THE MARMOUSI2 MODEL, ELASTIC SYNTHETIC DATA, AND AN ANALYSIS OF
IMAGING AND AVO IN A STRUCTURALLY COMPLEX ENVIRONMENT
A Thesis
Presented to
the Faculty of the Department of Geosciences
University of Houston
In Partial Fulfillment
of the Requirements for the Degree
Master of Science
By
Gary Stuart Martin
May 2004
THE MARMOUSI2 MODEL, ELASTIC SYNTHETIC DATA, AND AN ANALYSIS OF
IMAGING AND AVO IN A STRUCTURALLY COMPLEX ENVIRONMENT
Gary Stuart Martin
APPROVED:
Dr. Kurt Marfurt
Dr. Hua-Wei Zhou
Dr. Fred Hilterman
Dr. Chuan Yin
Dean, College of Natural Sciences and
Mathematics
ii
ACKNOWLEDGEMENTS
There have been many contributors to this project and I wish to thank everyone
involved for their support, encouragement, and technical assistance. My thanks go to
Aline Bourgeois at the Institut Français du Pétrôle who provided an initial dataset from
which to begin construction of Marmousi2. Don Larson at GX Technology Corporation
deserves a special mention. His technical assistance throughout has been invaluable,
and his programming efforts to customize GXII to use the shear wave and density
functions enabled the model to contain the desired complexity. My appreciation also
goes to the management of GX Technology for their support of my efforts, and for the
use of company algorithms and facilities. John Faragher, John Crowther, Susan Collins,
and Mohamed Dolliazal, all from GX Technology also deserve my thanks for their
technical contributions. Dr. Robert Wiley at the University of Houston warrants a special
mention for his work which involved the setup and execution of the elastic modeling
using the Sun cluster. He has also been involved with the storage, duplication, and
distribution of data copies. My appreciation also goes to Cory Hoelting at the University
of Houston for his assistance with the numerical modeling and other various tasks. This
work would not have been possible without the contributions of my committee, and
special thanks are due to Dr. Kurt Marfurt and Dr. Fred Hilterman. Last, but certainly not
least, my deepest gratitude goes to my wife Rebecca and my daughter Kate who have
had to endure many lonely evenings and weekends over the past three years.
iii
THE MARMOUSI2 MODEL, ELASTIC SYNTHETIC DATA, AND AN ANALYSIS OF
IMAGING AND AVO IN A STRUCTURALLY COMPLEX ENVIRONMENT
An Abstract of a Thesis
Presented to
the Faculty of the Department of Geosciences
University of Houston
In Partial Fulfillment
of the Requirements for the Degree
Master of Science
By
Gary Stuart Martin
May 2004
iv
ABSTRACT
The reliability of AVO analysis is well established in relatively simple structural
areas, but there remains some doubt as to the applicability of the technique in areas of
complex velocity structure. To study this problem, I have created a 2D synthetic
numerical model containing realistic hydrocarbons in a variety of structural settings. The
new model, Marmousi2, is based on the structure and velocity of IFP’s acoustic
Marmousi model, but has been extended in width and depth, and is fully elastic.
High frequency, high fidelity elastic modeling was performed using state of the art
modeling code and computational resources. Synthetic streamer, OBC, and VSP multi-
component shot records were collected, including offsets up to 17km. Analysis of the
data indicates that it is suitable for a wide variety of geophysical research including
conventional imaging, AVO analysis, multiple attenuation, multi-component imaging,
inversion, etc. The model and dataset have been made available to other researchers
throughout the world.
Using a marine streamer subset, I applied some basic processing and surface
multiple attenuation. I imaged the data with a suite of imaging algorithms using the
known velocity model. The complex nature of the velocity dictates that for a good overall
solution prestack depth imaging methods are required. The wavefield prestack imaging
method produced the most impressive result.
In complex areas prestack imaging and AVO analysis are inextricably linked
since more simple methods such as NMO and stacking are not sufficient to produce
meaningful data. Events must be well imaged on the migrated stack section before AVO
analysis is possible due to the much lower signal to noise ratio present in the image
v
