Deltas are dynamic natural and socio-economic environments that are hypersensitive to the effects of urbanization, natural disasters and climate driven sea level change. Deltaic groundwater resources are often susceptible to degradation from saline water through seawater encroachment, tidal channel mixing and interaction with brackish paleowaters. The Ganges-Brahmaputra-Meghna River delta, located in Bangladesh, is a particularly vulnerable area with an estimated twenty million coastal inhabitants directly affected by saline drinking water. The shallow groundwater of the coastal regions is primarily brackish with pockets of fresher water. A small scale hydrologic investigation of groundwater salinity beneath a tidal channel island was undertaken to explore possible hydrogeological explanations of the distribution of water salinities in the shallow aquifer. This study employs a combination of 3H and 14C dating, electromagnetic (EM) subsurface mapping, and a 2D solute transport model. The EM and isotope data suggests the development of freshwater lenses, sustained by meteoric recharge, above paleo-brackish estuarine water that was deposited during the Holocene. The transport model indicates an advection dominated system which is insensitive to small changes in aquifer architecture.

Mathematical descriptions of sediment transport are essential for our understanding of the evolution and form of Earth's surface. Recent work has highlighted the potential strengths of a nonlocal mathematical description of the hillslope sediment flux in steepland settings. Theory for nonlocal hillslope sediment transport has largely outpaced field or empirical studies designed to inform it. Here I present topographic data from tectonically and climatically diverse regions that is largely consistent with theory from nonlocal sediment transport. I use these data to show how one may extract the ratio of uplift rate to transport activity, which is the central factor in determining relief. I further isolate transport activity and show a direct connection between transport mechanisms and mean annual precipitation. This ultimately offers a physically-based connection between climate, sediment transport mechanisms, and hillslope relief. A second part of this thesis explores the stability behavior of hillslopes evolving by nonlocal sediment particle motions. A stable system unconditionally returns to an initial condition whereas an unstable system amplifies perturbations at characteristic wavelengths. Results suggest that nonlocal formulations of sediment transport with non-uniform entrainment rates are unstable. Numerical analyses suggest that hillslope roughness scales of 1-4 meters may be caused by instability introduced by nonlocal transport. Indeed these scales are consistent with $approx10$ meter scale roughness observed in natural settings. This result has implications for how geomorphologists interpret spatial variability of soil thickness which may reflect unstable hillslope behavior.

Here sediment provenance from 48 drill cores across a fluvial valley is used to highlight fluvial channel dominance, anticlinal Pleistocene structures, Holocene channel occupations, and Holocene tectonic channel steering in the Meghna valley, which is alongside a convergent margin in the Ganges Brahmaputra Meghna Delta. Sediment provenance shifted in the late Pleistocene and early Holocene from being supplied by more local rivers that were part of the Barak, Surma, and Kushiara rivers that drain the Indo-Burman ranges to being mainly supplied by the Brahmaputra River, which accounts for the majority of sedimentation in the Holocene. Places in the Meghna valley that do have Pleistocene Brahmaputra sediment, appear to be rare, older than the latest Pleistocene channel-levee sand deposits, and are linked to anticlinal folds. The center of Lalmai anticline, which outrcrops in the Meghna valley, is comprised of ~50 m of Brahmaputra braidbelt sands, while the western side of the valley has several cores with Pleistocene Brahmaputra sediments that are likely buried anticlinal structures with deformation tilt rates allowing for the Brahmaputra to be excluded from the Meghna valley in the late Pleistocene. The Holocene occupation history of the Meghna valley by the Brahmaputra River of several occupations of varying time from this study agrees with previous work. Based on Brahmaputra Holocene deposits and cross valley gradient that slight tilting of the land surface due to deformation alone managed to steer the main channel at least 50 km westwards during the Holocene during the 7.5 to 5 kyr occupation.

Soil (n=45) and water (n=111) samples were collected in SW Bangladesh in the wet (May) and dry (November) seasons in 2016 to investigate the effects of irrigation by tidal channel water and groundwater on concentrations of arsenic and salts in soil. Between 1.4% and 2.1% of arsenic in soils is soluble, indicating that most arsenic is bound to soil minerals. Bulk soil arsenic is positively correlated with arsenic in irrigation water, suggesting that tube well irrigation adds arsenic to soils. Soluble As in rice paddy soil extracts is positively correlated with concentrations of sulfur and organic carbon in extracts and negatively correlated with S concentration in bulk soil, suggesting that As is released from soil by sulfide dissolution and possibly complexes with organic carbon in solution. Bulk soil arsenic is higher in rice paddies during the wet season, suggesting that reducing conditions present in waterlogged soils cause As-rich sulfide precipitation. Rice paddy soils irrigated with As-rich tube well water do not have higher bulk or soluble arsenic concentrations than paddies irrigated by tidal channel water. Salts in water are concentrated during the dry season due to decreased precipitation and increased evaporation rates. Soluble log median Na soil concentrations fall between 158 ppm and 1,259 ppm, while log median Na concentrations in bulk soils range between 6,310 ppm and 10,000 ppm, indicating the binding of Na to clay minerals. Comparison of soil/sediment samples from rice paddies, tidal channels, and the Sundarbans suggests that cultivation reduces dry season soluble salt and As concentrations in rice paddy soils.

