We provide a realistic assessment of the uncertainties associated with plate reconstructions by creating empirical probability density functions (PDFs) for rotations between plate pairs based upon uncertainties in the positions of observed fracture zone and magnetic anomaly crossings. We determine the PDF for a sequence of rotations through Monte Carlo sampling of the PDFs for these plate pairs. We assign confidence limits within PDFs and find that worst-case uncertainty regions for global reconstruction poles can be twice as large as 95% confidence regions. We illustrate our technique with the anomaly 6 rotation between the Pacific and North American plates and demonstrate that realistic uncertainties for summations of rotations are apt to be significantly smaller than those previously predicted by earlier techniques. We develop revised apparent polar wander paths for the North American and Pacific plates through paleomagnetic Euler pole analysis. We determine motions of the Pacific plate within the hotspot reference frame for the past 80 Ma through a similar analysis and extend the model back to 145 Ma through study of anomalous basement depths and application of the paleomagnetic constraints. The match between the paleomagnetic and hotspot reference frames within uncertainty limits imposed by the available data, contradicts the notion of true polar wander during the late Cretaceous and/or motion between distinct Pacific plates. Applying realistic limits on the expected error in global plate reconstructions, we determine that either the South Atlantic hotspot track is improperly defined, or that there has been more than 1,000 kilometers of relative motion between the Pacific and South Atlantic hotspots since 80 Ma. Our model of plate interactions for the western Americas indicates subduction perpendicular to the margins from 78 Ma to 33 Ma. Northward transpression and transport of allochthonous terranes is limited to regions north of the evolving Pacific-Izanagi (Kula) ridge. We present a series of reconstructions which allows quantification of subduction and lateral transport along the margins. We model the generation of possible oceanic plateaus by known hotspots in the Pacific region and find little evidence to indicate that their subduction had an impact on the magmatic history of North America.

The Tibetan plateau is arguably the most important geological feature on Earth, yet its formation and evolution are poorly understood. This investigation utilizes Cretaceous sedimentary strata exposed in the Lhasa terrane of southern Tibet in order to constrain the paleogeography and tectonic setting of the region prior to the Indo-Asian collision. Lower Cretaceous strata consist of clastic sedimentary units that were deposited in shallow marine and fluvial environments. In northern Lhasa these sediments were deposited in a peripheral foreland basin that formed in response to the Lhasa-Qiangtang collision. The lower Cretaceous sediments in southern Lhasa are quartzose and were derived from cover strata exposed by local uplifts. A marine limestone of Aptian-Albian age overlies the lower Cretaceous clastic strata and was deposited in a shallow continental seaway. The paleogeography of the Lhasa terrane during deposition of the carbonate units was dominated by the effects of the Lhasa-Qiangtang collision, although other conditions, such as a high eustatic sea-level, influenced sedimentation as well. The Upper Cretaceous Takena Formation is composed of a basal member of marine limestone and an overlying member of fluvial red beds. The arkosic strata of the Takena Formation were deposited in a retro-arc foreland basin that formed to the north of the Gangdese magmatic arc. Collectively, the Cretaceous sedimentary strata indicate significant tectonic activity occurred in southern Tibet prior to the Indo-Asian collision. Moreover, the data suggest the crust of southern Tibet was thickened and possibly at high elevations before the Cenozoic.

The buildup of the cosmogenic nuclides ³He and ²¹Ne in surficial rocks permit exposure ages and erosion rates to be estimated. This dissertation extends the cosmogenic exposure technique to garnets, plagioclase with significant nonspallation ²¹Ne components, and alluvial fill terraces. Garnets from Nanga Parbat, Pakistan have low nucleogenic ³He and moderate radiogenic ⁴He concentrations. ³He exposure ages from garnets in glacial erratics indicate glacial advances at Nanga Parbat at about 16 ka and 55 ka. 3He in alluvial garnets suggests that denudation in small unglaciated basins proceeds 5 to 7 times slower than glacial erosion, and 10 to nearly 100 times slower than regional rock exhumation and surface uplift. Rocks older than several million years possess nucleogenic and mugenic ²¹Ne and ³He components that must be resolved for accurate exposure ages. These nonspallation components in plagioclase and clinopyroxene from the Miocene Columbia River Basalt Group are best isolated with shielded samples. Analyses of ⁴He, U, Th, and Li systematically underpredict the amount of nonspallation ²¹Ne and ³He present in shielded samples, probably because of mugenic production. Step heating experiments suggest that ²¹Ne diffusive loss from plagioclase is possible, but most samples do not exhibit such ²¹Ne loss. The ratio of ²¹Ne in plagioclase and ³He in clinopyroxene is generally constant after correction for the nonspallation component, indicating that little or no ²¹Ne loss has occurred. The last highly erosive floods at Grand Coulee occurred at about 21 ka, early in the cycle of Missoula flooding. Nuclide inheritance must be resolved for accurate exposure ages of stream fill terraces. Depth profiles of cosmogenic ²¹Ne in quartz from terraces on the Pajarito Plateau, northern New Mexico resolve nuclide inheritance. Three patterns of depth profiles are recognized: (1) downward decreasing; (2) downward increasing; and (3) uniform; types 2 and 3 are associated with cumulate deposits and bioturbation, respectively. Inheritance corrected exposure ages for the terraces agree with independent radiocarbon and soil development ages. Denudation rates estimated from the profiles are higher for fill terraces than for straths.

