Seismo Lab Seminar

The Statewide California Earthquake Center (SCEC) Community Velocity models (CVMs) provide a great resource for ground motion modeling, but the accuracy of the resulting ground motions is limited by their generally coarse resolution. However, recent studies have shown that multi-scale models (e.g., models with varying resolution due to inversion strategies, ray path coverage, and data integrity) generated by combining locally imaged models into coarser reference models can significantly improve the fit between simulations and data (e.g., Yeh and Olsen, 2023; Zhang and Ben-Zion, 2023). We have started to assemble and test the ground motion prediction efficacy for a multi-scale CVM generated by embedding a series of high-resolution models from central and southern California into the SCEC CVM-S4.26M01 as the reference model. The high-resolution velocity models are generated from travel time tomography, joint inversion of short-period Rayleigh wave ellipticity, phase velocity, teleseismic receiver functions, and ambient noise cross-correlation from land and ocean-bottom recordings. We validate the assembled multi-scale CVM by comparison of seismic recordings from a series of local M4-5 earthquakes to physics-based simulations using the scalable 4th-order finite-difference methods AWP-ODC (Cui et al., 2013) including surface topography, anelastic attenuation, and a discontinuous mesh.