Seismic Tomography of the Sacramento -- San Joaquin River Delta: Joint P-wave/Gravity and Ambient Noise Methods

The Sacramento -- San Joaquin River Delta (SSJRD) is an area that has been identified as having high seismic hazard but has resolution gaps in the seismic velocity models of the area due to a scarcity of local seismic stations and earthquakes. I present new three-dimensional (3D) P-wave velocity (Vp) and S-wave velocity (Vs) models for the SSJRD which fill in the sampling gaps of previous studies.;I have created a new 3D seismic velocity model for the SSJRD, addressing an identified need for higher resolution velocity models in the region, using a new joint gravity/body-wave tomography algorithm. I am able to fit gravity and arrival-time residuals jointly using an empirical density-velocity relationship to take advantage of existing gravity data in the region to help fill in the resolution gaps of previous velocity models in the area. I find that the method enhances the ability to resolve the relief of basin structure relative to seismic-only tomography at this location. I find the depth to the basement to be the greatest in the northwest portion of the SSJRD and that there is a plateau in the basement structure beneath the southeast portion of the SSJRD. From my findings I infer that the SSJRD may be prone to focusing effects and basin amplification of ground motion.;A 3D, Vs model for the SSJRD and surrounding area was created using ambient noise tomography. The empirical Green's functions are in good agreement with published cross-correlations and match earthquake waveforms sharing similar paths. The group velocity and shear velocity maps are in good agreement with published regional scale models. The new model maps velocity values on a local scale and successfully recovers the basin structure beneath the Delta. From this Vs model I find the maximum depth of the basin to reach approximately 15 km with the Great Valley Ophiolite body rising to a depth of 10 km east of the SSJRD. We consider our basement-depth estimates from the Vp model to be more robust than from the Vs model.