In mountain regions, wind is known to cause inhomogeneous snow distribution. Recently, it has been shown that this snow distribution can be explained by considering high-resolution flow fields created by numerical flow models. It has also been shown that small-scale deposition features are mainly caused by saltation and snow redistribution, larger-scale inhomogeneities are caused by preferential deposition, and the total deposition patterns correlate well with the mean wind speeds. On the basis of the strong correlation between wind fields and snow depth, we develop a new empirical parameterization of preferential deposition of precipitation, which is only governed by the mean flow field. The parameterization is based on case studies, where the physically based model for wind-induced snow transport ALPINE3D was combined with wind fields from the regional atmospheric model ARPS (Advanced Regional Prediction System). The parameterization, along with ARPS wind fields and a distributed energy balance snowmelt model, is used to simulate the 2006/2007 accumulation season over the catchment of Haut Glacier d'Arolla, a glacierized Alpine basin. The results were validated with a Light Detection and Ranging (LiDAR)-derived fully distributed high-resolution dataset of snow depth and show that the developed parameterization, despite being simple, is able to largely reduce the difference between modelled and measured snow depth and is suitable for application in glacier mass balance models. Copyright (C) 2010 John Wiley & Sons, Ltd.
