An excitation potential imaging condition for elastic reverse time migration

Abstract

Elastic reverse time migration (ERTM) has been demonstrated to be more accurate than scalar RTM. However, low efficiency (large storage and heavy calculated amount) and strong artifacts caused by the crosstalk between different wave modes are the two primary barriers to the application of the ERTM during the processing of real data. The scalar (P) and vector (S) potentials of the elastic wavefield and the arrival times corresponding to the first energy extremum of the wavefield are saved at each grid point during the forward modeling of the source wavefield. The angle-dependent reflection coefficient images are subsequently obtained by dividing the scalar and vector potentials of the backward extrapolated receiver wavefield by the saved scalar and vector potentials at the grid points that satisfy the image time at each time step, respectively. The proposed imaging condition does not need to store the snapshots of the source wavefield, while it can considerably improve the computational efficiency and decrease the amount of storage and Input/Output manipulation (compared with the cross-correlation imaging condition) in addition to suppressing the crosstalk between compressive and shear wave modes. Compared with the excitation time imaging condition, the proposed imaging condition reduces the energy loss caused by the opposite polarity of the horizontal component at opposite sides of the source in stacked images. Numerical tests with synthetic data of the Sigsbee2a model have demonstrated that this imaging condition is a cost-effective and practical imaging condition for use in prestack ERTM.