Distribution of Snow Water Equivalent Climatology for the Western United States From Snow Telemetry (SNOTEL) Data

A quantitative assessment of the spatial distribution of snow data provides valuable insight and information for water resource management and many other environmental studies. Hydrologic models that include spatial distribution of snow water equivalent (SWE) and snowmelt are shown to significantly improve estimates of spring and summer discharge where snowfall and snowmelt components are dominant. In the western United States, SWE data mainly comes from the snow telemetry (SNOTEL) network of the United States Department of Agriculture, Natural Resource Conservation Service. Unfortunately, SNOTEL observation sites are generally sparse, and therefore cannot provide adequate information about the spatial distribution of the snow data. Two methods for estimating spatial distribution of SWE in mountainous basins in the western United States are developed and evaluated in this research. The first one is based on the observation that high correlation exists between precipitation and SWE in mountainous regions during winter months. The method takes advantage of high quality, monthly climate maps of precipitation and temperature derived using the Parameter-elevation Regression on Independent Slopes Model (PRISM) methodology developed by Oregon State University. Since the PRISM methodology is designed to account for effects of varying terrains, topographic facets, proximity to coastal regions and other complex climatic regimes on precipitation and temperature, it was hypothesized that SWE climate maps derived through PRISM maps will account for the same effects. For the second method, the western U.S. is divided into nine homogeneous snow climate regions. For each region, a multivariate regression technique was used to estimate the SWE normals at 1 km grid from physiographic parameters of elevation, slope, aspect and forest density; as well as solar radiation, precipitation and temperature normals. Different statistical measures were used to evaluate and compare two methods. Although both methods produced satisfactory results, the SWE estimates were far superior when second method was used.