Date of Award

Fall 9-2015

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Civil and Environmental Engineering


David A. Dzombak

Second Advisor

Mitchell J. Small


In 2015 California entered its fourth year of record-breaking drought which led many to question the sustainability of future water supply in the state, especially in Southern California. With over 90% of its water imported from outside the city, Los Angeles has been greatly impacted by this instability. Due to climate change, population growth, groundwater contamination, and competing demands, Los Angeles’ water sources (Los Angeles Aqueduct, Colorado River Aqueduct, California Aqueduct, local groundwater, and reclaimed water) are subject to various stressors, which make projections of water supply and planning for system sustainability challenging. Reductions in availability from each water source will influence the price and availability of residential water and inevitably lead to greater need for conservation, and for the development of new sources of water supply. Understanding the future Los Angeles water system in the context of a growing population and climate change thus merits careful investigation. The overall objective of this work was to evaluate the factors that impact water demand in Los Angeles as well as apply and compare various modeling techniques to forecast water demand. This was accomplished by (1) assessing the sustainability of each of the water sources that supply Los Angeles under present and future conditions using a system characterization and analysis. (2) Analyzing the importance of various factors influencing water demand in Los Angeles using multiple linear regression and artificial neural network models. (3) Projecting water demand in Los Angeles until 2050 under various scenarios of price, precipitation, temperature, conservation and population. (4) Developing an agent based model of Los Angeles iii water demand that provides insight into interactions among consumers and the Los Angeles Department of Water and Power (LADWP) water management system. First, the sustainability of each of the water sources that supply Los Angeles under present and future conditions was analyzed using a system characterization and analysis. The results of the study showed that of the five main water sources that supply Los Angeles, a majority will be impacted by climate change, water quality, energy, and cost stressors. While the expansion of water demand management and agricultural water transfers can help address the challenge of increasing demand, the impacts of climate variability and competing demands are likely to limit their potential. Next, multiple linear regression (MLR) and artificial neural network (ANN) models were used to evaluate historical (1970-2012) water demand data for Los Angeles to assess the importance of water price, population, conservation methods, temperature, and precipitation on influencing water demand. Results indicated that the multiple linear regression models were comparable to the artificial neural network models in ability to describe historical water demand data. Models developed for and fitted to monthly data were more accurate in estimating water demand compared to models based on yearly data. Temperature, precipitation, conservation, population, and water price were all significant in impacting monthly water use in Los Angeles. Additionally, fitting of the data with the MLR models revealed that price and conservation impacts have significantly counteracted the impact of population growth on water demand. iv After the variables significant in impacting water demand were identified through the modeling of the historical water demand data, MLR modeling with the same variables, water price, population, temperature, precipitation, and conservation, were used to project Los Angeles water demand until 2050. The model used projections of four global climate models with two CO2 emission scenarios, as well as high, medium, and low scenarios of population, water pricing, and conservation to generate an envelope of forecasts of water demand in Los Angeles from 2013 to 2050. Results of the forecasting with the MLR models under the various scenarios indicated that the effects of climate change on water demand (not supply) are projected to be modest. Without the introduction of increased water pricing and conservation methods, water demand in Los Angeles has the potential to nearly double from 130 billion gallons per year in 2013 to 250 billion gallons in 2050. However, more likely scenarios of population growth, conservation implementation, climate change, and pricing structures yielded predicted increases in water demand of approximately 30 percent, to a level of 170 billion gallons per year by 2050. Finally, an agent based model for water demand was developed to understand the interactions between Los Angeles water consumers and management actions by the LADWP. The model was calibrated with historical (1970-2012) data for water demand. In the model, consumers react to changes in water supply variability and adjust their conservation levels. These changes in conservation help reduce the dependability of imported water in Los Angeles. Projections of various water variability scenarios showed that consumers respond to advertisement intensity and shifts in contact rates with other consumers, which allows for more system resiliency. This is evident in the present day drought in Los Angeles in which water demand is being reduced. Results of the agent based model also demonstrated that high variability in water supply can v increase water demand by up to 2.67 times from 2013 to 2050. Validation of the model indicates the agent based model can adequately project water demand in the future under changes in population, climate, and water supply reliability. Water management in Los Angeles needs to be understood as a highly integrated social, engineered infrastructure, economic, and ecological system. Previously, research efforts in Los Angeles have been focused on understanding the hydrological components of the problem without the consideration of social impacts and human behavior. Consideration of the human decision making and adaptive responses involved is important because it affects water resources management and planning. Adaptive agent modeling can provide insights into how consumer-manager-supplier interactions impact overall water system performance and evolution. The projection models proposed in this research offer a means of assessing potential impacts of water conservation initiatives, restrictions on water demand, pricing incentives, and other water system management options. They also provide insight into the relative importance of environmental conditions that affect water supply on water demand. By viewing the Los Angeles water supply system as a complex system, identification of areas in which there are problems as well as possible solutions can be evident. The city of Los Angeles is still growing and dependency on transferred water is inevitable, so understanding the complexity of the system is imperative to develop sustainable solutions to water scarcity and reliability.