Design and implementation of a snow measurement network using ground-based wireless networks and space-borne measurements in the American River Basin of California

Seasonal snowcover in the Sierra Nevada of California is the primary source of streamflow tributary to the Sacramento-San Joaquin Delta. The blending of strategically placed on-the-ground measurements with broad-coverage satellite and aircraft measurements offers unprecedented estimates of snowpack, soil moisture, vegetation state and energy balance, and snowmelt. A prototype project is planned for the American River basin that includes: i) establishing a network of instrument clusters to provide spatial estimates of these variables across the basin, ii) blending ground measurements with satellite snow-cover data to estimate basin-wide snow water content and more accurately predict runoff, and iii) generating real-time data that will reduce key uncertainties, make snowmelt runoff forecasts more reliable, and inform water resource management decision making. There are currently 12 snow-pillow sites and an additional 21 meteorological stations (above 1,200-m) reporting to the California Data Exchange Center. Adding snow-depth measurements in a 1-km2 area surrounding each snow-pillow site would partially capture the terrain variability in the catchment and provide for representative statistical sampling across physiographic differences. There is not a significant gain in extending instrumentation beyond the 1-km plot at these sites. Instrument clusters will use recent advances in wireless technologies and have 10-20 snow-depth sensors, depending on local variability, temperature, solar radiation, and soil moisture at a subset of the nodes. Results from a prototype instrument cluster at the Southern Sierra Critical Zone Observatory indicate that a spatially distributed wireless network of 23 active nodes, measuring snow depth and soil moisture, successfully relayed data along a 0.75 km transect across complex terrain. Network failures occurred due to drops in radio connectivity, as 20% of the radios failed resulting from a design flaw, which was repaired, but no data was lost as data is collected independent of connectivity with the network. Fractional snow-covered area per 500-m pixel from MODIS can be obtained daily and processed in near-real time, as demonstrated in monthly climate reports generated for the California Department of Water Resources. Initial results from incorporating MODIS SCA into monthly operational water volume forecasts for 2000-2008 has improved the overall model fit for monthly reservoir in-flows.