CONCLUSIONS

The results of the tests presented in this paper indicate that, when the velocity function increases with depth or there are NMO-velocity conflicts, AMO is a more robust process than constant-velocity DMO to transform prestack data before performing partial stacks. AMO performs better than constant-velocity DMO because it is a residual operator, and thus does not overcorrect data that have been moved out with a too fast NMO velocity.

The results of these tests encourage further comparison of AMO with V(z) DMOs, and possibly in data sets collected in areas where transverse anisotropy plays an important role.