Simple Snowdrift Model for Distributed Hydrological Modeling

Abstract

A simple snowdrift model was developed, incorporated into a distributed winter-time hydrological model, and tested against snow measurements from a hillside in eastern Washington State. Snow movement can be an important factor in the distribution of spring soil moisture and runoff. Although current hydrological models often attempt to account for heterogeneities in precipitation distribution, they do not account for snowdrift. Snow melts and accumulates during the same times that it is redistributed. Therefore, evaluation required a snowmelt/accumulation model to be coupled with the snowdrift model. The snowmelt/accumulation model used the standard energy balance approach and performed well, i.e., standard errors of snow water equivalent ≈1 cm. The snowdrift model’s simulated snow distribution generally agreed with observed snow distribution across a hill. Most notable were the model’s ability to correctly place a snowdrift on the lee side of the hill and its ability to predict snow removal from nondrift areas. The effects of snow redistribution and the model’s ability to reproduce these were obvious when overlaid on model results that ignored snowdrift.