
inversion methods.
➢ To understand the distinction between impedance and reflectivity we can see the
different ways in which geologists and geophysicists look at their data
➢ The figure below shows the actual earth’s layer which geoscientist try to characterize.
Geologist generally analyzes borehole well log measurements of some layer
parameter P, whereas the geophysicist, when using exploration seismology, analyzes
changes in this layer parameter at the interfaces between successive layers, called the
reflectivity (R).

SLIDE 2
➢ There is a relationship between this P and R, written as
where subscript i refers to the ith layer.
➢ The reflectivity R of parameter P of the ith layer is found by dividing the change in the
parameter P by twice its average value.

SLIDE 3
➢ So we get a relationship between refelectivity R and parameter P.
➢ And the problem is even if our processing is extremely good, we are always left with a
band-limited reflectivity which means it strongly depends on the effect of the seismic
source band
➢ It can be modeled using the convolutional model, shown here as matrix multiplication
between
R is an m-sample reflectivity,
W is the convolutional matrix in which each column contains an n-sample wavelet
shifted by one sample from the previous column,
and S is the n+m−1 sample result, or seismic trace
 ➢ What is actually convolution? We can think of convolution as shifting the wavelet so
that it is centered at each reflection coefficient, scaling the wavelet by the amplitude
of each reflection coefficient, and then summing the result.

➢ Next questions what is parameter P we are interested in? We are limited to

parameters that are related to the propagation and reflection of the seismic signal,
such as

SLIDE 4
➢ Before we talk about real seismic data used in this study, it worth to review the
concept of CMP seismic gather. In this case traces recorded at different offsets are
grouped by increasing offset around a CMP point.
➢ Notice that each offset Xi corresponds to an angle of incidence θi at the reflector.
➢ Figure below shows the single reflector which will create a seismic reflection due to
the contrast between the density (ρ), P-wave velocity (V P), and S-wave velocity (VS ) in
each layer.

SLIDE 5
➢ Figure on the left shows a set of real gathers which were acquired and processed over
a shallow gas field in central Alberta. This data set will be used to explain many of the
methods described in this tutorial. In particular, notice the amplitudes change as a
function of offset on the trough and peak of the seismic reflection event below 620
ms.
➢ Our goal in this tutorial is a physical understanding of this amplitude change, and how
it relates to the gas sand.

SLIDE 6
Reflectivity methods
Normal incidence reflectivity
➢ In an elastic earth, we observe mode converted reflected and transmitted P- and S-
raypaths for an incident P-wave raypath at arbitrary incidence angle
➢ Figure shows the conversion of an incident P-wave at a non-zero angle of incidence
into its reflected and transmitted P- and S-components
➢ As we already know S-wave travel slower than the velocity of the P-waves

➢ Zoeppritz in 1919 has derived equations to calculate the amplitudes of the reflected
and transmitted waves shown in Figure
➢ The complete solution for an arbitrary angle leads to a complicated set of equations
 ➢ The normal incidence P-wave reflection coefficient can be determined in a
straightforward way from the Zoeppritz equations.

SLIDE 7
➢ In order to understand the P- and S-impedance, reflectivity we discussed before with
our example data case we also provide here data from log measurements
➢ Also shown are logs that have been derived from the velocity and density logs such as
VP∕ VS ratio, Poisson’s ratio, extended elastic impedance (EEI) at 56°, λρ, and μρ.

SLIDE 8
➢ In reflectivity-based method we use Aki-Richards equations to approximate the P-
wave reflection coefficient by the linearized sum of three terms

➢ Note that wavelet is angle-dependent, losing its higher frequencies as the angle
increases.
➢ Equation above is also angle-dependent approximation which starts to deviate from
the full Zoeppritz solution between 30° and 35°.

➢ We can rearrange the equation, as shown by Fatti et. al (1994) and also by Wiggins et
al (1983) and end up with the other three terms, consist of the gradient, intercept and
curvature
➢ Eq 12 serve as the basis of analysis of gradient and intercept as commonly used for
AVO analysis in the industry

➢ We also can find gradient and intercept by crossplotting the picked amplitudes of an
angle gather at a given time as a function of sin 2θ , fit a straight line to the points, the
intercept of the plot gives RAI and the slope, or gradient, gives RGI.