Earth and Environmental Sciences Perceptions of Water Quality in Southwestern Bangladesh Rachel Shumaker Thesis under the direction of Professor Jonathan M. Gilligan Increasing stress on Southwestern Bangladesh has raised concerns about the well-being of the many people living in rural villages living in this region. This has led researchers to focus on the primary source of stress: water quality and water scarcity. However, the dynamic nature of the land and people of Bangladesh require a more interdisciplinary approach when exploring perceptions of water quality access to safe drinking water is affected not only by the chemistry of the water, but also the politics governing access to water and the psychology of peopleâs perceptions of drinking water. Lack of reliable year-round drinking water sources, coupled with salinity, arsenic contamination, and conversion of paddy fields to shrimp farms all contribute to water scarcity in coastal regions of Bangladesh. In our study area, we found that most people do not have access to safe drinking water, regardless of social or economic status. Despite the poor quality of the drinking water, most people think their water tastes good, so there is a mismatch in perceptions of both taste and saltiness, and water quality. The perception of water quality is not affected by socio-economic status, religion, or gender, and water security remains a challenge even for the relatively well to do. Approved: Jonathan Gilligan, Ph.D.

The removal of sediment mass to deposition in sedimentary systems creates predictable patterns of downstream fining and facies changes. Using Holocene lithology and stratigraphic architecture from a dense borehole network, the mid-Holocene sedimentary record of the Ganges-Brahmaputra-Meghna Delta (GBMD) is examined in a mass balance framework to demonstrate a shift from bypass-dominated to extraction enhanced modes of sediment dispersal. A prolonged (~3000 year) occupation of the tectonically-influenced Sylhet basin occurred from ~ 7.5-4.5 ka, producing thick (20-30 m) packages of channel sands and splay deposits. Sediment was initially (~7.5-6.0 ka) routed along the western edge of Sylhet basin, where confinement in antecedent valleys promoted bypass to downstream depocenters. At ~6.0-5.5 ka, the system shifted to an extraction enhanced mode of sediment dispersal, as a large (3500 km2) splay developed northeast of the main sediment delivery pathway. This pattern of deposition persists through the end of the occupation (~4.5 ka) as reflected in another large (~4000 km2) splay along the Shillong Foredeep at the basinâs northern margin. The persistence of sediment bypass suggests that seasonal flooding of the basin by local runoff exerts a hydrologic barrier to overbank flow and is thus a principal control on river path selection. The progradation of two 10-30 m thick, sandy mega-splays into the basin interior reflect an alternative mode of sediment dispersal that appears to operate only in the mid Holocene (~6.0-4.0 ka) during a regional weakening of the summer monsoon. The reduced water budget at this time would have lowered seasonal water levels in the basin, temporarily lessening the hydrologic barrier effect and facilitating splay development into the basin interior. The spatial distribution of sediment mass contained in each of the three pathways was apportioned spatially to calculate ï£, a scale-independent dimensionless downstream distance that represents the total percent of supplied sediment flux lost to deposition at any given point in the system. Both of the splay deposits contain distinct facies shifts and a reduction in net:gross at a ï£ value of ~0.7. A statistical analysis of the field data yields another quantitative measure of downstream facies shifts associated with mass extraction, as well as a measure of stratigraphic order. The runs metrics c and f are employed to describe thickness and grain size trends within vertical sequences of borehole data. A Monte Carlo simulation estimates the probability that a given succession of strata could be generated by chance. The results of both the statistical analyses and Monte Carlo simulations indicate spatial patterns in order and completeness, whereby areas associated with enhanced mass extraction exhibit more stratigraphic order. These findings are consistent with experimental and field results from a variety of depositional settings that demonstrate the utility of a scale-independent mass balance framework for the prediction of downstream facies changes.

I have conducted a modern precipitation and cave drip water analysis in Northern California and developed precisely dated, high-resolution records of stable oxygen (δ18O) and carbon (δ13C) isotope variability from a Lake Shasta Caverns speleothem (LSC3) from Lakehead, California (40.8043°N, 122.3040°W). Analysis of six years of weekly rainfall isotope data (2010-2016) indicates that temperature and moisture source are significant controls on the δ18O and δ2H signature of precipitation that falls in Northern California. U-Th dating of the LSC3 stalagmite reveals that it precipitated from ~35.9 to ~14.3 ka, spanning Heinrich Stadials 1-3, and the last deglaciation, a period when regional records suggest large hydroclimatic changes in western North America (WNA). Increased δ18O and δ13C in LSC3 during the last glacial period (Marine Isotope Stage 2) including the Last Glacial Maximum suggest an increase in subtropical moisture but overall dry conditions in Northern California. This interpretation is supported by slower stalagmite growth rates during the glacial relative to the deglaciation and Marine Isotope Stage 3. Wetter conditions during the BA are suggested by decreased δ13C values and faster stalagmite growth rates. Comparison of the LSC3 record with coeval paleoclimate records from WNA will assist in refining our understanding of regional hydroclimatic variability and will help to determine the stationarity of spatial patterns of change between the north and south across the western United States. Understanding the driving mechanisms of California precipitation variability is critical in this heavily-populated, water stressed region.

