Earthquake Wave Propagation in Nevada Sedimentary Basins
AuthorScalise, Michelle Elyse
AdvisorLouie, John N
AltmetricsView Usage Statistics
Nevada is one of the most seismically active of the U.S. states, with large active faults adjacent to urban areas on sedimentary basins posing significant seismic hazard. The frequent background seismicity also provides a natural laboratory for event discrimination and nonproliferation research. We explore the seismic wave propagation effects associated with earthquake hazard and nuclear explosion monitoring in the four independent chapters of this dissertation. Focused in Reno, Nevada, urban ambient noise recording coupled with surface wave dispersion analysis to derive shear-wave seismic velocity profiles across the urban basin. Refraction Microtremor (ReMi) applied to long geophone arrays and extended recording times, images up to 1.7 and 3 km deep using 15- and 22-km-long geophone transects extending across the basin. Phase velocity uncertainties are 200 m/s below 1 second period and 600 m/s at longer periods. Dispersion data correlate well with adjacent smaller-scale deep ReMi surveys that imaged up to 1 km deep. Gravity derived basin thickness models correlate with the 2.0 km/s velocity boundary on the west side of the east-west oriented transect. ReMi results on the east side of Reno suggest basin depths are greater than what is modeled from gravity. Results inform future survey design and highlight the importance of linear array geometry and precise geophone spacing. Observed ground motions of the 2008 MW 4.9 Mogul earthquake in northwest Reno test the performance of three gravity derived basin geometry models of the Reno-area urban basin. Physics based 3D waveform modeling simulates ground motion from 0-3 Hz through 3D velocity models incorporating alternative basin geometries. The source model and seismic velocity assumptions are consistent across all models to isolate the effects of basin geometry on ground motion. Results indicate the Widmer basin model performs best near the Mogul subdivision, where it is more finely characterized and integrated with surface geological investigations. The Widmer model reproduces spectral velocity amplitudes better than the alterative models, but all lack the velocity heterogeneity to reproduce spectral velocity amplitudes above 1 Hz. The velocity models are too smooth and lack the scattering mechanisms that increase duration at both rock and proximal basin stations. Anomalous synthetic misfits near the Hidden Valley Golf course and extended observed durations suggest current basin models insufficiently characterize the ground under the Hidden Valley subdivision. All results emphasize the strong 3D wave propagation effects of shallow seismic sources, such as the 3.6 km deep Mogul mainshock. To investigate the wave propagation effects that generate shear energy from explosive sources, ground motions of the Source Physics Experiment are simulated from 0-5 Hz using high-performance computing and physics based 3D waveform modeling. Sensitivity tests of small-scale velocity heterogeneity represented by correlated random velocity perturbations using a Von Karman correlation function, show that the length scale and depth of scattering control the scattering efficiency. Models incorporating 3D geologic structure and small-scale 3D heterogeneity generate significant shear wave energy from isotropic sources at local distances. The small-scale heterogeneity improves the fit at high frequencies. Alternative source models test shear energy generated at the source, and scattering from lateral velocity heterogeneity is a larger contributor of shear motion at local distances (< 25 km). The 3D basin structure of Yucca Flat explains some of the inconsistent P/S ratio behavior at these distances, but cannot be fully reproduced with flat earth models. In preparation for the final phase of the Source Physics Experiment, the Rock Valley Direct Comparison, seismicity in the Rock Valley Fault Zone is relocated to quantify location accuracy and image subsurface structure. Absolute relocations locate events with median 1.09 and 0.5 km vertical (depth) and horizontal errors, respectively. Relative relocations locate events with an average vertical error of 180 m. The relocation effort utilizes two 1D regional velocity models. Absolute relocation results indicate 1D velocity models can accurately locate shallow event above a prominent velocity interface (or refractor) but are insufficient for locating shallow events between 1.5- 2 km depth in geologic settings with strong lateral heterogeneity. Relative relocation results image fault structures of the Rock Valley Fault Zone and confirm the high angle fault geometry assumed in the Rock Valley Geologic Framework Model. The independent studies demonstrate and quantify the challenges associated with shallow seismic sources in structurally complicated regions with strong lateral velocity heterogeneity. Results inform future earthquake hazard, event location and discrimination efforts.