Micro-scale spatial variability and the timing of snow melt runoff in a high mountain catchment

Abstract

A process-based distributed model of snowmelt processes was developed to simulate snowpack depletion during the melt season in a small headwater catchment in the Spanish Pyrenees. The grid scale model was calibrated and evaluated using meteorological and snowpack data from the catchment. The model was then run in distributed mode to examine the relative importance of spatial patterns of snow accumulation and melt on snowpack depletion. Model results demonstrated that the spatial distribution of snow water equivalent (SWE) at the start of the melt season is the primary control on patterns of snow disappearance within the catchment, and that representation of spatial variability in melt rates is of minor importance. Accurate simulation of the rate of snow cover depletion during the melt season was shown to be important for accurate simulation of melt runoff rates at the catchment scale, because of the importance of snow covered area in determining the catchment average melt rate. The practical implication of these results is that micro-scale spatial variability in SWE (i.e. of the order of 10–100 m) needs to be taken into account for accurate simulation of snowmelt runoff at larger scales. However, it appears that variability in energy inputs at these scales can be neglected without significant loss of accuracy. It is suggested that a frequency distribution approach could be adopted to parameterise SWE variability into grid or catchment scale models.