Evaluation of snow cover fraction for regional climate simulations in the Sierra Nevada

Mountain snow cover plays an important role in regional climate due to its high albedo, its effects on atmospheric convection, and its influence on runoff. Snowpack water storage is also a critical water resource and understanding how it varies is of great social value. Models are often employed to reconstruct snowpack and explore and understand snow cover variability. Here, we use a new, accurate satellite-derived snow product to evaluate the ability of the Weather Research and Forecasting (WRF) regional climate model, combined with the Noah land surface model with multi-parameterization options (Noah-MP), to simulate snow cover fraction (SCF) and snow water equivalent (SWE) in a 3-km domain over the central Sierra Nevada. WRF/Noah-MP SWE simulations improve on previous versions of the Noah land surface model by removing an early bias in snow melt, though a 2-day positive melt bias in SWE timing remains significant at the 90% confidence level. In addition, WRF/Noah-MP identifies the areas where snow is present to within 94.3% and captures large-scale variability in SCF. Temporal root mean squared error (RMSE) of the domain-average SCF was 1938.6km(2) (24%). However, our study shows that WRF/Noah-MP struggles to simulate SCF at finer spatial scales. The parameterization for SCF fails to produce temporal variations in grid-scale SCF, and depletion occurs too rapidly. As a result, the WRF/Noah-MP SCF parameterization reduces to a binary function in mountain environments. Sensitivity tests show that adjustment of the parameterization may improve simulation of SCF during accumulation or melt but does not remove the bias for the entire snow season. Although WRF/Noah-MP accurately simulates the presence or absence of snow, high-resolution, reliable SCF estimates may only be attainable if snow depletion parameterizations are designed specifically for complex topographical areas.