Warping prestack imaged data to improve stack quality and resolution

Abstract

Accurate seismic imaging requires that a geologic feature be located at the same lateral and vertical position in images obtained by 3D prestack migration from different data bins, such as common-offset or common-angle subvolumes. Misalignment of those images degrades the quality of the stack. For dipping reflectors and lateral discontinuities, imperfect imaging causes both lateral and vertical misalignment. In current practice, the vertical component of the misalignment is used to estimate updates in velocity and other imaging parameters; the lateral component is largely ignored. We show that recent developments in seismic-attribute analysis allow us to examine the lateral misalignment of prestack volumes with similar resolution to that achieved in examining vertical moveout. To measure lateral moveout, we pick maxima from local 2D crosscorrelations computed between slices from 3D attribute volumes. We then use these measurements to correct for the lateral misalignment by applying a warping procedure to the corresponding slices in the prestack migrated seismic data. We apply our technique to a 3D land survey acquired over the Fort Worth basin in Texas, and obtain subtle, but potentially important, improvements in the quality and resolution of the stack as well as in the attribute images computed from the corrected data.