My dissertation research poses two overarching questions. First, how do climate change, land use change, and population growth compare and interact as drivers of change to future water demands and supply in California's Central Valley? Second, how can risk be integrated into such assessments of climate impacts?
To address the first of these questions, I built and calibrated an integrated hydrology and water operations model to simulate the historical water system operations in the Stanislaus, Tuolumne, and Merced river basins in California's Central Valley. I then drove this model over the course of the century using simulations of climate change, population growth, land use change and water use efficiency to compare the effects on water demands and water supply reliability.
Model results indicate that in the rapidly urbanizing study area, with projected low-density growth displacing farmland, the impacts of population and urbanization on water demands are greater than that of climate change alone. The net effect throughout the study area is decreasing water demands, driven by removal of acreage from agricultural production. Although climate change considered alone results in decreasing water supply reliability, population growth and land use change mute the effect.
To address the second question with risk analysis, I developed a method for quantifying risk preferences of water managers, using the economic concepts of risk aversion (the desire to avoid and manage risks) and loss aversion (a tendency for people to strongly prefer avoiding losses to acquiring gains). I applied the method in interviews with managers responsible for water supply to irrigation districts in the study area. My interviews revealed high levels of both risk aversion and loss aversion when it comes to their duties in water provision for agricultural customers.
I then combined the risk preferences with output from the climate-driven hydrology modeling to estimate expected utility under climate change. Model results for water supply under climate change give lower expected utility for managers than when assuming historical conditions, indicating that impacts of climate change will be negative for the water sector in this region regardless of the degree of managers' risk aversion. However, the expected utility for decision makers is strongly influenced by their risk preferences, and these risk preferences are stronger determinants of results for expected utility than are climate conditions.
The results highlight the importance of considering land use as a driver of water system change, especially when invoking population growth as a driver of change. They also show the limitations of climate impacts assessments that do not incorporate other major stressors, complementing previous path-breaking global-scale efforts and highlighting the importance of place-specific, spatially explicit analyses.
The analysis reported in this dissertation supports the notion that managers' risk preferences may be underutilized in impacts assessment, and in particular that ignoring them may understate estimates of climate change impacts. If results using this and other methods on water managers stand to scrutiny and repeated application, and particularly if variants produce congruent results in other sectors of resource management, they may also indicate that loss aversion on the local scale could drive the use of risk analysis in global scale integrated assessment.