SEARCH

SEARCH BY CITATION

Keywords:

  • SWE;
  • snow depth;
  • hydrological modeling;
  • SNODAS;
  • WEBB

Abstract

The National Weather Service's Snow Data Assimilation (SNODAS) program provides daily, gridded estimates of snow depth, snow water equivalent (SWE), and related snow parameters at a 1-km2 resolution for the conterminous USA. In this study, SNODAS snow depth and SWE estimates were compared with independent, ground-based snow survey data in the Colorado Rocky Mountains to assess SNODAS accuracy at the 1-km2 scale. Accuracy also was evaluated at the basin scale by comparing SNODAS model output to snowmelt runoff in 31 headwater basins with US Geological Survey stream gauges. Results from the snow surveys indicated that SNODAS performed well in forested areas, explaining 72% of the variance in snow depths and 77% of the variance in SWE. However, SNODAS showed poor agreement with measurements in alpine areas, explaining 16% of the variance in snow depth and 30% of the variance in SWE. At the basin scale, snowmelt runoff was moderately correlated (R2 = 0.52) with SNODAS model estimates. A simple method for adjusting SNODAS SWE estimates in alpine areas was developed that uses relations between prevailing wind direction, terrain, and vegetation to account for wind redistribution of snow in alpine terrain. The adjustments substantially improved agreement between measurements and SNODAS estimates, with the R2 of measured SWE values against SNODAS SWE estimates increasing from 0.42 to 0.63 and the root mean square error decreasing from 12 to 6 cm. Results from this study indicate that SNODAS can provide reliable data for input to moderate-scale to large-scale hydrologic models, which are essential for creating accurate runoff forecasts. Refinement of SNODAS SWE estimates for alpine areas to account for wind redistribution of snow could further improve model performance. Published 2012. This article is a US Government work and is in the public domain in the USA.