Dip-angle image gathers computation using Poynting vector in elastic reverse-time migration and their application for noise suppression

Angle-domain common-imaging gathers (ADCIGs) are important in analyzing the subsurface discontinuities where the reflection-waves take place. In elastic reverse-time migration (E-RTM), the dip-angle ADCIGs can be computed post-migration via the subsurface offset extension. We here obtain dip-angle ADCIG pre-migration in E-RTM by using Poynting vector, which is likely to compute during wavefield propagation. The reflection normal of PP imaging is the bisector of the scattering angle while that of PS imaging is not. We derive formulas to estimate PP and PS dip-angle ADCIGs, respectively, with some straightforward vector operations. Similar to the subsurface-offset method, our method also has dip-angle ADCIGs with the appearance of blocky horizontal coherence. According to the theory of local semblance analysis, the signal with better horizontal coherence has a higher semblance score, and vice versa. We can thus design a specularity filter to suppress incoherent noises according to their corresponding local semblance scores. We validate our methods with numerical examples. Both the Graben data and Marmousi data show that our methods work effectively in dip-angle ADCIG computation and the following noise suppression in E-RTM. We also apply our methods to the field data.