Towards Inclusion of Sub-grid Variability of Snow in Distributed Permafrost Models

High altitude environments are often exposed to strong winds, thus drifting snow can create a small-scale pattern of highly variable snow heights. This has profound implications for the ground thermal regime. Variability in mean annual ground surface temperatures of up to 6°C within areas of 1 km² have been documented from Svalbard and Norway earlier, and can to a large degree be explained by variation of maximum snow heights. Land surface models or permafrost models employing an average snow height per grid cell of e.g. 1 km²or larger are not capable of representing such local-scale variability.

This study presents a robust scheme where we statistically represent the sub-grid variability of ground surface temperatures and demonstrate that we can reproduce the distribution within a grid cell. This is achieved by (1) applying a simple terrain parameter (Winstral et al., 2002) on a 10 meter DEM, estimating exposure from a selected wind direction. The degree of exposure is estimated for main wind directions over the winter season, derived from HINDCAST 10 m wind direction data (ERA40 reanalysis downscaled using HIRLAM, Norwegian Meteorological Institute), available at approximately 11 km (0.1 degree) spatial resolution and 2 hour temporal resolution back to 1957. Here we obtain a statistical representation of the distribution of maximum snow heights. (2) we use this information for calculating ground temperature distributions within the grid cells using a simple equilibrium permafrost model (CryoGRID 1). Finally, we test this approach at four 100 km²sized areas where we have extensive field observations, located along a latitudinal transect in Norway. We demonstrate that the inclusion of a snow distribution function within each grid-cell clearly improves the representation of permafrost coverage in high latitude areas.

Winstral, A., K. Elder, and R. E. Davis. 2002. Spatial snow modeling of wind-redistributed snow using terrain-based parameters, Journal of Hydrometeorology, 3: 524-538.