Crustal shear-wave velocity anisotropy measurements determined from P-wave receiver functions: A possible new tool for determining upper crustal stress orientation

In this study we assessed the viability of using teleseismic receiver functions to infer the orientation of the maximum horizontal principal stress in the earth's crust from the direction of crustal shear-wave anisotropy. Two different methods were tested. The first utilized shear-wave splitting measurements of Psphases which sampled the entire crust. Data from thirty-eight US Array stations in the Permian Basin of west Texas and north-central Oklahoma were analyzed. Only 6%of the receiver functions passed stringent quality criteria that we established to identify crustal anisotropy. Unfortunately, nearly all of data analyzed had a high standard deviation (>25 degrees). Several stations that did pass the quality criteria had splitting directions that disagreed with the abundant independent stress indicators in both areas. It is possible that competition between fabric-induced anisotropy in the lower crust and stress-induced anisotropy in the upper crust may have prevented us from detecting a consistent direction of shear-wave splitting. The second method we investigated was an inverse approach that used the Neighborhood Algorithm to conduct a non-linear search in order to find best-fitting crustal models that allowed us to isolate stress-induced velocity anisotropy in the upper crust. This method was applied to teleseismic earthquakes recorded on four US Transportable Array stations in the Central Valley of California, four US Array stations in northern Oklahoma, four permanent stations in the vicinity of the San Andreas Fault in southern California, and one permanent station near Parkfield, California, essentially within the San Andreas Fault zone. Of the thirteen stations to which we applied this method, only three - all in California - had both a good apparent signal and a zero-lag cross-correlation coefficient of at least 0.24. The one station that was far from the San Andreas Fault aligned to the independent stress indicators and appears to indicate stress-induced anisotropy. The other two stations were close to the San Andreas Fault and seem to represent a mixture of structure- and stress-induced anisotropy.