Enhanced Cascaded Migration

Migration deals with an inverse problem of geophysics. It collapses by wave field extrapolation methods the surface recorded seismic data into a depth acoustic reflectivity map. The upward compressional waves travel with velocity dependent on local density and pressure parameters of the medium. On the other hand, the Stolt method solves exacty wave migration only with constant velocity, but it is unable to take into account lateral velocity variations. This work shows the possibility of solving this point, merging general Stolt scheme (cascaded migration) with a space-frequency method. A physically accurate and computationally effective parallel algorithm is presented here, and some results on an implementation for a supercomputer are presented.