Hybrid tomography based on beam migration

We propose a method of tomography that utilizes the time residual and other attributes associated with each wavelet in beam migration. This time residual is referred to as 3D residual normal moveout (3DRNMO) and is related to error in the velocity model. The spatial dip components of the wavelet with respect to time enable very efficient ray tracing. The proposed method possesses the features of both postmigration tomography and stereotomography. Furthermore, not only reflections but also diffractions and turning waves are included in the tomography. The method is validated using a TTI synthetic data set from a 3D syncline model. The method is also applied to fine tune the velocity model for a marine field data set.