As global populations grow and climate becomes more extreme, cities in drought prone regions of the world such as South East Australia are faced with escalating water scarcity and security challenges. Approaches geared towards addressing these challenges are diverse, and detecting “success” against background climate variability is difficult. Here we use 72 years of urban water consumption data from Melbourne, Australia, a city that recently overcame a 12 year “Millennium” drought, to 1) evaluate the utility of wavelet-based approaches for deconstructing climatic and anthropogenic drivers of urban water demand and 2) assess the relative contribution of various water saving strategies to overall demand reduction during the Millennium Drought. Our analysis points to conservation behavior as a dominant driver of municipal water savings (~69%), followed by non-revenue water reduction (~29%), and potable substitution with alternative sources like rain or recycled water (~3%). Demand exhibited both climatic and anthropogenic signatures, with temperature and rainfall impacting consumption significantly at all frequencies except 0.09-0.2 cpy. Residual patterns in demand (after removing climate effects) mapped to anthropogenic controls like outdoor water restrictions, which damped seasonal variability in consumption during the Millennium Drought, and changing technologies/social norms, which impacted consumption at sub annual frequencies, particularly from 1960-1990.

Despite major advances in water efficiency, urban water resources are increasingly strained due to population growth and severe droughts. In arid regions, water managers often promote water conservation rebate programs: the most common is known as a “turf rebate” program – offering participants a fixed rebate amount per unit area (e.g., $2 per square foot) in exchange for replacing turf grass with drought-tolerant landscaping. In southern California where outdoor watering can comprise more than half of residential water use, hundreds of millions of dollars worth of turf rebates were distributed during the 2011-2016 drought.

In my thesis, I focus on single-family residence (SFR) participation in a turf rebate program in the Irvine Ranch Water District (IRWD) in southern California. Spatial analyses at the village (i.e. neighborhood) level and at the parcel level reveal that drivers of participation are tied heavily to aspects of the built environment, whether it’s home age (at the village level) or outdoor area (at the parcel level). From these results we propose a probabilistic framework for estimating, and potentially optimizing, the water savings achieved by cash-for-grass programs, taking into account constraints of outdoor area imposed by the built environment, climate, demographic drivers of conservation behavior, and financial incentives.

I also explore temporal patterns of participation and find that a universal temporal pattern captures the majority of the monthly participation probability across all villages (i.e. neighborhoods) in the IRWD service area. This temporal pattern is highly correlated with Google search rates for the phrase “turf rebate”, which serves as a proxy for mass media coverage on the California drought and the turf rebate programs that became more heavily advertised as a result. In conjunction with California Governor Jerry Brown’s 2014 emergency drought proclamations, IRWD conducted water conservation and education programs. These local and state actions led to very high SFR participation in the IRWD turf rebate program.

Together, these built environment, demographic, and mass media coverage variables help explain water savings potential and spatiotemporal patterns of participation in IRWD’s turf rebate program.

Bioavailable forms of nitrogen, such as nitrate, are necessary for aquatic ecosystem productivity. Excess nitrate in aquatic systems, however, can adversely affect ecosystems and degrade both surface water and groundwater. Some of this excess nitrate can be removed in the sediments that line the bottom of rivers and coastal waters, through the exchange of water between surface water and groundwater (known as hyporheic exchange).

Several process-based models have been proposed for estimating nitrate removal in aquatic systems but these (1) do not consider the multiscale nature of hyporheic exchange flows; (2) rely on simplified conceptualizations of mixing within streambed sediments (e.g., a well-mixed box); (3) neglect important steps in the N-cycle (e.g., nitrification and ammonification); and/or (4) adopt pseudo-first-order kinetic descriptions of denitrification. On the other hand, a number of empirical correlations have been published based on in-stream measurements of nitrate uptake using reach-scale stable isotope tracer experiments. While these correlations are noteworthy in many respects, they do not account for physical processes known to play an important role in nutrient processing, such as the exchange of water between a stream, sediments, and groundwater.

In this thesis, I develop and test a simple and scalable process-based model for estimating the nitrate uptake velocity that addresses the limitations identified above. In particular, my model accounts for: (1) hyporheic exchange at multiple scales together with ambient groundwater flow; (2) the broad residence time distributions characteristic of hyporheic exchange; (3) key biogeochemical reactions associated with N-cycling, including respiration, ammonification, nitrification, and denitrification; and (4) the nonlinear nature of the pertinent biogeochemical reaction rates, including Monod kinetics for aerobic respiration and denitrification, and second-order kinetics for nitrification. Using this modeling framework I systematically evaluate primary controls on stream N-cycling and evaluate how multi-scale and regional factors are likely to affect this process. I also demonstrate how my model predictions compare with previously published reach-scale measurements of nitrate removal, develop scaling relationships by which my process-based model can be applicable to larger scales (i.e., watershed and regional scales), and provide some mechanistic explanations for previous observations.