This dissertation explores the uses for data collected at broadband seismic stations to investigate upper mantle structures. In the Barents Sea region, we use seismic waveform modeling on data collected from arrays in Norway and Finland to investigate the nature of the Hales discontinuity in this area. We find that the unusually high velocities required by the move-out of the diffracted first arrival requires a discontinuity below the Moho, which we believe is probably caused by a phase transition from spinel to garnet peridotite. In Chile and Argentina, we use data collected during the Chile Argentina Geophysical Experiment to perform a regional travel time tomography in order to investigate the nature of the mantle above this unusual subduction zone. The northern half of the study area (between 30° and 33°S) is characterized by the central Chilean flat slab segment, where the descending Nazca slab dives to 100 km depth and then flattens, traveling horizontally for hundreds of kilometers before resuming its descent into the mantle. The Nazca plate in the southern half of the study area has a relatively constant dip of about 30°. The southern half exhibits normal arc volcanism roughly above the 100 - 125 km contours of the downgoing slab. The northern half has had no active volcanism in the past 2 Ma, and underwent an eastward displacement of arc volcanism beginning ~10 Ma. The northern half is also remarkable for the basement-cored uplifts of the Sierras Pampeanas. Our study of the upper mantle above the southern half indicates low P wave velocities, low S wave velocities, and high Vp/Vs ratios below the arc, consistent with partial melt. Above the flat slab segment we find low Vp, high Vs, and low Vp/Vs ratios. While the nature of the material responsible for these velocities cannot be uniquely determined, the velocities indicate it must be dry, cold, and depleted. In the transition from flat to normal subduction geometries, we find velocities consistent with frozen asthenosphere, which may have been displaced by the advancing flat slab during the Miocene.

The study of jadeite-hedenbergite binary pyroxenes has profound implications for interpreting the P/T history of eclogites and, in particular, ultra high pressure metamorphic rocks. Because binary pyroxenes are extremely rare in nature, a synthesis method was developed, and various compositions along the jadeite-hedenbergite join were produced. Synthesis products were analyzed using an X-ray diffractometer and microprobe to verify purity and homogeneity. Synthesized binary pyroxene was then mixed with quartz and albite and subjected to eclogitic conditions in order to determine the equilibrium reaction Ab <=> Jd + Qtz as a function of pressure, temperature, and X(jd). Two methods of reaction have been developed; one employs a graphite sample cup, and the other requires the use of welded gold sample cups, which reduce the temperature of reaction and allow for better pyroxene growth. All experiments were performed in a 1/2" or 3/4" piston cylinder apparatus at pressures from 15-39 kbar and temperatures from 650-1300°C.

Analysis of the alignment of geologic features and the use of GPS strain measurements are very different approaches to understanding crustal deformation histories and crustal and upper mantle properties. In this dissertation, two study areas with markedly different environments are evaluated using these approaches. The first study area includes the Guinea Plateau offshore West Africa that is part of a complex passive-margin system formed during two phases of rifting during the Jurassic and Cretaceous. Circular features revealed in two 3D seismic reflection surveys are interpreted to be extrusive volcanic features or vents emplaced after the cessation of Cretaceous rifting and opening of the Equatorial Atlantic Ocean. Statistically significant alignments of the vents implies that their distribution was influenced by faults or fractures not obvious in the seismic data alone. The existence of inferred alignments provides additional information about possible structures in the area of the volcanic vents that can be compared to more regional structures, giving better insight to magma migration and extrusion and structure of the Guinea Plateau. The alignment in one of the 3D survey areas is sub-parallel to oceanic fracture zones and continental lineaments that may extend into the survey and could have influenced the distribution of the volcanic features. The alignment in a separate 3D survey area is sub-parallel to the shelf-break and thought be related to inferred oceanward crustal thinning. Employing a different approach to the analysis of deformation, the second study area focuses on the Southern Basin and Range and Colorado Plateau, a weakly deforming area that is still capable of producing large earthquakes such as the 1887 Mw 7.5 Sonoran earthquake. To better constrain crustal motions and investigate the distribution of strain rates several hundred kilometers from the Pacific-North American plate boundary, an expanded GPS network of 34 sites was installed to complement existing continuous and campaign networks. Coseismic and postseismic deformation associated with earthquakes outside the study area, including the 4 April 2010 Mw 7.2 El Mayor–Cucapah, affected the GPS time series resulting in time-varying crustal surface velocities that obscured the background tectonic deformation. Through a deformation model, viscosities of the lower crust and upper mantle are estimated and the effects of earthquakes dating back to 1887 are removed from the time series to yield a time-independent or background secular velocity. A total velocity uncertainty is calculated that includes uncertainty of the time-independent velocities related to uncertainty in the viscosity estimates. Displacement histories are used to illustrate how earthquakes along the Pacific-North American plate boundary can temporarily impede extension in the Southern Basin and Range, particularly in southwestern Arizona. The time-independent velocities are used to calculate strain rates using latitudinal and longitudinal velocity profiles on one-degree increments. On a statistically significant basis, the velocity profiles are modeled with two linear segments rather than a single linear segment. Using the break points dividing the segments, the study area can be separated into a relatively lower-strain-rate eastern domain and a relatively higher-strain-rate western domain. The break points are interpreted to signify a boundary zone approximately 1000 km in length that overlaps tectonic and deformational boundaries described in previous studies. Comparing the time-invariant velocities with cumulative extensional slip rates of Quaternary faults across the area reveals a discrepancy in a portion of the area that is difficult to explain, but may be related to the time-varying velocities resulting from earthquakes on the Pacific-North American plate boundary.