This project attempts to identify the source of surface water and groundwater salinity observed in a polder in southwest Bangladesh through geochemical analyses. All surface water samples are mixtures of tidal channel water and meteoric water concentrated by evaporation; B, K, Mg, Na, Sr, Cl, Br, and S behave conservatively. Surface water types include rice paddies, freshwater ponds, shrimp ponds, and tidal channels. Thirty percent of rice paddy samples exceed the upper salinity limit for rice crops of 1.56 parts per thousand (ppt). Although compositionally similar to rice paddy water, field observations show that freshwater ponds are sourced from wet season tidal channels and then concentrated by evaporation. Evidence indicates that salts in freshwater ponds were not added by inundation of brackish water following Cyclone Aila. Brine shrimp ponds are sourced from dry season tidal channels, which contain water that is significantly more saline than in the wet season. Groundwater is also identified as tidal channel water mixed with meteoric water, but is chemically unique from surface waters in that B and S do not behave conservatively, dissolved organic carbon concentrations are higher (except for shrimp ponds) and oxidation-reduction potential values (Eh in millivolts (mV)) are lower. Groundwater compositions show high spatial variability, suggesting low flow velocities. Groundwater conservative element concentrations are 30-50% higher in the dry season, suggesting limited recharge in the wet season. Nearly 50% of groundwater samples surpass the Bangladesh drinking water limit for arsenic of 50 micrograms per liter (µg/L). Sixty one percent of groundwater samples exceed the salinity limit of 3.25 ppt for safe consumption. Tidal channel water is the primary source of salinity in both surface water and groundwater.

The origins of gravel bearing sediments in the upper Holocene stratigraphy of the Brahmaputra-Jamuna River in northern Bangladesh are explored in this thesis through sediment geochemistry analyses and characterization of gravel types sampled from 41 boreholes along a 120 km transect (BNGA) that crosses two major paleo-alluvial valleys in the Ganges-Brahmaputra-Meghna Delta. The modern Brahmaputra River (BR) is primarily a sand bedload system even though flow velocities are sufficient to transport significant amounts of larger sediments (i.e. gravel). However, gravels are common in the upper Holocene stratigraphy. Through the use of strontium concentration (Sr ppm) as a provenance indicator for the sands associated with gravel samples, we confirm that the gravels in question belong to the Holocene aged paleo-alluvial valley fill deposits of the BR. This thesis presents evidence that the previously defined range of Sr ppm for BR sediments (>140 ppm) should be revised to include sediments with >120 ppm Sr. The lower tail of the BR Sr ppm range (120-140 ppm) is interpreted as new evidence of variability in sediment mixing budgets throughout the Holocene between the BR and tributary systems with lower known Sr ppm signatures (e.g. Tista River). However, no direct correlation is found between gravel bearing strata and any specific range of BR Sr ppm values. It is determined that the model for the origin of gravels in the upper Holocene stratigraphy of the BR relies on the availability of gravel bearing sediments to be eroded by the BR through lateral bank migration, not direct input from any one particular tributary system.

Due to the climate effects and aviation threats of volcanic eruptions, it is important to accurately locate ash in the atmosphere. This study aims to explore the accuracy and reliability of training a neural network to identify cases of ash using observations from the Moderate Resolution Imaging Spectroradiometer (MODIS). Satellite images were obtained for the following eruptions: Kasatochi, Aleutian Islands, 2008; Okmok, Aleutian Islands, 2008; Grímsvötn, northeastern Iceland, 2011; Chaitén, southern Chile, 2008; Puyehue-Cordón Caulle, central Chile, 2011; Sangeang Api, Indonesia, 2014; and Kelut, Indonesia, 2014. The Hybrid Single Particle Lagrangian Integrated Trajectory Model (HYSPLIT) was used to obtain ash concentrations for the same archived eruptions. A back-propagation neural network was then trained using brightness temperature differences as inputs obtained via the following band combinations: 12-11-μm, 11-8.6-μm, 11-7.3-μm, and 11-μm. Using the ash concentrations determined via HYSPLIT, a flag was created to differentiate between ash (1) and no ash (0) and used as output. When neural network output was compared to the test dataset, 93% of pixels containing ash were correctly identified and 7% were missed. Nearly 100% of pixels containing SO2-rich ash were correctly identified.

