An elastic reverse-time migration method for vertically transversely isotropic (VTI) media

Summary Following the isotropic elastic reverse-time migration (RTM), an elastic RTM method for VTI media is proposed. The proposed method exploits polarization vector projection to separate anisotropic elastic wavefields. These polarization vectors, estimated at each point of the medium, can be treated as localized convolutional operators, which resemble conventional derivative operators. This separation method perform well in heterogeneous VTI media, but it is very computationally time consuming. In order to improve the efficiency of wavefield separation, lowrank singular value decomposition (SVD) is used to approximate space-domain separation operators. Another problematic issue is polarity reversals in converted-wave image. A polarity reversal correction method for isotropic elastic RTM is extended to solve this problem. First, a sign factor, representing the polarity distribution of the S-wave component, is computed during the wavefield reconstruction for every imaging point. Then, the polarity reversal is corrected by multiplying the PS and SP images with the sign factor at every time step when an elastic imaging condition is applied. Simple flat layer model and complex model are used to test the proposed anisotropic elastic RTM and to demonstrate the potential of the proposed method.