NASA Satellite Data Fuels Machine-Learning Forecasts for Western Water Management

• NASA satellite data is powering machine-learning models to help water managers and utilities navigate extreme water conditions.
• During the 2026 western U.S. snow drought, Tacoma Power used AI-driven river-flow forecasts to manage hydroelectric operations on the Cowlitz River.
• The integration of space-based observations helps utilities balance power generation, flood prevention, and water supply stewardship.

The 2026 Water Extremes

• A warm 2025-26 winter led to record-low mountain snowpack, with levels reaching only 20% to 50% of normal.
• Precipitation often arrived as rain rather than snow, causing rapid runoff instead of steady spring melt.
• A December 2025 atmospheric river event triggered one of the largest one-day inflow surges on record for Tacoma Power’s Cowlitz River project.
• By April 8, 2026, Washington state declared a drought emergency for all watersheds.

Technology and Operations

• Upstream Tech’s HydroForecast tool combines satellite data from NASA (via MODIS and VIIRS instruments) with weather models to predict river flow every two hours.
• The machine-learning models use historical data on snow cover and vegetation to improve prediction accuracy in areas with limited ground monitors.
• Tacoma Power utilized short-term forecasts to prepare for storm-related surges and seasonal models to optimize reservoir levels for summer drought conditions.

Broader Application

• NASA partners with various agencies to incorporate satellite data into water-supply forecasting, including:
  • The U.S. Department of Agriculture for groundwater and snow information.
  • The National Oceanic and Atmospheric Administration for Colorado Basin modeling.
  • The Bureau of Reclamation for California’s San Joaquin Basin reservoir operations.
• NASA became a formal partner of the U.S. Drought Monitor in 2026, contributing to the weekly assessment of national drought conditions.