Impact of improved snow canopy and frozen soils on mesoscale simulations of the wintertime boundary layer

Different configurations of a 10-km version of the non-hydrostatic MC2 model were integrated daily over northeastern North America during the whole winter season of 1998-1999 (i.e., from October 1998 to April 1999). Two surface schemes were tested in this study: a simplified force-restore scheme that is used operationally at the Canadian Meteorological Centre, and the ISBA (Interactions-Soil-Biosphere-Atmosphere) scheme which includes more physical representations of vegetation and snow. In the operational scheme, the heat capacity of snow is climatologically modulated and the effect of snow melting on the surface temperature is parameterized in a crude manner. In ISBA, melting-freezing of snow is derived from more physically-based energy and water budgets; prognostic equations determine the evolution of the density and albedo of snow as a function of age; and the snow-water equivalent (or depth of the snow pack) evolves according to the simulated precipitation and evaporation-sublimation at the surface. In a third model configuration, new reservoirs for liquid water in the snow pack and for frozen water in the soil have been included in ISBA. Results and statistics from the large number of 24-h integrations conducted with these three configurations show the positive impact of including more sophisticated snow processes. For instance, surface air temperature over relatively ‘old’ snow was, in certain cases, about 10 oC warmer in the ISBA simulations as compared to the operational setup, due to more realistic heat coefficients. Air temperature differences between the two ISBA versions were also found to be large over forest regions, especially in late-winter/early-spring, when sensible heat fluxes are large due to both increasing solar radiation and control of evapotranspiration by frozen water in the soil.