Two important factors that control snow albedo are snow aging (grain growth) and presence of light-absorbing impurities (aerosols) in snow. However, most current regional climate models do not include such processes in a physically-based manner in their land surface models, which can have serious implications to simulated surface energy and water budgets, and ultimately to water resources and regional climate. We improve snow albedo calculations in the Simplified Simple Biosphere (SSiB) land surface model coupled with the Weather Research and Forecasting (WRF) regional climate model (RCM), by incorporating the SNow ICe And Radiative (SNICAR) scheme. SNICAR is a snow radiative transfer model that simulates snow albedo evolution due to snow grain growth and presence of aerosols in snow in a physically-based manner. The land surface model is further modified to account for deposition, movement, and removal by meltwater of such impurities in the snowpack. The newly modified SSiB-3 land surface model (LSM) is validated offline with in situ observations at a location in Western U.S. (WUS), and shows significant improvements in simulated snow albedo and depth when dust in snow is considered, and reproduces snow grain size and vertical distribution of dust in snow that are comparable to those observed.
The online, coupled version of the new model, WRF/SSiB-3/aer, is then employed to conduct two 10-year long simulations over North America - aerosols-loaded (AER) snow and clean snow (NOAER) cases - to investigate the impact of aerosols in snow (AIS) on surface energy (SEB) and water (WB) budgets on a regional scale. GOCART surface aerosol deposition data is used in AER scenario. Comparisons between AER and NOAER simulations reveal albedo is reduced the most during the ablation period in the presence of AIS, inducing a surface radiative forcing (RF) ranging 8.5 W/m2 over WUS to 13.6 W/m2 over NCan during MAM and MJJ (10-year means), respectively, but as high as 65 W/m2 during peak ablation. This corresponds to an increase in skin temperature (TSK) of 0.5 °C and a subsequent spring snow mass reduction ranging 12 - 45 mm in the aerosol-loaded snow case. Changes found in our study are higher than those found by GCM simulations, RF being an order of magnitude large in our RCM simulation, for example. On a sub-regional scale, our simulations reveal mountainous areas like the Sierra Nevada and Rockies see larger changes in TSK, runoff, and soil moisture (SM) due to AIS at higher elevation during the spring season. Furthermore, the Sierras see a net decrease in SM, which we show can have implications to wildfire vulnerability, while in the southern Rockies AIS cause shifts in runoff timing (9-year mean of 3.5 days earlier).