Prestack migration using arrival angle information

Abstract

When prestack depth migration is applied to field data, migration artifacts or false images resulting from limited observation geometry or aperture often make it difficult to interpret the migrated images. The occurrence of artifacts is especially a problem on VSP data that usually have poor geometry for migration. To reduce such unwanted artifacts in migration images, I propose here that information about incident angles or polarization directions of scattered waves at a receiving point obtained from multicomponent data be incorporated into the extrapolation of wavefields in migration. In this method, a given arrival of a wave from a subsurface scatterer is extrapolated in the direction focused on its incident direction at a receiving point to improve the migration image of the scatterer. This focusing is achieved by a projection of observed multicomponent scattered data onto the expected polarization direction at a receiving point for a subsurface scatterer. Numerical examples using simple 2-D and 3-D homogeneous models show that the multicomponent migration presented here can largely decrease migration artifacts or false images. Numerical experiments with more realistic models reveal that multicomponent migration can provide interpretable migration images even for very limited observation geometry such as a single source offset VSP, where single‐component migration fails. Additional tests indicate that this multicomponent migration method can be useful even in case where the observed data contain noise and a background velocity model deviates from the true velocity structure.