Examining the Representativeness of Current Climate Measurement Site Locations in the McKenzie River Basin, Oregon

Hydrologists and resource managers frequently use measurements from the Natural Resources Conservation Service (NRCS) Snowpack Telemetry (SNOTEL) system to forecast streamflows and direct resource planning strategies across the western United States. In the Columbia River Basin, where 60-70 percent of the region's power is generated through large-scale hydroelectric projects and several species of native anadromous fish are listed as endangered, the accuracy of streamflow forecasts directly impacts ecosystem and economic health. Anecdotal reports suggest that forecasts of late spring flows in alpine basins are frequently incorrect, perhaps due to imprecise measurements of basin-wide snow water equivalent (SWE). Here the spatially distributed SnowModel is used to distribute SWE across the McKenzie River watershed, a sub basin of the Columbia. A binary regression tree classification system is then used to explore relationships between spatially distributed SWE and a range of independent physiographic variables. We use this information to identify optimal locations that are representative of basin-wide SWE, at different times of the snow season. We also explore representative locations of SWE for different climate change scenarios The information provides a basis for identifying present-day and future SNOTEL and climate station locations for monitoring water resources.