Assessing the Heterogeneity of Snow-Water Equivalent During the Snowmelt Season: Spatial Variability and its Controlling Factors in an Alpine Setting

A distributed energy-balance model was developed for simulating the seasonal snowmelt at a grid scale of 30 meters in rugged alpine terrain and applied to a 45 km2 headwater catchment in the Tobacco Root Mountains, Montana. Micrometeorological data that were collected over the 1997-1998 snow-accumulation and snowmelt seasons were used as boundary conditions for estimating initial snow distributions, as well as calculating the timing and distribution of snowmelt from 47,000 individual cells in the basin. Heterogeneity of energy distribution was accomplished by coupling the general radiation algorithm of Olyphant (1986) with detailed topographic analysis and remotely sensed ground-cover classifications. Heterogeneity of the initial snow distribution was calculated using SnowTran-3D (Liston and Sturm, 1998). Simulated snowmelt volumes compare favorably with measured outflow from a stream gage located just downstream of the study area. Although the primary control on snowmelt patterns was net radiation in the early snowmelt season, the later season patterns and melt rates were more closely related to the initial distribution of snow-water equivalent. This is demonstrated by a decline in the standard deviation of radiation received by remnant snowcover from 2.80 MJ/m²d^-1 to 0.88 MJ/m²d^-1. Also by the end of the simulation, virtually no snow remained on west-facing (windward) slopes, whereas east-facing (leeward) slopes had the highest remaining number of snow-covered cells. Such results emphasize the critical importance of establishing an accurate initial snow-water equivalent distribution in addition to representing the evolving distribution of energy income. This modeling approach may lead to a more complete understanding of the impact of heterogeneity on the evolution of an alpine snowpack.