With growing urban populations and increasing concerns over the effects of climate change on water supplies, there has recently been a significant amount of interdisciplinary research focused on identifying the drivers of urban water use. Due to unavailability of individual or household level data, these studies are often limited to using spatially aggregated data. There is concern that this aggregation of data may be leading to misrepresentations of the drivers of urban water use, yet there have been few studies that have addressed this concern. As in all spatial quantitative analyses, studies in this area should consider how the spatial scales chosen for analysis are affecting the results. The purpose of this research is to use a case study of single-family residential (SFR) water use in Portland, Oregon to determine the extent to which scale variation significantly affects the patterns of SFR water use, and whether household scale water use is influenced by neighborhood and census tract characteristics. The results of this analysis provide evidence that aggregating household scale water use data can mask meaningful patterns in SFR water use and potentially provide misleading information on what is influencing water use habits. This research also shows that using the chosen exploratory variables, there is a statistically significant, but not substantial, cross-scale influence on household scale water use by neighborhood and census tract characteristics.

This dissertation presents a method for enhancement of the efficiency and scalability of photocatalytic water purification systems, along with an experimental validation of the concept. A 3-dimensional photocatalyst structure, made from a TiO2-SiO2 composite, has been designed and fabricated for use in a custom designed LED-source illumination chamber of rotational symmetry that corresponds with the symmetry of the photocatalyst material. The design of the photocatalyst material has two defining characteristics: geometrical form and material composition. The design of the material was developed through the creation of a theoretical model for consideration of the system's photonic efficiency. Fabrication of the material was accomplished using a Ti alkoxide solution to coat a novel 3D support structure. The coatings were then heat treated to form a semiconducting thin-film. The resulting films were evaluated by SEM, TEM, UV-vis spectroscopy and Raman spectroscopy. The surface of the material was then modified by implantation of TiO2 and SiO2 nanoparticles in order to increase catalytic surface area and improve the photoactivity of the material, resulting in increased degradation performance by more than 500%. Finally, the efficiency of the photocatalytic reactor was considered with respect to energy usage as defined by the Electrical Energy per Order (EEO) characterization model. The effects of catalyst surface modification and UV-illumination intensity on the EEO value were measured and analyzed. The result of the modifications was an 81.9% reduction in energy usage. The lowest EEO achieved was 54 kWh per cubic meter of water for each order of magnitude reduction in pollutant concentration -- an improvement in EEO over previously reported thin-film based photoreactors.

Waterborne disease is a significant contributor to the global burden of disease, in particular among high-risk populations in developing nations. State-of-the-art methods for the enumeration of microbial pathogens in drinking water sources have important limitations, including high initial cost, 24-48 hour delays in results, high staffing and facility requirements, and training requirements which all become especially problematic in the developing nation context. A number of alternative approaches to microbial water quality testing have been proposed, with the goal of decreasing the required testing time, decreasing overall costs, leveraging appropriate technology approaches, or improving sensitivity or specificity of the water quality testing method. One approach that may offer solutions to some of these limitations involves the deployment of sensor networks using fluorescent spectroscopy to detect intrinsic protein fluorescence in water samples as a proxy for microbial activity. In recent years, a number of researchers have found significant and meaningful correlations between indicator bacteria species and the protein fluorescence of drinking water samples. Additionally, advances in the semiconductor industry could be used to drive down the cost of such sensors. This technology may also be extensible to other water quality parameters, including dissolved organic matter or the presence of fluorescent pollutants. In this thesis, a literature review describes the fundamentals of fluorescence spectroscopy, historical and recent work regarding the fluorescence of the amino acid tryptophan and associated bacterial fluorescence, possible mechanisms for this association, and potential applications of this technology for drinking water quality monitoring and waste water process control. Extensibility of the technology is also discussed. Next, experimental methodology in reproduction of similar results is described. Samples were taken from seven (7) surface water sources and tested using membrane filtration and an off-the-shelf fluorescence spectrometer to help examine the association between the presence of indicator bacteria and the tryptophan fluorescence of the water sample. The results, showing an association of R2 = 0.560, are compared to the results of recent similar experiments. Finally, two prototypes are described, including their design requirements and data from prototype testing. The results of the testing are briefly discussed, and next steps are outlined with the goal of developing a low-cost, in-situ microbial water quality sensor using fluorescence spectroscopy principles.

Nutria are semi-aquatic rodents, non-native to Oregon. They are an invasive species that damage stream banks with burrowing and cause destruction of native vegetation; activity that may cause deterioration of stream water quality. I hypothesized that my study's duration and pattern of nutria trapping along South Johnson Creek would be sufficient to lead to a significant change in turbidity, dissolved oxygen (DO), total dissolved solids (TDS), total solids (TS), and total suspended solids (TSS).

