xii, 95 p. : ill. (some col.) / Vigorous convective activity in the western U.S. mantle has long been inferred from the region's widespread intra-plate crustal deformation, volcanism, and high elevations, but the specific form of convective activity and the degree and nature of lithospheric involvement have been strongly debated. I design a seismic travel-time tomography method and implement it with seismic data from the EarthScope Transportable Array and complementary arrays to constrain three-dimensional seismic structure beneath the western U.S. Tomographic images of variations in compressional velocity, shear velocity, and the ratio of shear to compressional velocity in the western U.S. mantle to a depth of 1000 km are produced. Using these results I investigate mantle physical properties, Cenozoic subduction history, and the influence of small-scale lithospheric convection on regional tectonic and magmatic activity, with particular focus on southern California and the Pacific Northwest. This dissertation includes previously published co-authored material. Chapter II presents a travel-time tomography method I designed and first implemented with data from southern California and the surrounding southwestern U.S. The resulting images provide a new level of constraint on upper mantle seismic anomalies beneath the Transverse Ranges, southern Great Valley, Salton Trough, and southwestern Nevada volcanic field. Chapter III presents tomographic images of the western U.S. mantle, identifies upper mantle volumes where partial melt is probable, and discusses implications of the apparently widespread occurrence of gravitational instabilities of continental lithsophere and the complex geometry and buoyancy of subducted ocean lithosphere imaged beneath the western U.S. In Chapter IV, tomography images are used in conjunction with geologic constraints on major transitions in crustal deformation and magmatism to construct a model for Pacific Northwest evolution since the Cretaceous. Accretion in the Pacific Northwest at 55-50 Ma is suggested to stimulate roll-back of the flat subducting Farallon slab. This change in convergent margin structure is further suggested to drive the short-lived Challis magmatic trend and trigger the southward propagating Eocene-Oligocene transition from the Laramide orogeny to widespread crustal extension and ignimbrite magmatism. / Committee in charge: Eugene Humphreys, Chair; Douglas Toomey, Member; Emilie Hooft Toomey, Member; John Conery, Outside Member

The structure and regional tectonic setting of an exhumed, 9.3-km long, left-lateral strike-slip fault zone eludicates processes of growth, linkage, and termination for strike-slip fault zones in granitic rocks. The Gemini fault zone is composed of three steeply dipping, southwest-striking, noncoplanar segments that nucleated and grew along preexisting joints. The fault zone has a maximum slip of 131 m and is an example of a segmented, hard-linked fault zone in which geometrical complexities of the faults and compositional variations of protolith and host rock resulted in nonuniform slip orientations, complex interactions at fault segments, and an asymmetric slip-distance profile. Regional structural analysis shows that joints and left-lateral fault zones have accommodated slip within a 4.8-km wide, right-lateral monoclinical kink band with vertical fold axes and northwest-striking axial surfaces. Geometric modeling of the kink band indicates that as little as 1.1 km of right-lateral displacement across the kink band may have produced the observed slip on kilometer-scale faults within the kink band.

The objective of this study is to evaluate the regional differences between the worldwide based NGA-W1 ground motion models and available Turkish strong ground motion dataset and make the required adjustments in the NGA-W1 models. A strong motion dataset using parameters consistent with the NGA ground motion models is developed by including strong motion data from Turkey. Average horizontal component ground motion is computed for response spectral values at all available periods using the GMRotI50 definition consistent with the NGA-W1 models. A random-effects regression with a constant term only is used to evaluate the systematic differences in the average level of shaking. Plots of residuals are used to evaluate the differences in the magnitude, distance, and site amplification scaling between the Turkish dataset and the NGA-W1 models. Model residuals indicated that the ground motions are overestimated by all 5 NGA-W1 models significantly, especially for small-to-moderate magnitude earthquakes. Model residuals relative to distance measures plots suggest that NGA-W1 models slightly underestimates the ground motions for rupture distances within 100-200 km range. Models including the aftershocks over-predict the ground motions at stiff soil/engineering rock sites. The misfit between the actual data and model predictions are corrected with adjustments functions for each scaling term. Turkey-Adjusted NGA-W1 models proposed in this study are compatible with the Turkish strong ground motion characteristics and preserve the well-constrained features of the global models. Therefore these models are suitable candidates for ground motion characterization and PSHA studies conducted in Turkey.