S-wave splitting: Key to earthquake prediction?

This paper presents a detailed analysis of S-wave data for two areas in western Nevada. Some of these data were used by Gupta (1973b, c) as the basis for his claim that large premonitory changes in the extent of stress-induced S-wave splitting are observed for moderate-sized earthquakes in Nevada. Analysis of particle motion for an earthquake sequence near Slate Mountain indicates that changes in S-wave splitting did not occur during that sequence. For the Mina area, comparison of S-wave signatures for 158 events occurring over a 3-year period resulted in the identification of numerous events that would be considered anomalous by Gupta's criteria, but these were not followed by larger earthquakes. The present state of knowledge on crustal structure and seismic source parameters in the western Basin and Range province is not sufficient to discriminate between stress-induced velocity anisotropy and the many other factors that contribute to the complexity of S-wave signatures.