Sub-grid-scale topography and the simulation of northern hemisphere snow cover

A new interactive snow hydrology scheme suitable for incorporating into a general circulation model (GCM) has been developed. The algorithm accounts for a prognostic calculation of snow depth and snow fraction and models the hydrological budget and energy balance over the snow surface. In addition, the scheme incorporates the effect of sub-grid-scale variations in topography with a novel statistical approach. A 10-year GCM simulation using the new interactive snow hydrology is analysed. The snow climatology and variability is compared over the Northern Hemisphere with observational data from NOAA/NESDIS. The model is shown to perform well. We have assessed the impact of parameterizing the effect of sub-grid topographic variations by degrading the model in the sense of removing these variations. Less snow is predicted, as a whole, in the (more sophisticated) control model. Also, the snow cover in the degraded model is found to lag that in the control model in both the broad-scale growth and retreat of snow. This is understood in terms of the 'seeding' and 'unseeding' of snow cover that only the full control model can represent. Some detailed comparisons in specific marginal regions are undertaken in order to expose the differences between the two versions. The model that does take sub-grid-scale variability of topography into account is superior to the degraded model.