Groundbreaking atmospheric river forecasting collaboration crosses the finish line
New system delivers more accurate forecasts of rainfall, runoff, and flooding.
A cutting-edge forecast system developed by NOAA, Colorado State University, and Scripps Institution of Oceanography at the University of California San Diego is now fully operational and providing improved precipitation forecasts of landfalling atmospheric rivers in the Bay Area.
When big storms approach California, accurate and timely precipitation estimates become critical for making decisions regarding public safety, infrastructure operations, and emergency response preparations. Atmospheric rivers provide up to 50% of total annual precipitation in California but also cause more than 80% of flood damages. Forecasting rainfall and runoff in California’s complex coastal terrain has been challenging because as these storms move ashore, heavy rains form at low elevations that sit beneath the level visible to the standard National Weather Service (NWS) radar network.
To solve this problem, scientists and engineers with NOAA Research, Colorado State University’s Cooperative Institute for Research in the Atmosphere (CIRA), and Scripps’ Center for Western Weather and Water Extremes (CW3E) developed the Advanced Quantitative Precipitation Information system (AQPI). AQPI is a new regional radar and advanced rainfall forecast system that gives emergency managers and water agencies in nine San Francisco Bay area counties more accurate and timely warnings about potential flooding and runoff impacts from heavy rainfall events.
On March 26, the San Francisco Public Utilities Commission held a ceremony to mark the handover of the system to Scripps and a local partner agency committee, which will run the system.
“AQPI represents the successful transition of cutting-edge atmospheric research into operational capability,” said NOAA Acting Chief Scientist Steve Thur, Ph.D., the Assistant Administrator for NOAA Oceanic and Atmospheric Research. “It demonstrates how sustained scientific innovation can directly improve hazard preparedness, water management, and community resilience.”
How the system and dashboard were created
NOAA Research and CIRA scientists worked together to design, build, and deploy supplemental radars, building an experimental system that could be tailored for operational use by water agencies in the San Francisco Bay area. The system’s foundation is an array of five short-range, low-elevation X-Band radars strategically located to fill radar gaps over populated, flood-prone urban areas and numerous burn scars throughout the region. One longer-range C-Band radar monitors atmospheric rivers from a mountaintop site as storms approach from over the ocean.
The dense radar coverage captures low-elevation rainfall that other systems miss, including short-lived events like intense precipitation from narrow, cold-front rainbands that occur during West Coast winter storms.
The AQPI radar data is combined with data from NEXRAD radars operated by NWS to generate detailed precipitation estimates for critical, very short-term forecast windows. These are updated every two minutes to provide real-time “nowcasts” between the hourly operational NWS High Resolution Rapid Refresh (HRRR) weather model runs. The HRRR estimates in turn drive runoff estimates generated by NOAA’s National Water Model and coastal flooding forecasts in and around the Bay coastline by the U.S. Geological Survey’s Coastal Storm Modeling System.
The new AQPI User Dashboard developed by CW3E provides both data visualization and delivery to Bay Area partners. It integrates a variety of observational and forecast data streams including the AQPI radar mosaic, rain and stream gauges, model precipitation forecasts, a radar-based, real-time nowcast, and additional CW3E forecast products. The dashboard functionality allows users to select points or areas to receive custom-formatted precipitation observations and forecasts.
Who is using AQPI?
During the past year, Sonoma, Marin, San Mateo and Santa Clara counties; the Union Sanitary District; and the San Francisco Public Utilities Commission have used AQPI to prepare for flooding and heavy runoff events and to protect water quality by managing combined sewer stormwater systems during rain events. The information also helped officials manage staffing, tailor public messaging and evacuation decisions, and keep the public informed before, during, and after events.
AQPI also provides data vital for managing reservoir levels and informing water supply decisions in a region shaped by both drought conditions and atmospheric rivers.
Initial funding for AQPI was provided by a $19.7 million grant from the California Department of Water Resources Integrated Regional Water Management Program awarded to Sonoma Water and participating Bay Area agencies.
“In the Russian River watershed, accurate forecasting of atmospheric rivers is critical for protecting communities and managing water resources,” said Sonoma County Supervisor David Rabbitt. “AQPI gives local agencies more precise, timely information to improve flood preparedness, response, and post-wildfire debris flow preparedness across the region.”
Within NOAA Research, the Physical Sciences Laboratory and Global Systems Laboratory led the project, overseeing design and deployment of the radar array, developing high-resolution monitoring tools, and running models to better predict storm-driven flooding. The AQPI radar systems were built, deployed, and are currently operated by CIRA, which produces the AQPI radar products that support decision-making by Bay Area agencies.
Future operations to maintain critical data delivery
Starting July 1, 2026, Scripps will lead and manage the AQPI system and the official user interface that delivers radar products and precipitation forecasts. CIRA will conduct operations and maintenance for the radar systems, and NOAA will continue to support research and development.
The AQPI system complements the Atmospheric River Reconnaissance program (AR Recon), which is also co-led by NOAA and Scripps Oceanography. AR Recon collects data to improve forecasts of atmospheric rivers from 1-15 days before a storm hits, while AQPI improves observations and forecasts of heavy rainfall both hours before and during the storm.