We perform Ediacaran-style decay experiments on both sea anemones (Condylactis gigantea) and sea hares (Dolabella auricularia) in order to test between several proposed models of control on âdeath maskâ style preservation. Specifically, we test comparisons of rates of decay, patterns of decay, and better quantify the black âdecay haloâ precipitate noted in previous work. The results demonstrate a necessary threshold of dermal rigidity for âdeath maskâ style preservation, as well as potentially help test between diploblastic vs. triploblastic affinities for Ediacaran organisms. We find that while rate of decay was decelerated under Ediacaran-style conditions, pattern of character loss in both diploblastic and triploblastic organisms was conserved from the baseline experiments. Geochemical analyses of sediment from around the carcasses indicate that previous models of âdeath maskâ preservation might need to be assessed further as precipitation of FeS or clay mineral precursors may be decay-rate-dependent or affected by overall tissue lability of the carcass. Patterns of character loss associated with both the anemones and mollusks demonstrate a bias against preservation of the most labile tissues, such as anemone tentacles and internal gut system of the mollusks. Such biases have likely affected previous interpretations and controversies of the affinities of some iconic Ediacaran and Cambrian organisms. According to our results, absence of evidence for some of these features might be more closely linked with dermal lability rather than absence of feature all together.

The ~1000 km3 Peach Spring Tuff (PST) is the product of an 18.8 Ma supereruption from Silver Creek caldera, southern Black Mountains, AZ (Ferguson et al 2013). Five zones are recognized in thick outflow sections of the ignimbrite near Kingman, AZ, based on welding characteristics and phenocryst, pumice, and lithics abundances. The lower four zones are uniformly high-silica rhyolite (HSR) (74-76 wt% SiO2; Pamukcu et al 2013; Barry et al 2015). Phenocryst content increases gradually upward from ~2% in the lowest zone to ~20% modally in the fourth zone (Ferguson & Cook, 2015), and then rapidly upward into the capping fifth zone, a densely welded, phenocryst-rich (~35%) trachyte (65-69 wt% SiO2) whose fiamme match those in intracaldera PST in texture and composition (Pamukcu et al 2013). Zone 5 fiamme are characterized by high Ba, Sr and Zr concentrations, compared to the highly evolved chemical signature of the HSR outflow. Phenocrysts in intracaldera and outflow trachytic pumice are large and exhibit evidence of resorption and reaction. Glasses in the trachyte are extremely enriched in elements consistent with the partial dissolution of the phenocryst assemblage. Previous elemental and isotopic data, plus new elemental data, support a model of the PST reservoir as a single, relatively simple, vertically stratified chamber with a crystal-rich base and massive, high-silica, crystal-poor upper zone (Frazier, 2013; McDowell et al 2016). Outflow and intracaldera trachyte represent the crystal-rich cumulate base proposed to have been partially remelted and remobilized, via injection of a hot mafic magma.

The capability to predict the final dimensions of an ‘a’a flow-field and the timeline for emplacement is key to effective lava hazard management. Levee breaching and the subsequent generation of secondary flows has been recognised as an intrinsic process in flow-field development. However, the conditions, locations and mechanisms for levee failure are not fully understood and so a multidisciplinary approach was taken utilising field studies of extant flow-fields on Mount Etna and Tenerife, analogue modelling and uniaxial compression tests to investigate levee failure. The experimental results were used to define three stages of flow-field development – lengthening, widening and thickening, a modification of the classification of Kilburn and Lopes (1991). Levee breaching was identified to be a key process by which flow-fields widen during stage two of emplacement. Four mechanisms for breaching are identified, providing a classification system for levee failure. Secondary flows originating from breaching consistently formed at an angle of 30 – 50o to the original channel, significantly contributing to overall flow-field width. Uniaxial compression tests were performed to quantify the mechanical properties of levees. Levee strength was shown to change with direction according to the size, orientation and distribution of vesicles. Samples with vesicle alignment orientated obliquely to the angle of principal stress were consistently weaker. A set of conditions, relating to the mechanical properties of the levee, is therefore identified under which levees are more susceptible to failure. The results of the study have been combined to produce a set of guidelines to the locations and conditions under which levee breaching is more likely to occur, aiding decision making for lava hazard management.