Thesis (Ph. D.)--Texas A&M University, 2005. / "Major Subject: Geophysics" Title from author supplied metadata (automated record created on Sep. 15, 2006.) Vita. Abstract. Includes bibliographical references.

Thesis (M. S.)--Texas A&M University, 2006. / "Major Subject: Geophysics" Title from author supplied metadata (automated record created on Feb. 23, 2007.) Vita. Abstract. Includes bibliographical references.

Thesis (M. S.)--Texas A&M University, 2005 / "Major Subject: Geophysics" Title from author supplied metadata (automated record created on Feb. 23, 2007.) Vita. Abstract. Includes bibliographical references.

Earthquake catalogues are usually dominated by diffusive behaviour consistent with the Omori law of aftershocks. This is investigated in terms of waiting times, i.e. the time between successive events in a time-sorted earthquake catalogue. The theoretical waiting time probability distribution for the Omori law is derived and shown to predict the numerically produced Omori aftershock sequence well. These results enhance our understanding of aftershock processes and demonstrate that previous waiting time interpretations were severely flawed. Iceland earthquake catalogues are studied in terms of waiting times. Omori aftershock sequences are shown to predict most datasets well but there are some significant exceptions. One of these is data from the glacier covered Katla volcano in South Iceland, with few aftershocks. This dataset can be further split into two geographical groups: Several hundred volcano-tectonic earthquakes occurring within the caldera, reaching depths down to 15 km, and thousands of emergent low frequency earthquakes with a poorly defined shallow source in Goðabunga, in the western part of Katla. These events are investigated further. The lp events at Goðabunga have been recorded for decades and show a clear seasonal and climate-related correlation where their number increases in the autumn as well as during warmer years. Many of them form groups with very with similar waveforms. New broad-band seismic data suggests that the lp events originate in a steep outlet glacier covering Katla. Here, ice movement leads to ice falls over the steep escarpment, and we now believe that the lp events are generated by large ice falls rather than being related to gas or magma movements within the volcano, and are not precursors to an eruption as previously suspected. This observation probably has major significance for hazard estimation at the many ice-covered volcanoes around the world. We report near-field (vlp) signals simultaneous with the largest lp events. Our data is partly consistent in character with surface deformation (displacement and tilt) due to the ice movements. However, in line with results from elsewhere, the magnitudes of the observed effects are large relative to those from mathematical modelling. Our analysis suggests that the signal is not an instrumental artefact. Possible explanations are discussed.

Tertiary age rocks are exposed along north-south trending structures throughout the hinterland Altiplano plateau, central Andes. The east limb of the Corque syncline (SW Bolivia) contains the thickest and most continuous successions of late EoceneOligocene age non-marine strata (Potoco Formation) on the Altiplano. The Potoco is up to ~6500 m thick and has continuous exposure >10³ km² making it the thickest and most extensive remnant of the mid-Tertiary Altiplano basin. Basin fill during late EoceneOligocene time remains the most rapid and sustained period of deposition since Andean orogenesis, recording a long-term sediment accumulation rate of ~0.5 mm/my. The Potoco consists of fine-grained sandstone, mudstone, and gypsum deposits and lithofacies indicate deposition during periodic flooding in broad, poorly confined channels, and floodplain and playa lake environments. Deposits coarsen upward and channel deposits are thicker, less extensive, and more lenticular upsection. The thickness, extent, and ephemeral depositional style exhibited by the Potoco are attributed to rapid sediment accumulation facilitated by sheetflow-dominated rivers in a region undergoing rapid subsidence and arid climate conditions. Provenance data from the succession, including paleocurrent indicators and sandstone composition, document a paleoflow reversal accompanied by a distinct variation in composition. The basal ~4200 meters coarsens up and was deposited by east-directed rivers containing lithic sandstone, shistose, gneissic, granitic, and volcanic fragments. The overlying succession coarsens up and contains lithic sandstone, shistose, and volcanic fragments deposited by west-directed rivers. Provenance trends indicate two source areas west and east of the basin during late EoceneOligocene time. Thrust faults, active west and east of the basin during mid-Tertiary time, substantiate applying flexural modeling to determine if crustal shortening can account for basin development. Models incorporating effective elastic thickness west (20 km) and east (13.5 km) of the deepest portion of the basin and basin depth (~4200 meters east-directed; ~3000 meters west-directed) and length (120 km east-directed; 75 km west-directed) reveal that subsidence from shortening, while non-unique, is one viable mechanism for basin formation. Potoco sedimentology, stratigraphy, provenance trends, and model results provide a testable tectonic model to explain evolution of the Altiplano Basin during mid-Tertiary orogenesis.

