An Evaluation of High-Resolution Regional Climate Model Simulated Snow Cover Using Satellite Data (With Implications for the Simulated Snow-Albedo Feedback)

Abstract:

Snow cover often exhibits large spatial variability over mountainous regions where variations in elevation, aspect, vegetation, winds, and orographic precipitation all modulate snow cover. Under climate change, reductions in mountain snow cover are likely to substantially amplify regional warming via the snow-albedo feedback. To capture this important feedback it is crucial that regional climate models (RCMs) adequately simulate spatial and temporal variations in snow cover.

Snow cover simulated by high-resolution RCMs over the central Rocky Mountains of the United States is evaluated. RCM simulations were conducted using the Weather Research and Forecasting (WRF) model on a 4 km horizontal grid forced by reanalysis boundary conditions over a seven-year time period. A pair of simulations is considered that differ in the domain size (regional vs. continental) and the land surface model (Noah vs. Noah-MP) employed. RCM output is compared with high-resolution gridded satellite analyses of surface albedo and fractional snow cover derived from the Moderate Resolution Imaging Spectroradiometer (MODIS).

Results reveal that both RCMs are generally successful at reproducing the observed seasonal cycle and interannual variability of snow extent over the high terrain of the Rockies. However, in simulations using the Noah land surface model (LSM), sub-grid scale fractional snow covered area of grid cells containing snow is systematically too high compared to observations, often exceeding observations by more than 0.2. This bias in fractional snow cover leads to a substantial positive bias in regional surface albedo. Simulations using the Noah-MP LSM produce more realistic variations in fractional snow cover and surface albedo, likely due to its more-realistic treatment of canopy effects.

We quantify how differences in simulated snow cover affect the strength of the snow-albedo feedback under climate change. Both RCMs were used to conduct representative 7-year simulations of a warmed climate using the pseudo global warming framework. These simulations show markedly different spatial and seasonal patterns of warming that are linked to differences in snow-albedo feedback strength and the treatment of fractional snow cover.