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

Abstract

Angle-domain common-imaging gathers (ADCIGs) are important in analyzing the subsurface discontinuities where reflection waves take place. In elastic reverse time migration (ERTM), dip-angle ADCIGs can be computed postmigration via subsurface offset extension. We have obtained dip-angle ADCIG premigration in ERTM by using the Poynting vector, which was easy to compute during wavefield propagation. The reflection normal of PP imaging is the bisector of the scattering angle, whereas that of PS imaging is not. We derive formulas for PP and PS dip-angle estimations, respectively, with some straightforward vector operations. Similar to the subsurface-offset one, our method also outputs dip-angle ADCIGs with the appearance of blocky horizontal coherence. According to local semblance analysis, the signal with a better horizontal coherence promises a higher semblance score, and vice versa. We can thus design a specular filter to suppress incoherent noises according to their corresponding local semblance scores. We validate our methods with numerical examples. The Graben and Marmousi data sets show that our methods work effectively in dip-angle ADCIG computation and the following noise suppression in ERTM. We also examine our methods with one field data set.