Implementation of digital terrain analysis to capture the effects of wind redistribution in spatial snow modeling

Though the effects ofwind-induced snow redistribution in mountainous locales are readily observed, the complexity of wind fields over rugged terrain has stymied attempts to model snow transport in these regions. The accuracy of spatial snow models that have been unable to include redistribution effects has been thusly compromised. This research designs, develops, and evaluates parameters that numerically represent terrain features that affect the degree of exposure to prevailing winds at each cell within a grid. The study area is the windswept 8.1-km2 Green Lakes Valley watershed along the Colorado Front Range. Comparisons are made among parameters representative of surface curvature, average upwind fetch, average upwind horizon, and average upwind slope gradient. All of these representative parameters are significant predictors of snow distribution representative of a process at least as important as radiation input in determining peak snow accumulations. Additionally, a slope divergence parameter is shown to successfully delineate regions prone to heightened accumulations in lee slope eddy zones. Ultimately it is demonstrated that the inclusion of a wind redistribution factor within a regression tree model produces a more accurate and very different spatial model of snow distribution compared to a model that does not include an eolian influence.