The response of streamflow and evapotranspiration to changes in snowmelt across the Western United States
PhD: University of Colorado Boulder, 2018.
Mountain regions disproportionately produce streamflow for downstream ecosystems and communities. In the western United States this snowmelt derived water is valued in the trillions of dollars. Given the high value of snowmelt-derived water, understanding how streamflow production and vegetation water use from mountain regions may change is of critical importance. Snowmelt rate, timing, and amount are forecast to change under future climate, potentially altering streamflow and evapotranspiration patterns. This dissertation investigates the relationship between snowmelt rate, timing, and amount and runoff or streamflow at the plot and regional scales across the western United States. Additionally, the effects of future land cover, precipitation, and air temperature changes on streamflow from a headwaters catchment are investigated.
At the plot scale, observations and hydrologic modeling were used to investigate how changes in snowmelt rate, timing, and amount affect snowmelt season runoff production and subsurface water storage in Colorado (CO) and California (CA). The snowmelt modeling experiment was designed to eliminate the observed collinearity between snowmelt rate, timing, and amount. Results indicate that runoff was most sensitive to snowmelt timing and rate at CO and CA, respectively (β=-0.31 vs. 0.22 and β= -0.31 vs. 0.67 for snowmelt timing vs. rate, respectively). Snowmelt season changes in subsurface storage were most sensitive to snowmelt timing at both CO and CA (β= -0.24 vs. 0.18 and β=-0.474 vs. 0.466 for snowmelt timing vs. rate, respectively).
At the watershed scale, the Landscape Disturbance and Succession (LANDIS) land cover evolution model was used in conjunction with the Regional Hydro-Ecologic Simulation System (RHESSys) to investigate how changes in climate and land cover may alter streamflow from 2000 to 2100 in a catchment on the Colorado Front Range. As forest cover in the catchment increased, counter intuitively, streamflow also increased by 29-44% by 2100 driven by reductions in wind-scour of snow out of the catchment and decreases in evapotranspiration. These changes in streamflow were partially attributed to land cover change but also to air temperature driven changes in snowmelt timing. At the regional scale, a long-term hydrometeorology data set was used to elucidate a possible mechanism linking snowmelt rate to streamflow production. An ensemble of Budyko streamflow anomalies (BSA), a measure of streamflow production, at ~20,000 Variable Infiltration Capacity model grid cells was computed. BSA was correlated with simulated baseflow efficiency (r2=0.64) and snowmelt rate (r2=0.42). A strong correlation between snowmelt rate and baseflow efficiency (r2=0.73) links these relationships and supports a possible streamflow generation mechanism wherein greater snowmelt rates increase subsurface flow and streamflow production.