Quantifying the impact of remotely-sensed precipitation, soil moisture and snow products for North American hydrologic hazard assessment

Reliable detection and prediction of hydrologic hazards, including floods, droughts, and rainfall-induced landslides, requires an understanding of the uncertainty in inputs (e.g. remotely sensed versus radar and rain gauge-based precipitation) with and without observational constraints on model state variables such as soil moisture and snowpack. In this invited talk, we will show results from NASA's Land Information System configured over the North American Land Data Assimilation System (NLDAS) domain, to demonstrate the impacts of remotely sensed precipitation on landslide triggering, as well as soil moisture and snow product assimilation on drought assessment (expressed via simulated ET and soil moisture profiles) and flood assessment (expressed via simulated streamflow). For our soil moisture assimilation experiments, we utilize two different surface soil moisture retrievals (NASA Level-3 product and the Land Parameter Retrieval Model (LPRM) product from VU Amsterdam) from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). For the snow assimilation experiments, we utilize the snow covered area from the Moderate Resolution Imaging Spectroradiometer (MODIS) and snow depth from AMSR-E. The results indicate that the assimilation of LPRM and bias-corrected AMSR-E SWE data help in further improving the soil moisture profile, snowpack and ultimately fluxes such as ET and runoff.