Efficient management of freshwater resources is critical as threats to water security increase due to changes in climate, population, and land use. The water and agricultural sectors of Asian countries have been identified as the most at risk for climate change impacts, and small-scale farmers are particularly vulnerable to changes in water availability and water quality. This research explored the tradeoffs between maximization of yield and minimization of environmental impact in rice production in Sri Lanka. The study used the DeNitrification-DeComposition (DNDC) model to examine how variations in climate, soil, and paddy management affect outputs of yield, greenhouse gas (GHG) emissions, and nitrogen (N) leaching in paddy systems in Sri Lanka from 1991 to 2010. Reducing fertilizer had little effect on yield, and N leaching and N2O emissions declined as a result. Reliable water inputs in irrigated systems increased yields over rain-fed schemes; alternate flooding techniques to mitigate water stress did not reduce yield. Alternate flooding increased N leaching as did rain-fed cultivation. A sensitivity analysis of soil parameters determined that high clay content reduced N leaching, low soil organic carbon increased yield, and the more basic the pH of the soil, the greater the reduction in GHG emissions. The results inform best practices for Sri Lankan farmers and decision makers on the supervision of water resources and agricultural inputs. This research demonstrates how cultivation in rice-growing regions in south Asia affects the environment and the nitrogen cycle on a global scale, in turn how informed management of these systems can adapt to a changing environment.

The Bishop Tuff, a 0.76 million-year-old rhyolitic deposit in California, contains two crystal phases, described by Pamukcu et al. (2012). The first contains long-lived crystal grown over 5000+ years; the second consists of microlites <50μm large, grown in <2 years as a result of eruptive decompression. In this study I combine vesicle size distribution (VSD) investigation with experimentally produced data on decompression crystallization to better understand the syn-eruptive evolution of a supereruption. Isothermal experiments were run by Jim Gardner using late-erupted Bishop Tuff material, which was water-saturated and melted, then decompressed at various rates and initial temperatures. Seven experiments have been analyzed; each began at 130MPa, and ended near 10MPa. All seven experiments had initial temperatures of 710⁰C or 785⁰C and the decompression rates were 20MPa/hr, 5.5MPa/hr, 1.7MPa/hr, or 0.75MPa/hr. I compared textures and mineral assemblages of Bishop Tuff and experimental products using Scanning Electron Microscope imaging. We were able to replicate growth of feldspar microlites and quartz rims in 6-3 day experiments. I used tomographic analysis of progressively larger pumice chips to obtain VSDs and calculated chip volumes and bulk density using image analysis. Plotting bulk porosity against volume generated a cumulative-like size distribution inconsistent with a true cumulative distribution. We demonstrate that this may result from censorship of large bubbles during sampling. We conclude that decompression crystallization took days to weeks and decompression rate fluctuated during eruption.

Soil samples were collected during three field campaigns to determine seasonal and spatial trends of soil salinity, soil acidity and arsenic concentrations on Polder 32 in coastal Bangladesh. Many farmers on Polder 32 use a crop rotation of rice cultivation in the wet season and shrimp farming in the dry season, and studies have shown that this rotation can increase soil salinity and acidity. Soil samples were collected from rice paddies and shrimp ponds on the polder, from adjacent tidal channels, and from the Sunderbans mangrove forest to the SE of the polder, and analyzed for both geochemical and physical parameters and then subjected to statistical tests and mapped using geographic information system software to find correlations. Results support the belief that soil salinity, acidity and arsenic concentration exhibit spatial variation, and soil salinity and acidity show seasonal variation. Rice grown in paddies should be unaffected by salt concentrations in the wet season, while arsenic concentrations in soil may be high enough to cause unsafe As levels in produced rice. Future studies should measure the As concentration in rice to determine the risk of bioavailable arsenic transmitted from soil in the area. No evidence of soil acidification was found, most likely due to the presence of soil carbonate.

