John Stockwell

Colorado School of Mines

This is part of a 400 level course on seismic data processing
Designed for a geophysics course

By the time that the students arrive at this problem, they have become familiar with using the CWP/SU:Seismic Un*x software package for viewing and filtering data, and they have done some constant background migrations.

This is part of the weekly exercises. The topic is revisited
when the students do their final project of doing a processing sequence on the same dataset.

Velocity analysis is a necessary part of seismic processing because an accurate background velocity profile is needed for performing a meaningful migration of seismic data.

Velocity analysis requires that the students understand the concepts of RMS (stacking) and interval velocities. Velocity analysis is done with and without noise suppression, so that the real-world difficulties are brought home to the student.

The software environment is not the point and click environment that most of the students have grown up with. In general, the CWP/SU package is an "expert user" environment that is commandline-shell script oriented package run under Linux. Students are free to interact with each other and are encouraged to help each other with the assignment, but ultimately, each student's work is their own.

One-on-one instruction is a must for this type of exercise.

The students each perform a velocity analysis over on a portion of the Viking Graben data that has been presorted into CMP gathers using a shell script called Velan that is written in Unix (bourne or bash) shell language. This procedure involves the hand picking of velocity as a funciton of traveltime on the section (RMS velocities).

The students are provided with additional shell scripts that take the results of the velocity analysis and generate interval velocities as a function of depth and time.

The students are expected to consider how consistent their velocity analysis is, as well as consider the geological reasonableness of the results.

Each student does a normal moveout correction followed by a stack of the seismic data, and maybe a post-stack migration. Typically there will be pathologies in the students' velocity picks that will introduce pathologies. Evaluating those pathologies present opportunities for demonstrating to the students the importance of velocity models, as well as showing the students how to use the pathologies to their advantage.
Has minimal/no quantitative component
Addresses student misconceptions

This is a weekly assignment which is emailed to the instructor with a short (one sentence discussion). The next class period, we compare the students' results with an eye on showing what can be learned from the errors and variations in each of the students' work. Because there is an element of interpretation, there is no one "right" answer, but rather rich diversity of answers.

Seismic Data Processing by Oz Yilmaz is reference that students may read. They get to see the messy reality of what Dr. Yilmaz's slick textbook figures.