Seismic Simulations of the August 24, 2014 South Napa and Nearby Earthquakes to Investigate the Effects of Rupture Details and Three-Dimensional Structure on Ground Motions and Static Displacements

We performed three-dimensional (3D) anelastic finite difference simulations to investigate the effects of rupture details and sub-surface structure (including surface topography) on motions for the August 24, 2014 South Napa M6.0 earthquake, its aftershocks and nearby earthquakes. We compared observed seismic ground motions and static displacements (from GPS and InSAR) with motions computed with 3D simulations. Calculations were performed with the SW4 finite difference code, which computes the transient motions and permanent displacements including the effects of 3D structure based on the USGS San Francisco Bay Area 3D model (version USGSBayAreaVM-08.3.0.etree) and surface topography. We investigated the response using both an average plane-layered (one-dimensional, 1D) and the USGS 3D structure. We computed motions for available models of the mainshock rupture (including the model reported by Dreger, 2014 from broadband seismic waveforms). To unravel the effects of earthquake rupture and propagation through complex 3D structure, we modeled both the M6 mainshock and smaller (M3-4) nearby events, which can be represented as simple point moment tensor sources. The computed ground motions were compared to observations. For the seismic ground motions, the simulated motions with the 3D model (up to 1 Hz) produce remarkably better fits to the data than the average 1D model, suggesting there are significant path effects due to sedimentary basins and other sub-surface features. Aftershocks are well modeled for most paths, including paths crossing sedimentary basin structures in the Napa Valley and San Pablo Bay, indicating that the USGS 3D model can reasonably predict complex path effects. The M6.0 South Napa mainshock ruptured toward the north and resulted in complex high-frequency response in the backward directivity direction to the south, which we have not yet been able to reproduce with simulations and may require addition of finer details in the rupture model. Preliminary comparisons of simulations with geodetic displacements reported from GPS data suggest 3D structure slightly alters directions and amplitudes of displacements. We will attempt to model the geodetic displacements reported from both GPS and InSAR data.