** Next:** THEORY
** Up:** Introduction
** Previous:** Motivations

The driving force behind the current research effort is to extract as much
information as possible about subsurface physical properties from
seismic reflection data.
The first and largest part of this effort is to
reconstruct the Zoeppritz reflection coefficients, *R*_{pp}, as
a function of specular reflection angle
at each subsurface pixel in the
designated target zone. This is accomplished by the Kirchhoff elastic
prestack depth migration, which is *elastic* in the sense that background
migration model values of *V*_{p}, *V*_{s}, and are specified.
Our method is similar in spirit to that of Parsons (1986), which in turn is
based on the methods of Beylkin (1985) and Bleistein (1987), except we
consider elastodynamic wave propagation instead of acoustics.
Additionally, values of the attenuation quality factor *Q* may be specified
as a function of a single dominant frequency to approximate anelastic
exponential amplitude attenuation (but not dispersion).
It is a true *depth* migration in the sense that heterogeneous
interval model properties, including lateral velocity variations, are
specified as a function of *spatial* coordinates, in direct contrast to
time migration methods.
Given the Zoeppritz specular reflectivity *R*_{pp}, information can be
further extracted regarding the polarity, magnitude, and spatial distribution
of *relative changes* in elastic parameters. Algorithmically, we
achieve this inversion by defining a matrix equation relating *R*_{pp} for
several specular angles to three elastic parameters using the linearized
Bortfeld approximation to the nonlinear Zoeppritz coefficient
(Bortfeld, 1961).
The matrix system is then inverted by Singular Value Decomposition (SVD)
at each subsurface location. (For a review of the SVD method we refer
the reader to Lines and Treitel, 1984.)
This allows the possibility
of producing elastic parameter depth images of the subsurface target
zone, similar in appearance to prestack depth migrations, each
exhibiting spatial variations in P or S impedance, for example.
Such images will be shown in this report for synthetic data, with the
corresponding field data images to be shown at a later date.
Ideally, this capability will
provide new or additional information about subsurface properties which
may be useful in exploration and production applications.

** Next:** THEORY
** Up:** Introduction
** Previous:** Motivations
Stanford Exploration Project

12/18/1997