Earthquake locations and three‐dimensional crustal structure in the Coyote Lake Area, central California

TLDR

The simultaneous inversion method has been refined to iteratively solve for earthquake locations and laterally heterogeneous crustal structure. The authors use approximate ray tracing with parameter separation to invert P‑wave arrival times from the USGS Central California Network, producing a three‑dimensional upper‑crust velocity model. The resulting velocity model correlates strongly with geological, gravity, and magnetic data, and the refined earthquake relocations achieve ~1 km epicentral accuracy, revealing a vertical San Andreas fault with no offset and two active surfaces on the Calaveras fault.

Abstract

Previous work on the simultaneous inversion method has been improved and extended to incorporate iterative solution for earthquake locations and laterally heterogeneous structure. Approximate ray tracing and parameter separation are important elements of the improved method. Application of the method to P wave arrival time data recorded by stations of the U.S. Geological Survey Central California Network yields a three‐dimensional model for the velocity structure of the upper crust in an area encompassing the rupture zone of the Coyote Lake earthquake of August 1979. Very strong correlations between the velocity model and the geology and gravity and magnetic anomalies are observed. Improved estimates of the locations of earthquakes in the study area are also determined. The relocation of explosions indicates epicentral accuracies of the order of a kilometer or better. Based on the revised hypocentral locations, it is concluded that the San Andreas fault is vertical in this area, with no actual offset between the epicenters and the fault trace. In contrast, the Calaveras has two (or more) active fault surfaces, one nearly vertical and another dipping 75° to the northeast.