A significant fraction of the Northern Hemisphere snow surface is sheltered by forests. Snow surface energy budgets of solar radiation, longwave radiation and turbulent fluxes in forested area, on which accurate prediction of snow ablation strongly depends, are dramatically modified by the overlying canopy when compared to open areas. The notorious delayed springtime snow melting in general circulation models might be resulted from inadequate treatment in vegetation sheltering.

The National Center for Atmospheric Research (NCAR) Land Surface Model version 1 (LSM1) canopy parameterization scheme coupled with a newly developed physically-based snow model (VISA) in winter is validated against BOREAS data observed at four stations SSA-OA (South Study Area, old aspen), SSA-OJP (Old Jack Pine), NSA-OJP (north study area, Old Jack Pine) and NSA-YJP (Young Jack Pine).

The two-stream approximation used in LSM1 for short-wave radiation parameterization basically captures the transferring of solar radiation through canopy and gives reasonable albedo simply by adjusting leaf area index (LAI) to match the observed. After LAI is adjusted, the albedo given by LSM1 in winter is still higher than observations, which can be reduced by increasing the maximum interception capacity of precipitation, guided by the fact that the canopy interception capacity for snow is larger than that for rain.

The lower estimation of long-wave radiation emitted from canopy, which is a function of leaf and stem area index, is subject to the uncertainty of leaf and stem area index. The turbulent flux parameterization within canopy overestimates the downward sensible heat flux from canopy to snow surface and lead to earlier snow melt than the open area. Sensitivity tests show that the turbulent sensible heat flux strongly depends on the canopy absorption coefficient of momentum, its stability adjustment, and higher canopy temperatures in daytime induced by neglect of phase change between liquid and solid water on canopy foliage. More detailed gradient and flux observations within forest are needed to better parameterize the sensible heat transfer within canopy.