**Abstract: Regional Velocity Reduction**

David Judson, Bruce Meadours, John Sherwood, Johann Sattlegger

In the conventional processing of exploration-seismic data from an offshore regional survey, many estimates are made of * velocities* which will optimize the stacking of common-depth-point data. Provided the geologic structure is not overly complex it is reasonable to utilize this stacking-velocity information to derive gross-

*velocity estimates. These facilitate the approximate conversion of seismic times to depth and can be utilized further in the interpretation of gross lithologic changes.*

**interval**A recently developed statistical method allows the extraction of gross-**interval*** velocities*, given stacking-velocity information for the region and an approximate time-space model for each geologic interface. A least-square error fit is made to the stacking-velocity observations by suitably adjusting the unknown

**interval***in the following manner.*

**velocities**The velocity within each geologic unit is allowed to change laterally in a continuous manner by representing its squared value with a piecewise continuous bicubic polynomial with unknown coefficients. Such a velocity model results in a large set of linear equations relating the unknown coefficients to each original stacking-velocity observation. Because of the possibly large number of equations and coefficients it is preferable to use an iterative approach to derive the least-square error solution. This final solution provides a compact representation of the lateral * interval*-velocity variations for the region. Its form is particularly convenient for seismic time-to-depth conversion and the production of maps.

AAPG Search and Discovery Article #90976©1976 AAPG-SEPM-SEG Pacific Section Meeting, San Francisco, California