--> Wave-Equation Migration in Mountainous Areas, by Jiao, Jianwu, Barry Newman, Stewart Trickett, Brian Link; #90030(2004)
[First Hit]

Datapages, Inc.Print this page

Previous HitWaveNext Hit-Previous HitEquationNext Hit Migration in Mountainous Areas

Jiao, Jianwu1, Barry Newman2, Stewart Trickett1, Brian Link1 
1 Kelman Technologies Inc, Calgary, AB
2 Kelman Technologies Inc, Denver, CO

Prestack shot-domain Previous HitwaveNext Hit-Previous HitequationNext Hit depth migration can produce better quality images than Kirchhoff methods. Most Previous HitwaveNext Hit-Previous HitequationNext Hit migrations based on downward continuation, however, are restricted to seismic data which is regularly and densely sampled in space, datumed to a planar surface, and has no significant anisotropic effects. This prevents application of Previous HitwaveNext Hit-Previous HitequationNext Hit migration to most land data, and in particular to areas with rugged topography. Here we describe how to overcome these limitations.

First, we regularize the traces beforehand using polynomial interpolation. Provided there is no spatial aliasing, polynomial interpolation has a number of attributes which make it well suited for the application at hand. Second, we migrate the seismic data directly from the acquisition surface, eliminating any datuming or elevation static corrections before migration. Finally, we add anisotropy and dip parameters to the depth-imaging velocity model, and modify the explicit finite-difference scheme to handle an anisotropic media with nonvertical symmetric axes.

The effectiveness of the resulting migration is demonstrated by comparing it to Kirchhoff migration on a synthetic data example simulating conditions in the Canadian Rockies, and on actual 2-D and 3-D data examples from the fold and thrust belt area of the Alberta foothills. These examples show that Previous HitwaveNext Hit-Previous HitequationTop migration can result in superior results over Kirchhoff migration. One remaining problem is with 3-D volumes, where the large interval between receiver lines can make regularization difficult. This requires further study.