PROJECTED EFFECTS OF EVAPOTRANSPIRATION AND CLIMATE CHANGE on WATER AVAILABILITY ACROSS THE CONTIGUOUS UNITED STATES

Evapotranspiration (ET) represents a major portion of the hydrologic cycle, and its magnitude is strongly correlated with air temperature (T). Thus an increase in global temperatures over the next century would lead to more ET and less water for base flow, ecosystems and human consumption. Many ET studies to date have focused on understanding field-scale physical processes. This has led to a focus on relating potential to actual ET and the problem of scaling up to the regional scales at which global climate models (GCMs) operate. The approach used in this study was to bypass the field scale, and instead use long-term mean watershed precipitation (P), stream flow (Q) and watershed areas (A) to calculate a proxy for actual ET. Mean P and T data for 838 watersheds across the United States were calculated from the PRISM climate national datasets for the period 1971-2000. Mean Q/A for these watersheds for this period was subtracted from the mean P to obtain estimates of mean watershed ET. A regression equation was developed for ET/P as a function of P, T, and the mean diurnal fluctuation in T, resulting in an R² value of 0.867. Land cover parameters were later added to the equation resulting in an R² value of 0.882.

The regression equation represents current ET conditions across various climatic settings of the contiguous United States. Thus forecasts of future ET conditions in a warmer climate can be projected based solely on the known current ET relation to T and P. In this study, average results from 16 different GCM models for three different CO₂ scenarios for the 1990s and 2090s were used to look at average projected changes in T and P over the 21^st century. These projected changes in T and P were used to calculate county-scale ET rates in 2100 and make comparisons to 2000 to project potential changes in ET and available water (P – ET). Results showed that the greatest impact of rising ET is likely to be in the northern High Plains and central Rocky Mountains, where a decline in available water of more than 30 percent is possible in places. Results also reveal areas where agriculture is currently required to be supplemented by irrigation (ET > P), mainly the Central Valley of California and western High Plains. The region that would require irrigation on the High Plains is projected to expand eastward by hundreds of kilometers, especially across the Dakotas.