UC San Diego

Hydroclimatic research has found that existing reservoir storage could be rendered ineffective at maintaining water reliability under anticipated climate change scenarios. However, little research has evaluated expanding reservoir storage as a means of adaptation to climate change. In my first thesis chapter, I develop a reservoir model of urban water supply for the city of San Diego, California to assess the ability of existing storage and storage capacity expansions to meet urban water demand under present and projected future climatic scenarios. Climate variability is explicitly treated through probabilistic formulations. I find that the climate change scenarios will be more costly to the city than scenarios using historical hydrologic parameters. The magnitude of the costs and the specific optimal adaptive policy are sensitive to projected population growth and the degree to which the model can accurately predict spills. My second thesis chapter explores how adjustment costs vary as a function of the adaptation strategy that the agent chooses. Adjustment costs can occur from imperfect knowledge about the future climate and the time it takes to adjust the capital stock once learning has occurred. Few climate change adaptation studies have evaluated the trade-off between scale and flexibility when considering precautionary investments in the presence of incomplete information. I find that for urban water in San Diego, changing the timing and/or magnitude of additional reservoir storage investments can mitigate, but not eliminate, water shortages attributable to climate change and that the investment strategy of expanding capacity incrementally lowers adjustment costs relative to adding larger increments of storage less frequently. My third thesis chapter solves for the optimal level of reservoir storage, conservation investment, and price of water as a function of the costs of providing water and social gains from obtaining water. The peak-load public utility pricing literature does not accurately characterize water reliability because it does not consider the case when supply and demand are both stochastic and correlated. I find that reliability is higher if the central planner requires additional water conservation; reliability is lower if price is constrained below the optimal level; and that reliability is higher under inefficient rationing