Combining Remote Sensing Data, Airborne Snow Observations and High Resolution Hydrologic Modeling to Improve SWE Simulation and Validation over Mountainous Terrain in Western US

The mountain snowpack is an important component of the hydrologic cycle and an essential resource for communities across the globe. From scientists to water resource managers and politicians, accurate information on snow amount, extent, and melt rate is key in understanding physical processes within the earth system and for planning a sustainable existence. Being able to correctly simulate snow, both historically and projecting into the future, is not only a great necessity, but also a challenge. Combining the power of satellite and airborne observations with numerical model simulations can bring us closer to having a more complete picture of the current and future state of the snowpack. In this study, the VIC (Variable Infiltration Capacity) macroscale hydrologic model is employed over Western United States (WUS) at a horizontal resolution of 0.0175°, or 3 km2, to simulated the snowpack during the recent drought this area has been experiencing (WY 2013-2015). Remote sensing data (PRISM, MERRA) are used as meteorological forcing, as well as in the assimilation process (MODSCAG) for a more optimal estimation of snow water equivalent (SWE). The model is run under two scenarios, with and without assimilation of MODSCAG snow cover, and the two cases are compared against in situ and airborne SWE products (Airborne Snow Observatory, ASO). Several questions are addressed: how does a spatially distributed snow product like ASO improve validation of high-resolution SWE model simulations, compared to validation against sparsely available in-situ measurements, which are often only available at low-mid elevations? ASO provides a unique and comprehensive view of the snowpack in both space and time, and over complex terrain of mountain watersheds, which has not been previously available. Such comparison can also help identify the level of improvement when assimilation of snow cover is used in estimating modeled SWE. These results can help improve the models we use, from which forcing datasets are best, to how and what type of assimilation might be most appropriate, to optimizing model parameterizations.