➢ We can extract reflectivities by considering the Rpp or Reflection coefficient as the

picked amplitude at time t and angle θi ,
➢ Once these reflectivity terms have been estimated from the prestack data at each
gather and time sample, they can be used as the basis for the impedance inversion
methods
➢ By using this way, we can estimate the intercept (RAI) and gradient (RGI)

SLIDE 9
➢ The red polygonal zone captures outliers that have negative intercept and gradient
values, and thus correspond to the top of the gas sand;
 ➢ the blue polygon captures outliers that have positive intercept and gradient values,
and thus correspond to the base of the gas sand.
➢ We can see this polygonal zones on the seismic section

SLIDE 11
➢ By this equation would recover the complete impedance profile of the earth layer
➢ However, given the band limitation of the actual seismic trace, we only recover the
impedance within the range of the seismic bandwidth
➢ To produce a full-bandwith estimate, we incorporate some external model, well log
impedance profiles, of course, at their correct position on a seismic volume and do
interpolation of the impedances

➢ We’ve consider the model, in order to perform inversion we need seismic volume as
an input. Traditionally, we use stacked seismic volume to perform acoustic impedance
inversion, but this is actually incorrect, because the stack is an average over all angles,
whereas the acoustic impedance reflectivity is the normal incidence P-wave
reflectivity.
➢ Thus, a more accurate input to acoustic impedance inversion would be the band-
limited acoustic impedance reflectivity (SAI = W RAI)

SLIDE 12
➢ Once the impedance inversion has been done, there are several approaches that can
be used to interpret the results.
➢ A straightforward method is to divide the P-impedance volume by the S-impedance
volume to create a VP ∕ VS ratio volume and to crossplot this against P-impedance
using the rock physics template (RPT) approach (Odegaard and Avseth, 2003)
➢ Anomalous fluid zones can then be picked on the crossplot and highlighted on the
seismic volume.
➢ Once the gathers have been inverted to P- and S-impedance, the results can be
transformed to lambda-rho (λρ) and mu-rho (μρ) volumes and Poisson (Goodway et
al., 1997)

SLIDE 13
➢ Figure on the right shows the region of the seismic stack that corresponds to the
polygonal zone in Figure on the left. This region correlates very well with the known
gas sand

SLIDE 14
➢ So far, we’ve talked several different types of impedances: acoustic, shear, Poisson,
lambda-rho, and mu-rho, which are derived from acoustic and shear impedance.
 There is another type of impedance underlying the angle dependent reflectivity, that
is EI
➢ To create the inverted EI volume, we use the methodology suggested before, where a
constant angle stack is the seismic input, the model volume is derived from equation
34, and your preferred inversion algorithm is used.
➢ The modified version of EI is EEI which is based on χ rotation angle in intercept-
gradient space instead of angle θ.
➢ Figure on the top right shows the optimum χ angle that is found by correlation
between EEI (χ) function and a parameter of interest, in this case we got a best
correlation value at angle of 560
➢ Figure on the left shows the computation of the EEI reflectivity at the computed angle
of 56° from the correlation before. This is the seismic section that is inverted to give
the final EEI result
➢ Finally, Figure on the bottom right shows the computation of V P∕VS ratio for our gas
sand example using the EEI log model with a rotation angle of 56°.

SLIDE 15
➢ So far, we only consider the isotropic case. Here we would like to extent our analysis
into anisotropic media. The most common cases to consider are the VTI and HTI
media.
➢ We can think their difference
➢ In order to involve the effect of anisotropy, we need to modified our the three-term
equation before
➢ The effect of anostropy can be seen here from the modeled AVO response where we
consider two cases, gas sand and wet sand.
➢ The results is incorporating anisotropy effect into our formulation will increase the
AVO effect for the gas sand case, while for the wet sand case it makes the wet
response look similar to the gas response