Snowpack is a critical freshwater reservoir that sustains ecosystem, natural habitat, hydropower, agriculture, and urban water supply in many areas around the world. Accurate estimation of basin scale snow water equivalent (SWE), through both measurement and modeling, has been significantly recognized to improve regional water resource management. Recent advances in remote data acquisition techniques have improved snow measurements but our ability to model snowpack evolution is largely hampered by poor knowledge of inherently variable high-elevation precipitation patterns. For a variety of reasons, majority of the precipitation gages are located in low and mid-elevation range and function as drivers for basin scale hydrologic modeling. Here, we blend observed gage precipitation from low and mid-elevation with point observations of SWE from high-elevation snow pillow into a physically based snow evolution model (SnowModel) to better represent the basin-scale precipitation field and improve snow simulations. To do this, we constructed two scenarios that differed in only precipitation. In WTH scenario, we forced the SnowModel using spatially distributed gage precipitation data. In WTH+SP scenario, the model was forced with spatially distributed precipitation data derived from gage precipitation along with observed precipitation from snow pillows. Since snow pillows do not directly measure precipitation, we uses positive change in SWE as a proxy for precipitation. The SnowModel was implemented at daily time step and 100 m resolution for the Kings River Basin, USA over 2000-2014. Our results show an improvement in snow simulation under WTH+SP as compared to WTH scenario, which can be attributed to better representation in high-elevation precipitation patterns under WTH+SP. The average Nash Sutcliffe efficiency over all snow pillow and course sites was substantially higher for WTH+SP (0.77) than for WTH scenario (0.47). The maximum difference in observed and simulated peak SWE was 810 mm for WTH and 380 mm for WTH+SP, which led to underestimation of snow season length and melt rate by up to 30 days and 12 mm/day, respectively, in WTH scenario. These results indicate that point scale snow observations at higher elevation can be used to improve precipitation input to hydrologic modeling in mountainous basins.
Sohrabi, M.; Safeeeq, M.; Conklin, M. H. (2017): Improving snow water equivalent simulations in an alpine basin using blended gage precipitation and snow pillow measurements. Fall Meeting, American Geophysical Union, December 2017. Abstract H43C-1658 .