The Upper Cretaceous-Paleocene El Molino Formation of the Bolivian Central Andes consists of mixed siliciclastic and carbonate strata alternately interpreted as syn-rift, post-rift thermal sag and foreland basin deposits. These deposits can be divided into two lithostratigraphic sequences. The first sequence consists of a carbonate, carbonate sand and mudrock lower member, a middle member consisting entirely of mudrock and an upper member containing carbonates and mudrocks. The second lithostratigraphic sequence contains a lower member composed of carbonate sands, carbonates and mudrocks, and an upper member consisting of a coarsening upwards sequence of sandstones and mudrocks. Within these lithostratigraphic sequences, five facies associations can be identified: 1) an open water facies; 2) a nearshore facies; 3) a beach, bar and shoal facies; 4) a floodplain facies; and 5) a fluvial facies. A regional study of El Molino Formation stratigraphic stacking patterns and facies association geographic distributions suggests that deposition occurred within a dominantly lacustrine basin. For most of El Molino Formation deposition, the lacustrine system remained hydrologically-closed and perennial, although evidence indicates that depositional systems experienced periodic ephemeral lacustrine conditions as well as hydrologically open lacustrine and/or shallow marine depositional environments. While lacustrine systems exist in syn-rift, post-rift thermal sag and foreland basin systems, sedimentological and stratigraphic data, in addition to an absence of key indicators of tectonic activity (e.g. faulting, growth strata) limit the El Molino Formation tectonic setting to post-rift thermal sag and/or foreland basin back-bulge settings. Paleocurrent and provenance data further support the interpretation for post-rift thermal sag and/or back-bulge basin deposition, indicating flow of continental block provenance sediment into a central depositional basin while clast count data show a simple unroofing sequence indicative of the tectonic quiescence associated with post-rift thermal sag and foreland basin back-bulge tectonic settings.

This thesis proposes a computationally efficient method for approximating the inverse of the normal operator arising in the linearized inverse problem for reflection seismology. The inversion of the normal operator using direct matrix methods is computationally infeasible. Approximate inverses estimate the solution of the inverse problem or precondition iterative methods. Application of the normal operator requires an expensive solution of large scale PDE problems. However, the normal operator approximately commutes with pseudodifferential operators, hence shares their near diagonality in a frame of localized monochromatic pulses. Estimation of a diagonal representation in this frame encoded in the symbol of the normal operator: (1) follows from its application to a single input vector; (2) suffices to approximate its inverse. I use an efficient algorithm to apply pseudodifferential operators, given their symbol, to construct a rapidly converging optimization algorithm that estimates the symbol of an inverse for the normal operator, thereby approximately solving the inverse problem.