The advent of molecular-based diagnostics has radically improved our ability to analyze the prevalence and activities of microorganisms in environmental systems. In this research. a new molecular tool (gene probe spectroscopy or GPS) was tested which may further enhance the capabilities of existing molecular techniques for carrying-out environmental analysis. With GPS, a microorganism (or a group or related microorganisms) is characterized by its hybridization to a set of non-specific oligonucleotide probes in a dot-blot format. The hybridization intensities for each probe/microorganism combination is determined experimentally and stored in a matrix database (A). Samples containing unknown mixtures of microorganisms are then hybridized with the same set of probes to obtain a set of composite hybridization intensities (b). By employing linear inverse theory, the concentration (c) of each microorganism present in the unknown sample may be computed by matrix inversion techniques: C = A -lb. With an award from the Water Resources Center, we investigated this approach using a model system consisting of four different bacteriophage (T2, M13, lambda, and 1'7) and six non-specific oligonucleotide probes each of which cross-reacts with two or more phage. In a series of tests, phage concentrations were correctly predicted by the method. illustrating its efficacy for identifying and enumerating microorganisms in fluid samples. Its application to "real" environments will require the use of DNA chip technology in which oligonucleotide probes are arrayed on glass slides. Future work will be directed toward the merging of the techniques developed in this project, namely GPS, with the DNA chip format,

As water imports dwindle, and climate changes, southern California is progressively turning to non-traditional water resources, including water reclamation, ocean water desalination, and the capture and fit-for-purpose use of dry and wet weather urban runoff. All three approaches are being employed to various degrees in the region, but capture and use of urban runoff has the potential to not only increase water supply but also provide habitat for rare and endangered flora and fauna, pollutant removal services, urban greening, and mitigation of the urban heat island effect.

This thesis explores both the water supply and water quality challenges and opportunities associated with using green stormwater infrastructure (GSI) to capture and treat urban runoff in Southern California. On the supply side, Chapter 2 presents an analysis of streamflow trends across the region, with the goal of assessing how reliable this source of water is likely to be into the future, given current trajectories in water conservation and climate change. The results suggest that, generally speaking, streamflow was increasing across the region until around 1990, and has been steadily falling since. In most of the 37 streams analyzed here, current summertime flows are less than 50% of 1990 levels. This precipitous decline reflects a combination of human and hydrological factors, including reduced water imports to the region and outdoor water conservation measures initiated during, and following, the drought of 2011-2016.

On the water quality side, two complementary mathematical fate and transport frameworks are presented for assessing, and predicting, the removal of pollutants in a form of GSI called bioretention systems. The two frameworks take distinctly different approaches for addressing a primary challenge for modeling pollutant removal in these systems; namely, the intermittent nature of storms implies that flow through these systems is inherently unsteady.

The first mathematical framework (Chapter 3) hypothesizes that, by flow weighting time, pollutant breakthrough in these systems can be represented by a standard one-dimensional advective-dispersion model for pollutant transport through porous media. This hypothesis is validated using bromide breakthrough data collected during a pilot-scale bioretention experiment at Orange County Public Works (OCPW). This experiment, which was conducted in Spring of 2019, explicitly accounts for the highly variable (flashy) storm flows generated from impervious areas in Southern California.

The second mathematical framework (Chapter 4) hypothesizes that transient transit-time distribution theory (T-TTD) can be coupled to a biokinetic model of nitrogen cycling, to estimate the fate and transport of nitrogen species (ammonium and nitrate) in the same pilot-scale bioretention system described above. A particular advantage of T-TTD theory is that it can account for loss of water by both gravitational drainage and evapotranspiration--both of which can exert significant controls on nitrogen cycling during the long antecedent dry periods typical of Southern California's summers. The breakthrough of both ammonium and nitrate in OCPW's pilot scale facility can be reproduced, after accounting for adsorption and nitrification (ammonium) and nitrification (nitrate) in the soil media component of this system.

By shedding light on both the supply and water quality dimensions of urban stormwater runoff, this thesis supports Southern California's long-term goal of reducing its reliance on imported sources of water. It also informs adaptive water management approaches under changing conditions and guides bioretention design enhancements for improved stormwater quality.