NASA Satellite Data Powers Machine-Learning River Forecasts Amid Western Snow Drought
NASA’s space-based observations are feeding machine-learning models that guide river-flow forecasts for Washington’s hydroelectric system during the ongoing western snow drought.
The lingering 2026 snow drought across the western United States has spurred the use of NASA’s satellite-derived snow-cover and vegetation data in Upstream Tech’s HydroForecast system. Updated every two hours, the model blends weather forecasts, river gauges and NASA’s VIIRS and MODIS products to predict flow on the Cowlitz River, supporting Tacoma Power’s hydroelectric operations that serve over 151,000 homes. After an unprecedented inflow surge in December 2025 and a drought emergency declared in April, the utility relied on short-term forecasts to anticipate water arrivals and on seasonal forecasts to keep reservoirs higher for the dry summer.
NASA Water Resources program manager Erin Urquhart highlighted the agency’s mission to turn space observations into actionable information for flood control, power generation and water stewardship. Similar NASA data collaborations extend to the USDA’s Natural Resources Conservation Service, NOAA’s Colorado Basin River Forecast Center and the Bureau of Reclamation, and NASA now contributes directly to the weekly U.S. Drought Monitor.
Why it matters
Space-derived river forecasts help utilities balance electricity, water supply and flood safety during extreme drought.
In this story
