The influence of lateral snow redistribution processes on snow melt and sublimation in alpine regions

Summary

Snow plays a crucial role in determining moisture and energy exchanges between the land surface and atmosphere. Numerous studies have demonstrated that the accurate simulation of the Alpine snow cover is difficult due to the manifold interactions of preferential snow deposition, wind induced snow transport, and snow slides. The presented study discusses the spatial snow distribution as well as the associated sublimation and melt rates within the Berchtesgaden National Park area by using a collection of basic and medium complex models (SnowModel, SnowTran-3D, SnowSlide). A comparison between model results and classified satellite images was done for a general model validation. It was shown that the inclusion of snow transport processes into the model environment, can improve the accuracy of the predicted snow distribution. Sublimation from turbulent-suspended snow was analysed and it could be seen that the inclusion of gravitational snow transport processes is able to reduce the predicted sublimation rates at the crest regions by 44%. Furthermore, an analysis was performed to identify critical spatial scales, associated with snow redistribution and melt-water generation to define when snow transport processes should be accounted for. It was seen that the effect of snow cover redistribution on the predicted melt rates is only detectable if the observed area is small and in the range of a slope or a face. There was no effect detectable if the total catchment area was considered.

Introduction

The presented study is focused on defining an adequate model to represent snow patterns formed by the accumulation and ablation dynamics of the snow cover as well as by lateral snow transport processes and on the effect of these patterns on calculated sublimation from snow in turbulent suspension. Furthermore, a scaling study is realised with respect to the calculated snow melt rates. The scaling study is focused on the question whether blowing snow processes which are mentioned as the dominant factor, which is influencing the snow distribution at scales from tens to hundreds of meters (Liston, 2004) and gravitational snow transport which can modify the snow distribution at tens of meters (Bernhardt and Schulz, 2010) can impact the predicted melt rates over the same order of scales. The necessary model runs were accomplished by applying an established snow evolution model (SnowModel; Liston and Elder, 2006a), that includes model routines for the description of snow accumulation and ablation as well as wind induced snow transport, to the German Alpine catchment Berchtesgaden National Park (BNP).

Sublimation from suspended particles as firstly mentioned by Dyunin (1967) is a moisture flux from the snow particle to the surrounding undersaturated atmosphere. The intensity of the sublimation process depends of the availability of transportable snow, on the wind speed and on the air humidity (Hood et al., 1999). Sublimation rates were firstly quantified by Schmidt, 1972, Lee, 1975 and later on estimated through model calculations (Dery and Yau, 2001, MacDonald et al., 2010; Pomeroy, 1991; Schmidt, 1982, Strasser et al., 2008). A literature review showed that the calculated seasonal rates of sublimations do have a very high range, reaching values from 4 mm within 6 months at a snow covered Antarctic ice shelf site (King et al., 1996), to 37–85 mm within a snow season at a Canadian tundra site (Essery et al., 1999, Pomeroy, 1997), to 135 mm in 5 month in the Colorado front range (Meiman and Grant, 1974), up to 1265 mm within a snow season at an Alpine ridge (Strasser et al., 2008). Given this uncertainty, it is of importance to know if the model results indicate that sublimation has a significant impact on the water balance of a whole catchment, or if the sublimation effect is only of local importance. Within the presented work, the results of Strasser et al. (2008) were extended by the inclusion of gravitational snow transport processes. This is needed, because gravitational transport removes snow from crests and steep walls which do usually show enhanced wind speeds and accumulates snow at the foothills and wind sheltered areas. The effect of this transport process on the calculated sublimation rates is discussed and the results are compared to established results of Dery and Yau (2001). Furthermore, given the current computational constraints for describing snow hydrological processes within complex, coarse-scale hydrological and climate models, it is of great interest to analyse and define which processes, what type of model complexity, and which process interactions are required to be represented at the scales associated with each application (Bloschl, 1999). The Snow Models Intercomparison Project (SnowMIP) (Etchevers et al., 2004) showed that present-day 1-dimensional snow models are able to reproduce the snow cover evolution if they are applied in close vicinity of a meteorological station. Results become less confident for spatially distributed 2- or 3-dimensional models (see Fig. 1), especially in areas with complex terrain or in forested areas (Bernhardt and Schulz, 2010, Bernhardt et al., 2009, Liston, 2004, Liston and Elder, 2006a, Pomeroy et al., 2002, Rutter et al., 2009). This can be attributed to errors and uncertainties in the calculated meteorological fields which drive the models, as well as to oversimplifications of the model formulations with respect to snow–canopy interactions, windblown snow, gravitational snow transport, and preferential snow deposition (Bernhardt et al., 2009, Gruber, 2007, Lehning et al., 2006, Liston, 2004).

Hence, the predicted melt water generation is analysed with the help of two different model set-ups and at different scales for analysing for which scales and which model resolution the calculation of lateral snow transport processes is needed.