NEXT GENERATION OBSERVATION AND MODELING SYSTEMS

Atmospheric Rivers (AR) transport large volumes of water vapor outside the tropics and, when making landfall, produce large quantities of rain that replenish aquifers, contribute to beneficial increases in snowpack, yet can cause flooding and significant damage to property and lives. Accurate forecasts of precipitation during landfalling ARs are critical because of their large role in water supply and flooding along the US west coast. The intensity, location, and duration of atmospheric rivers are poorly forecast in all currently operational numerical weather prediction models beyond forecast Day 7, resulting in a significant decision support services gap. During Phase-I we used Machine Learning (ML) technologies to develop and demonstrate a robust decision support approach—with a viable path from research to operations. The resultant near real-time LEARN2 tool integrates Numerical Weather Prediction (NWP) forecast guidance products, remotely sensed satellite observations, and Sub-seasonal-to Seasonal (S2S) teleconnection indices to demonstrate the theoretical basis for a family of decision-support products. Our Phase-II SBIR project goals are to remove remaining biases, systematically refine and operationally validate the ML technique across a decade of observations and extend the state of science to deliver skillful landfalling ARs forecasts beyond Week 1 under realistic operational conditions.