Physiographic Drivers of Snow Water Storage: Modeling the Spatial Distribution of Our Water & the Suitability of Our Monitoring Network

Topography controls elevation, slope, aspect and exposure and can influence snowpack dynamics across a watershed. Snow accumulation and ablation processes are also highly influenced by vegetation structure at multiple temporal and spatial scales, particularly in forested regions. Spatially-distributed physiographic data and modeled snow water equivalent data were used to develop a binary regression tree model to characterize snowpack conditions across the McKenzie River Basin, Oregon. A GIS analysis model was used to characterize expected snowpack and predict the depth and spatial variability of the snowpack across the watershed using the physiographic drivers classified by the statistical model. This model was applied to the North Santiam and Middle Fork Willamette River Basins, as well as the greater Willamette River Basin to predict the spatial distribution and quantity of peak snow water equivalent (SWE). The model estimated the total volume of 1.1 km3 SWE in the McKenzie River Basin, 0.45 km3 SWE in the North Santiam River Basin, 1.4 km3 SWE in the Middle Fork Willamette, and 3.3 km3 SWE in the greater Willamette River Basin. The model tends to predict slightly less snow water equivalent than observed at the NRCS - SNOTEL sites throughout the greater Willamette River Basin. The current SNOTEL snow monitoring network captures current average snow conditions well, but may not be representative in the future as climate impacts snowpack distributions. This nested model approach was used to develop an objective and representative monitoring network design that may better capture the full range of snowpack variability across the complex topography and densely forested watersheds of the Western Oregon Cascades.