The Virginia Institute of Marine Science (VIMS) and its partners have launched an upgraded Chesapeake Bay Environmental Forecasting System (CBEFS), adding new artificial intelligence-powered harmful algal bloom (HAB) forecasting tools and extending water quality forecasts to help resource managers, water users and the public better prepare for changing conditions.

Supported by NOAA, the updated platform expands operational forecasts from two days to five days and incorporates a wind-wave model to improve storm forecasting. The system also includes automated alerts for harmful algal blooms and marine heat waves, providing earlier warning of conditions that can affect water quality, aquatic ecosystems and coastal communities.

Researchers developed the Chesapeake Bay's first machine learning-based forecasting model for Prorocentrum minimum, one of the estuary's most common harmful algal bloom species, and additional forecast models for six other HAB groups. By combining environmental observations with artificial intelligence, the system produces a unified forecast designed to improve bloom prediction accuracy.

The enhanced platform also includes a digital atlas of Chesapeake Bay conditions based on model data from 1985 through 2025, providing historical information on 27 physical and biogeochemical variables to support water quality research and management.

According to VIMS, the upgraded forecasting system will provide more actionable information for anglers, shellfish growers, coastal managers and scientists by improving real-time monitoring of dissolved oxygen, water quality and bloom conditions. The project was led by VIMS in partnership with the University of Maryland, FlowWest and NOAA through the U.S. Integrated Ocean Observing System's Coastal Ocean Modeling Testbed program.

For stormwater and watershed managers, the enhanced forecasting tools offer improved situational awareness of water quality conditions and algal bloom risks that can be influenced by nutrient runoff, helping inform monitoring efforts and long-term watershed management decisions.