Hydrologic sensitivities of western U.S. rivers to climate change

Abstract

As the climate continues to change, increasing temperatures and changes in precipitation will lead to fundamental changes in the seasonal distribution of streamflow, especially in the western United States where snowmelt plays a key role. These changes will inevitably lead to challenges for water resource managers. There is, however, considerable uncertainty as to the character of these hydrologic changes, especially at local and regional scales (10^2 - 10^5 km^2). My research aims to better understand how climate influences hydrologic processes, with a particular focus on variations in runoff sensitivities to changes in precipitation and temperature, and the use of this information in water management. Using land surface model simulations, I explore the sensitivity of runoff to changes in precipitation (defined as precipitation elasticities, E, the fractional change in runoff divided by the fractional change in precipitation), changes in temperature (defined as temperature sensitivities, S, percent change in runoff per degree change in temperature) and to the combined effect of temperature and precipitation changes. The character of these sensitivities varies considerably depending on how the land surface is simulated (e.g., type of land surface model), the particulars of the location (e.g., elevation, vegetation, soil types), and the season in which changes in temperature and precipitation occur. I explore these variations through hydrologic model experiments in the Colorado and Columbia River basins - two basins which can be considered end points of hydroclimatic variability in the West, and which also have diverse management concerns as existing reservoir storage in these systems varies strongly. The total storage relative to annual inflow ratio of over four in the Colorado River, results in a management focus on total (annual) magnitudes in streamflow, whereas this ratio is about 0.3 in the Columbia River and hence changes in the seasonal distribution of streamflow is the primary driver there. Within this body of work, I use the nature of these hydrologic sensitivities (e.g., spatial and temporal variability, superposition, and the linearity of their underlying functions) to develop two complementary methodologies that can be applied to generate viable first-order estimates of future change for long-term (e.g., 30-year) annual change (applied in the Colorado River basin) and seasonal change (applied in the Pacific Northwest). My results show that these sensitivity-based estimation approaches to future change compare well with the more common, computationally intensive full-simulation approaches that force a hydrologic model with downscaled future climate scenarios. These methods can be applied to newly released climate information to easily assess underlying drivers of change and to bound, at least approximately, the range of future streamflow uncertainties for water resource planners.