The time-varying stress field of fault systems at seismogenic depths plays the mort important role in controlling the sequencing and nucleation of seismic events. Using seismic observations from repeating earthquakes, controlled active sources and seismic ambient noise, five studies at four different fault systems across North America, Central Japan, North and mid-West China are presented to describe our efforts to measure such time dependent structural properties. Repeating and similar earthquakes are hunted and analyzed to study the post-seismic fault relaxation at the aftershock zone of the 1984 M 6.8 western Nagano and the 1976 M 7.8 Tangshan earthquakes. The lack of observed repeating earthquakes at western Nagano is attributed to the absence of a well developed weak fault zone, suggesting that the fault damage zone has been almost completely healed. In contrast, the high percentage of similar and repeating events found at Tangshan suggest the existence of mature fault zones characterized by stable creep under steady tectonic loading. At the Parkfield region of the San Andreas Fault, repeating earthquake clusters and chemical explosions are used to construct a scatterer migration image based on the observation of systematic temporal variations in the seismic waveforms across the occurrence time of the 2004 M 6 Parkfield earthquake. Coseismic fluid charge or discharge in fractures caused by the Parkfield earthquake is used to explain the observed seismic scattering properties change at depth. In the same region, a controlled source cross-well experiment conducted at SAFOD pilot and main holes documents two large excursions in the travel time required for a shear wave to travel through the rock along a fixed pathway shortly before two rupture events, suggesting that they may be related to pre-rupture stress induced changes in crack properties. At central China, a tomographic inversion based on the theory of seismic ambient noise and coda wave interferometry clearly reveals a coseismic velocity decrease region with the strike and length strikingly matching the fault zone of the 2008 M 7.9 Wenchuan earthquake at depth. We speculate the imaged decrease velocity region resulted from decreased crustal stress around the fault zone at upper crust.

This investigation into Antarctic Peninsula Ice Sheet (APIS) development represents research from the stratigraphic record of three geographic areas: The James Ross Basin (northwestern Weddell Sea), the Pacific continental margin of the Antarctic Peninsula, and the Joinville Slope (northwestern Weddell Sea). The stratigraphic architecture of the James Ross Basin, NW Weddell Sea continental shelf, shows three major phases of deposition: pre-glacial, ice sheet growth, and ice sheet dominated. Each stratigraphic unit is characterized based upon seismic facies and stratigraphic architecture, and the ages are inferred from a seismic stratigraphic age model. A total of 34 grounding events of the Antarctic Peninsula Ice Sheet (APIS) are recorded on the continental shelf. The seven oldest glacial unconformities are believed to pre-date all previously identified unconformities on the peninsula continental shelf. An expanded section of Late Pliocene/Pleistocene deposits show a minimum of 10 grounding events. Isopachs of sedimentary sequences on the Antarctic Peninsula Pacific continental margin show shifting depocenters through time. Chronostratigraphic and seismic depth-converted data from ODP 178 cores allow the calculation of sediment flux for shelf units S3-S1 and rise units M6-M1. Sediment flux to the margin increases from the Late Eocene until the Late Pliocene and then decreases slightly from Late Pliocene to present. Significant increases in sediment flux coincide with early development of the APIS and during the early Pliocene warming period (Barker and Camerlenghi, 2002). Minimum glacial denudation rates for the Antarctic Peninsula are in the range of 0.06 to 0.13 mm yr -1. The Joinville Slope sediment wedge located in the northwestern Weddell Sea shows seismic stratigraphic evidence of mixed turbidite/contourite/hemipelagic deposition. A prominent seafloor unconformity and the exposed and eroded basement of the adjacent continental shelf indicate erosion by grounded ice during the Plio-Pleistocene. SHALDRIL recovered core at three drill sites, 12A, 5C, and 6D, and sampled sediments from the upper Oligocene, middle Miocene, and lower and upper Pliocene which are constrained by diatom and calcareous nannofossil assemblages. The sediment wedge shows no apparent hiatuses or large unconformities from Late Oligocene to the Lower Pliocene. Regional sedimentation rates show continuous sedimentation throughout the Late Paleogene and Neogene.

This thesis develops an SH-wave version of frequency-domain, full waveform tomography, and applies it, together with traditional acoustic waveform tomography, to a multicomponent seismic data set acquired over a shallow contaminated aquifer at Hill Air Force Base, Utah. The study combines the high resolution provided by waveform tomography with inherent advantages of SH-wave imaging, such as reduced seismic velocity and independence of pore fluid content. Presented are synthetic tests of the method, its application to the field data, and interpretation of the resulting P- and S-wave velocity models. Synthetic tests reveal fundamental differences between acoustic and SH waveform tomography, and demonstrate, together with the field data inversions, improved resolution for SH-wave imaging due to smaller velocities. High-resolution velocity models from inversion of the field data are interpreted in terms of lithology and water saturation, which are better constrained by the availability of both P- and S-wave velocity.