Kinematic determination of Vp/Vs in multicomponent seismic data

Osareni Christopher Ogiesoba, Robert R. Stewart

This paper discusses a method of determining the zero-offset velocity ratio VP/SS (0) via velocity analysis, using Thomsen's (1999) non-hyperbolic traveltime equation. Processing of synthetic shot gathers from a four-layer geologic model provides the P-wave velocities (VP) and PS velocities (VS) used to stack the data sets. A semblance program written in Matlab code to compute semblance in 3-dimention is employed to scan for depth varying 0. It was observed that post-critical angle reflected events adversely affect the VP/VS values determined by this method. Thus, it is necessary to mute the data prior to velocity analysis. Accuracy increases as the depths to reflecting horizons increase. This is probably because the conversion point moves closer towards the midpoint with increasing depth. Furthermore, accuracy increases by decreasing the sampling interval in each variable used in the computation of semblance. That is increasing the number of samples in each variable. This however, linearly increases the computational cost. By converting the 0 and the PS zero-offset times to the corresponding P-wave times, the correlation of P-wave and PS stacked sections can be implemented. Correlation of the transformed PS data and the P-wave stacked data sets is found to be good (the difference being less than 5%).