P-S Wave separation in 3-D elastic seismic data

In geophysical exploration, reflected P- and S-wave have different velocities but are always coupled in solid earth. Receivers sense only specific displacement components without distinguishing P- or S-wave. An accurate velocity distribution is required to image the seismic energy at its reflecting point. A strategy is to separate reflected P- and S-waves before imaging such that the P- and S-wave can be imaged separately. This study intends to develop an algorithm to separate the reflected P- and S-wave in three-dimensional (3-D), three-component, prestack elastic seismic data. The 3-D elastic seismic data is downward extrapolated from the receiver locations into a vertically homogeneous elastic model (with P- and S-velocities) using the 3-D elastic wave equation. During downward extrapolation, divergence and curl of the displacement are computed and recorded independently at a certain depth. The P- and S-velocities in the elastic model are then split into two independent acoustic models. The divergence is upward extrapolated, using the acoustic wave equation, into the P-velocity model and extracted at the receiver locations to obtain the separated P-wave. The curl is extrapolated upwards, using the acoustic wave equation, into the S-velocity model and extracted at the receiver locations to obtain the separated S-wave. Tests are successful on synthetic seismic data computed for simple laterally heterogeneous 3-D models and if the velocities used in the extrapolations are incorrect.