Researchers Develop First-of-Its-Kind AI Model to Forecast Water Contamination
Scientists have created a groundbreaking artificial intelligence (AI) tool capable of predicting water quality changes across the United States, marking a major breakthrough in environmental monitoring.
Developed by researchers at the University of Vermont (UVM), this new technology enhances the National Water Model, a government system traditionally used for forecasting stream flow. By integrating AI and real-time sensor data, the tool can now predict potential water contamination, helping water treatment plants, city planners, and environmental agencies better prepare for disruptions.
“With the first-ever application of the National Water Model to forecast water quality, we’ve opened a new window that can really benefit the country as a whole moving forward,” said Dr. Andrew Schroth, the study’s lead researcher and a professor at UVM.
How the AI Tool Works
This innovative system was tested in New York City’s water supply network, one of the largest unfiltered water supply systems in the world. The AI-enhanced model was used to predict turbidity levels—an indicator of water clarity that determines how much sediment is present. High turbidity can make water unsafe and requires treatment adjustments.
New York City sources about 40% of its drinking water from the Ashokan Reservoir, fed by Esopus Creek in the Catskill Mountains. During storms, the creek erodes glacial sediments, leading to increased turbidity. When sediment levels exceed safety limits, the city must alter its water management strategies, which can be costly and inefficient.
The AI tool processes real-time data from sensors monitoring water flow and sediment levels. It then predicts when turbidity levels will rise, allowing authorities to take preventive actions such as adjusting filtration processes or sourcing water from alternative reservoirs.
A Game-Changer for Water Management Nationwide
Beyond New York, this AI-driven model has national implications. Across the U.S., cities and rural communities face water quality challenges from pollutants, industrial runoff, and climate-related changes such as increased storms and droughts.
By applying this forecasting system, local governments, farmers, and water treatment facilities can anticipate contamination risks before they occur.
- Public water suppliers can predict and mitigate contamination threats, ensuring safe drinking water.
- Farmers can plan fertilizer use based on water quality forecasts, reducing pollution.
- Coastal authorities can close beaches in advance if an algal bloom is predicted, protecting public health.
“Turning a streamflow forecasting tool into a water quality forecasting tool paves the way for increasingly available forecasts to serve community needs,” said Dr. John Kemper, co-author of the study from Utah State University.
Future Applications of AI in Water Quality
The success of this model opens the door for further AI applications in environmental science. Researchers plan to expand the system to track additional water quality indicators such as phosphorus, nitrogen, and chloride levels. The technology could help communities nationwide adapt to climate change-driven water issues and improve infrastructure planning.
By integrating AI into water management, the U.S. moves one step closer to ensuring clean, safe water for all—turning scientific innovation into real-world solutions.
