Illuminating seismic waveguides using noise cross correlations and numerical simulations

Abstract:

Prediction of peak ground motions for scenario earthquakes may differ between theoretical and empirical approaches, adding uncertainty to seismic hazard estimates and risk mitigation. For example, the predictions of shaking in Los Angeles due to a large, southern San Andreas earthquake depend on the presence, or absence of, a seismic waveguide in the northern basins of the greater Los Angeles. Here, theoretical predictions of peak motions for northeastward ruptures depend on the (somewhat poorly constrained) details of the basin connectivity from the San Andreas fault to downtown Los Angeles in the Earth models. On the other hand, predictions from empirical methods, such as the Virtual Earthquake Approach (VEA), are produced independently of the velocity models. This study attempts to cross validate two established methods for ground motion prediction: 1) numerical simulations of the seismic wavefield in a 3D velocity and attenuation model (CVMS4.26) using the AWP-ODC solver, and 2) VEA using ambient noise cross correlations. We compare the empirical and numerical Green tensor for small earthquakes located on the San Andreas Fault and dense arrays of receivers in the basins. In particular, we explore and compare the peak amplitude of the correlation and Green’s function and the leakage of Love and Rayleigh waves on the off-diagonal terms of the correlation tensor.