S23B-2700
Converted-waves Imaging Condition for Elastic Reverse-Time Migration with Decomposed Wavefields

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Hyungwook Choi, Soon Jee Seol and Joongmoo Byun, Hanyang University, Seoul, South Korea
Abstract:
To successfully deal with responses from the elastic earth, imaging techniques need to incorporate the elastic wave equation. Elastic Reverse-Time Migration (ERTM) with separating-while-imaging approach is capable of yielding physically meaningful PP, PS, SP, and SS images from multicomponent data. Even in PP images, ERTM has brought enhancements comparing to those from acoustic RTM because ERTM can handle converted waves. Converted-wave images, core results of ERTM, however, have two major problems related to characteristics of S-waves. First, polarity reversals according to propagation directions of S-waves cause destructive effect to final PS and SP images while each migrated result is stacked over the shots. In addition, non-existent spurious events which are produced by crosscorrelating downgoing S-waves in source wavefields and reflections associated with downgoing P-waves in receiver wavefields lead masking effects over true reflection events in SP and SS images. In this study, we adopt a wavefield decomposition method to solve the polarity problems and derive a new converted-wave imaging condition for SP and SS images to alleviate the generation of spurious events. The acceleration vector wavefield decomposition method used in our ERTM has advantages over the conventional wavefield separation method based on the Helmholtz decomposition because the wavefield decomposition can automatically compensate polarity changes in PS and SP images when the zero-lag crosscorrelation for vector wavefields is applied. To suppress spurious events in SP and SS images, our imaging condition is designed to make images only where S- and converted P-waves from source wavefields are coexisted with decomposed wavefields from receiver wavefields at reflection boundaries. To verify our new imaging condition, we tested our algorithm with OBC (Ocean Bottom Cable) data from elastic Marmousi-II model and compared the migrated images with those from ERTM with the zero-lag crosscorrelation imaging condition. The results evidently show that (c) SP and (d) SS images from our new imaging condition provide much clearer images without masking effects caused by spurious events than (a) SP and (b) SS images from the conventional imaging condition.