Snow accumulation patterns in Idaho

We obtained snow course data from 230 stations in Idaho and western Montana, and our research focused on the transformation of these point data to a statistical surface so that the data could be better utilized for water resources applications. The transformation question was approached at two levels. At the first level, we used the snow course data directly with different spatial interpolation algorithms including the inverse distance squared and universal kriging methods. The first approach yields poor results because of the non-random, non-regular spatial sampling that the snow courses represent. At the second level, we used a geographic information system (GIS) to collect several layers of information such as elevation, slope, and aspect and built a regression model to predict monthly SWE values for the Idaho snow courses. These models are formulated at the watershed scale, and generally yield very good results (R2 = 0.70 -0.98). The regression models were then employed to build a denser network of points for spatial interpolation. Future research will examine the reliability andutility of these 'snopographic' surfaces.