Å·ÃÀ¿Ú±¬ÊÓƵ

Barnhart, Theodore ¹ ; Molotch, Noah  ² ; Tague, Christina  ³

¹ INSTAAR / Å·ÃÀ¿Ú±¬ÊÓƵ Boulder
² INSTAAR / Å·ÃÀ¿Ú±¬ÊÓƵ Boulder
³ University of California Santa Barbara

The mountain snowpack is important for runoff generation across the western United States. Near surface warming alters the amount of precipitation that falls as snow changing the amount, rate, and timing of snowmelt. Recent work linked snowmelt rate to streamflow production across the western United States and linked snowmelt rate to snowpack magnitude and snowmelt timing. We seek to reveal the dominant runoff production factor by disentangling the relationships between snowmelt rate, timing, and amount. We use observations of evapotranspiration and snowmelt from Niwot Ridge, CO (CO), Providence Creek, CA (CA), and the Valles Caldera, NM (NM) as well as the Regional Hydro-Ecologic Simulation System (RHESSys) to assess the linkage between snowmelt rate, amount, timing, and runoff production. We conducted 250,000 RHESSys simulations at each site varying the rate, timing, and amount of snowmelt based on the observational record across a range of soil water storage and drainage parameter sets calibrated for each site. Analyses of observational data shows that runoff production from each site is most sensitive to a unit change in snowmelt rate. Snowmelt experiment multiple regression results reveal that snowmelt rate is also the most important driver of runoff production at CO and CA (r2=0.72 and 0.68, respectively with p<0.001 for both sites) with coefficient values of 0.62 and 0.72, respectively (Figure 1). Snowmelt timing is similarly important but with the opposite sign with model coefficients of -0.57 and -0.44, respectively (Figure 1). At NM, neither snowmelt rate nor snowmelt timing were good predictors of runoff production (r2=0.05, p<0.001, Figure 1). The snowmelt rate and timing coefficients showed sensitivity to soil water storage at CO while the response surface for both coefficients were flat at CA. This suggests that soil water storage is more important than soil water drainage at CO and that CA is insensitive to changes in both soil water drainage and storage. Simulation output suggests that rapid snowmelt partitions more water to the groundwater store suggesting a subsurface streamflow generation pathway. These results suggest that, as climate warming decreases the mountain snowpack and causes earlier, slower snowmelt, runoff production is likely to decrease primarily due to the change in snowmelt rate.

Fig 1. Snowmelt experiment multiple regression coefficient comparison showing higher coefficients for snowmelt rate for both the CO and CA sites.