Application of surface wave travel times and amplitude ratios interpreted through a 3D crustal model to locate and characterize regional seismic events in the US

Abstract:

The error in the epicentral location of crustal earthquakes across the contiguous US is on the order of 10 km due to the inability of 1D seismic velocity models to capture regional body wave travel time variations. New high resolution 3D models of the crust and uppermost mantle have been constructed recently across the US by inverting surface wave dispersion from ambient noise and earthquakes, receiver functions, and Rayleigh wave H/V ratios using USArray data [e.g., Shen et al., 2013]. These are mostly S-wave models of the lithosphere, however, which are not optimal for predicting regional P-wave travel times. We explore the use of observations of surface waves to improve regional event characterization because the new 3D models are constructed explicitly to model their behavior. In particular, we use measurements of group and phase time delays and the amplitude ratio between different periods of surface waves to estimate the moment tensor, the epicentral location and the earthquake depth. Preliminary estimates of these variables are determined through a simulated annealing algorithm. Afterward, a Bayesian Monte Carlo method is applied to estimate the posterior distribution of all variables in order to assess uncertainties in source characteristics. The reliability and limitations of the location method are tested by systematic relocation of earthquakes across the contiguous US.