Polluted stormwater, if not treated, can compromise water quality throughout our hydrologic cycle, adversely affecting aquatic ecosystems. Common stormwater pollutants, copper and zinc, have been identified as primary toxicants in multiple freshwater and marine environments. For small-scale generators, stormwater management can be cumbersome and implementation of common BMPs impractical thus catch basins are popular though not the most environmentally conscious and sustainable option. This study aims to characterize the potential of a mobile media filter operation for the treatment and on-site recycling of catch basin stormwater. The removal capacities of various commercially available filter media (e.g. a common perlite; Earthlite™, a medium largely composed of biochars; and Filter33™, a proprietary porous medium) were measured using binary injection solutions modeled after local catch basin stormwater characteristics. The results of filtration experiments, rapid small-scale column tests (RSSCTs), indicate that the transport of metals in Perlite is primarily impacted by nonspecific sorption whereas in Earthlite™ and Filter33™ both nonspecific and specific sorption are present. For all media and experimentation, there was a consistent preferential uptake of copper such that copper displayed delayed arrival and/or greater removal than zinc. Moreover, the observed snow plow effects and concentration plateaus in Earthlite™ and Filter33™ RSSCTs suggest rate limited ion exchange and specific sorption in addition to ion competition. Earthlite™ exhibited an approach velocity dependent removal efficiency in the RSSCTs and pseudo second order uptake behavior for zinc in kinetic batch experiments. At the lab scale equivalent of the proposed field scale flow rate, Filter33™ displayed the greatest average zinc removal of 8.6 mg/g. In all, this research indicates that test parameters (i.e. pH, competitive ions solutions, empty bed contact time, flow rate) based on the natural environment and field scale operation can greatly impact removal efficiency in filter media.

Protection of water resources in a changing climate depends on bottom-up stewardship and adaptive management. From the ground up, a vital component is maintaining soil ecosystem services that regulate water, recycle nutrients, sequester carbon, provide food, and other benefits. Interacting spatial, social, and physical factors determine agricultural and stormwater management, and their impact on water. This dissertation explores these dimensions within a complex social-ecological system. The first chapter evaluates a participatory process to elicit solutions to complex environmental problems across science, policy, and practice. The second chapter studies on-farm soil assessment and its role in informing management decisions and supporting adaptive capacity. The third chapter investigates cross-scale dynamics of residential green stormwater infrastructure (GSI) for improved water resource management in a broader social-ecological context. Integrating participant feedback into current science, research, and decision-making processes is an important challenge. A novel approach that combines a Delphi method with contemporary “crowdsourcing” to address water pollution in Lake Champlain Basin in the context of climate change is presented. Fifty-three participants proposed and commented on adaptive solutions in an online Delphi that occurred over a six-week period during the Spring of 2014. In a follow-up Multi-Stakeholder workshop, thirty-eight stakeholders participated in refining and synthesizing the forum’s results. The stakeholders’ interventions from the crowdsourcing forum have contributed to the current policy dialogue in Vermont to address phosphorus loading to Lake Champlain. This stakeholder approach strengthens traditional modeling scenario development to include priorities that have been collectively refined and vetted. Healthy agricultural soils cannot easily be prescribed to farms and require knowledge and a long-term commitment to a holistic and adaptive approach. The second chapter addresses the questions: “to what extent do farmers use indicators of soil health, and does feedback inform management decisions?” A survey of farmers in two Vermont watersheds was conducted in 2016 showed relatively high use of fourteen soil indicators and high rankings of their importance. The finding that there were differences in use and perceived importance of soil indicators across management and land-use types has implications beyond the farm scale for agriculture, and the provision of ecosystem services. Soil management relates to broader adaptation strategies including resistance, resilience, and transformation that affects adaptive capacity of agroecosystems. Bottom-up adoption of environmental behaviors, such as implementing residential GSI, need to be understood in the context of the broader social-ecological landscape to understand implications for improved water management. A statewide survey of Vermont residents paired a cross-scale and spatial analysis to evaluate how intention to adopt three different GSI practices (infiltration trenches, diversion of roof runoff, and rain gardens) varies with barriers to adoption and household attributes across varying stormwater contexts from the household to watershed scale. Improved stormwater management outcomes at the watershed and local levels depend on management strategies that can be implemented and adapted along the rural-urban gradient, across the bio-physical landscape, and according to varying norms and institutional arrangements.

Silage bunker runoff is a form of agricultural pollution that contributes to aquatic ecosystem degradation. Current handling and treatment methods for this process wastewater are often ineffective or expensive. A woodchip bioreactor is an emerging treatment technology designed to facilitate denitrification through the provision of an anaerobic, carbon rich environment. A wood chip bioreactor treatment system, consisting of three pre-treatment tanks, two wood chip bioreactors, and one infiltration basin, was constructed at the Miller Research Complex in South Burlington, Vermont in 2016. Runoff and leachate from an adjacent silage storage bunker is directed into the system. The pre-treatment tanks include two settling tanks and one aeration tank. The former allows for sedimentation of organic matter, while the latter is designed to allow for nitrogen transformations that will help maximize nitrogen removal in the bioreactors. During the summer and fall of 2017, sampling occurred at four points within the system in order to determine the efficacy of various treatment steps. Samples were analyzed for nitrate (NOx—N), ammonium (NH4+-N), total nitrogen (TN), soluble reactive phosphorus (SRP), and total phosphorus (TP) in order to compare inflow and outflow pollutant concentrations and loads. Results indicate that this treatment system significantly reduced nutrient loads in the runoff. Over the entirety of the sampling period, the influent TN and TP mass load were both reduced by approximately 44%.