Currently, Greenland ice sheet mass loss is increasing at a rate near 1000 Gt yr-1, making the ice sheets top contributors to current sea level rise. Errors in constraining the mass balance of the Antarctic and Greenland ice sheets are primarily from measurement of frozen precipitation. In this study, we use the Multi-Angle Snowflake Camera (MASC) to capture images of falling snow in Boulder, Colorado. The MASC contains three cameras positioned on adjacent edges of a decagon, which provides one of the first instruments capable of collecting three-dimensional information of frozen hydrometeors. We created autonomous image analysis software designed specifically to measure snowflake particles from image triplets recorded by the MASC. After processing snowflake measurements are organized and presented in particle size distributions, which show the expected number of particles with a given size. Particle size distributions for precipitation have previously been presented as exponential distributions that vary with precipitation rate (mass/time). We found evidence indicating the shape of the distribution to exhibit a dependence on the presence of aggregation. Our observations of non-aggregate particles follow a log-normal distribution, while aggregates adhere to an exponential distribution. Therefore, we propose snowfall size distributions to be a combination of log-normal and exponential size distributions depending on the relative fraction of aggregated versus non-aggregated particles.

The landscape of Middle Tennessee is characterized by relatively thick soils overlying Ordovician and Mississippian bedrock that is predominantly limestone with a high proportion of siliciclastic materialoften >20%that remains behind after dissolution as an insoluble residue contributing to soil formation. Zircon U-Pb geochronology, element mass fluxes, and other methods were used with soil-bedrock pairs collected from two different sites located atop high flat surfaces to test the hypothesis that soils within the region are derived solely from insoluble residue left behind by dissolution of the underlying bedrock. At Site 1, Mississippian chert of the Fort Payne Formation was collected along with the B1 and B2 horizons of the overlying ultisol. At Site 2, a composite sample of the A and B horizons of an alfisol and a sample of the underlying Ordovician limestone of the Hermitage Formation were collected. Zircon was recovered from the samples, imaged using cathodoluminescence, and analyzed using a 193 nm laser and quadrupole ICP-MS. Discordant analyses were discarded and 206Pb/238U ages are reported. Similarities in diagnostic mineral content and zircon U-Pb age spectra and predominantly negative element mass fluxes indicate that soils at both sites primarily formed by weathering of underlying bedrock, with a small exotic component that may account for ~25% of the ages. The presence of several Cenozoic zircon grains as young as 29.8 ± 0.9 Ma in the B2 soil from Site 1 require a loess component. Dissolved Bigby-Cannon limestone may be the source of exotic material at Site 2.

Factors such as lithology, tectonics and climate have been identified as exerting control on the form of stream channels and on the evolution of watersheds. The degree to which these factors interact to influence landscape erosion and its evolution has been an area of extensive research. Here, we use Digital Elevation Models to extract metrics of basin morphology to determine the form and the controls on river long profiles in the Harpeth River Watershed, which drains the Paleozoic carbonate bedrock of the Nashville Basin and the Western Highland Rim. We use the integral approach, which transforms the horizontal coordinate by integrating drainage area over flow distance to produce the transformed horizontal coordinate, χ, along with river long profiles to determine the relative influence of lithology on the form of the river profiles. The results show that drainage area, a proxy for water discharge, exerts the greatest influence on channel geometry in the Harpeth River Watershed as is indicated by m/n values that center around 0.5. Deviations from the linear, χ-transformed elevation profiles that are produced at m/n = 0.50±0.05 were found to be as a result of local differences in bedrock lithology that affect erodibility and sediment production.

The state of Meghalaya in NE India is a vibrant landscape where peoples livelihoods are intimately tied to the regional geology and seasonal climate. The climate of Meghalaya is defined by the Indian Summer Monsoon (ISM), which brings immense rainfall during the summer months. Uplift of the Shillong Massif, which dominates the topography of Meghalaya and contributes to orographic ISM precipitation in this region, is associated with historic mega earthquakes. However the long-term seismic history of the region is not well understood. Caves on the southern edge of the Massif, including Mawmluh Cave, provide unique opportunities to both improve understanding of the seismic history of NE India through uranium-thorium (U-Th) dating of tectonically broken speleothems and to study relationships between the ISM and other global climate indices through δ18O analysis of annually laminated speleothems. U-Th dating indicates that new speleothem growth on top of large broken speleothems in Mawmluh Cave initiated within 7 ± 2.2 years following the magnitude 8.7 1950 Assam-Tibet earthquake. These results suggest that dating of older breakdown structures could provide a stronger record of the seismic history of this region. Furthermore, a high-resolution modern speleothem δ18O record from Mawmluh Cave reveals ISM strength and frequency variations correlated with known changes in behavior of the El Niño Southern Oscillation over the past ~60 years, providing a new understanding of the influences on precipitation in NE India and a valuable calibration dataset for long-term speleothem paleoclimate records from this region.