Elevated nutrient concentrations present significant challenges to surface water quality management globally, and dissolved organic matter mediates several key biogeochemical processes. Storm events often dominate riverine loads of nitrate, phosphorus, and dissolved organic matter, and are expected to increase in frequency and intensity in many regions due to climate change. The recent development of in situ optical sensors has revolutionized water quality monitoring and has highlighted the important role storms play in water quality. This dissertation focuses on improving the application of in situ optical water quality sensors and interpreting the high-frequency data they produce to better understand biogeochemical and watershed processes that are critical for resource management. We deployed in situ sensors to monitor water quality in three watersheds with contrasting land use / land cover, including agricultural, urban, and forested landscapes. The sensors measured absorbance of ultraviolet-visible light through the water column at 2.5 nanometer wavelength increments at 15-minute intervals for three years. These deployments provided a testbed to evaluate the sensors and improve models to predict concentrations of nitrate, three phosphorus fractions, and dissolved organic carbon using absorbance spectra and laboratory analyses through multivariate statistical techniques. In addition, an improved hysteresis calculation method was used to determine short-timescale storm dynamics for several parameters during 220 storm events. Goals of each dissertation chapter were to: (1) examine the influences of seasonality, storm size, and dominant land use / land cover on storm dissolved organic carbon and nitrate hysteresis and loads; (2) evaluate the utility of the sensors to determine total, dissolved, and soluble reactive phosphorus concentrations in streams draining different land use / land covers, and perform the first statistically robust validation technique applied to optical water quality sensor calibration models; and (3) analyze storm event dissolved organic matter quantity and character dynamics by calculating hysteresis indices for DOC concentration and spectral slope ratio, and develop a novel analytical framework that leverages these high frequency measurements to infer biogeochemical and watershed processes. Each chapter includes key lessons and future recommendations for using in situ optical sensors to monitor water quality.

The diffuse runoff of agricultural nutrients, also called agricultural nonpoint source pollution (NPS), is a widespread threat to freshwater resources. Despite decades of research into the processes of eutrophication and agricultural nutrient management, social, economic, and political barriers have slowed progress towards improving water quality. A critical challenge to managing agricultural NPS pollution is motivating landowners to act against their individual farm production incentives in response to distant ecological impacts. The complexity of governing the social-ecological system requires improved understanding of how policy shapes farmer behavior to improve the state of water quality. This dissertation contributes both theoretically and empirically to NPS pollution governance by examining the impacts of water quality policy design on farmer nutrient management decision making and behavior. In the first study, I theoretically contextualize the issue of agricultural NPS pollution in the broader discussion of environmental public goods dilemmas to suggest that an increased focus on the link between policy and behavior can improve sustainable resource management. I propose two empirical approaches to study the policy-behavior link in environmental public goods dilemmas: 1) explicit incorporation of social psychological and behavioral variables and 2) utilization of actor mental models, or perceptions of the world that guide decision making, to identify behavioral drivers and outcomes. In the second and third studies, I then use these approaches to examine how water quality policies for agricultural NPS collectively change farmer behavior to reduce nutrient emissions. The second chapter uses a quantitative, survey-based approach to examine the relationship between mandatory policy design and behavior change in New Zealand. I find that a shift to mandatory policy is not immediately associated with increased adoption of nutrient management practices, but the mandatory policy design is important for potential future behavior change and long-term policy support. In the third study, I combine qualitative methodology with network analysis of qualitative data to examine a spectrum of agricultural NPS pollution policies in Vermont, USA and Taupo and Rotorua, New Zealand. I use farmer mental models to examine behavior change within each of the regions, the perceived drivers of behavior change and perceived outcomes of the policy. In this study, farmers across all three regions cite mandatory water policy as a key behavioral driver, but in each region, policy design interacts with the social-ecological context to produce distinct patterns of behaviors and perceived outcomes. Taken together, this dissertation demonstrates that agricultural NPS pollution policy design must consider the interactions between policy and other social-ecological behavioral drivers in order to achieve long term water quality improvements.

Areas of California's Central Valley are sinking at rates up to 1 foot per year due to subsidence caused, in part, by the state's years-long drought, challenging growers to locate additional water sources for their crops. Supply and demand theory guided this correlational study. The purpose of the study was to examine the financial impact of drought on almond demand. This study included annualized historical almond industry data for the United States (N = 97), downloaded from a United States Department of Agriculture database. The results of multiple linear regression analysis indicated that the model was capable of predicting almond demand, F(3,92) = 483.579, p < .001, R2 = .940. Both supply and price were statistically significant in the final model, with supply (p < .001) accounting for a higher contribution to the model than price (p = .015). Fine effect's contribution (p = .267) to the model was not statistically significant. The results of this study could enable almond industry leaders to increase profit margins through market predictability understanding and mitigate fiscal risks associated with variable labor and groundwater pumping costs. The implications for positive social change include the potential to restore employment opportunities, stabilize migratory worker prospects, and reduce water utilization to preserve natural resources.