HOBOKEN, N.J. – Springtime floods caused by river ice have long threatened northern communities, damaging infrastructure, cutting off water supplies and putting lives at risk. Now, a CIROH-funded project is helping to change that.
Researchers at the Stevens Institute of Technology, led by Dr. Marouane Temimi, have developed a cutting-edge system that uses satellite imagery and machine learning to monitor river ice conditions in near real-time. The technology provides faster, more accurate detection of river ice—offering crucial data to agencies like NOAA and the National Weather Service.
“Before this system, we relied heavily on manual observations or satellite images that often missed narrow rivers and lacked the frequency needed to inform flood response,” Temimi said. “Now, we can deliver a much clearer, faster picture of what’s happening on the ground.”
The system blends high-resolution radar and optical satellite data with machine learning algorithms to generate detailed ice maps that are publicly accessible. It has already proven successful during testing on the Saint John River in Maine, where it helped simulate ice conditions and predict potential flooding more effectively.
The research team worked closely with partners at NOAA, CIROH and forecasting centers to ensure the system’s usability in real-world scenarios. That feedback loop helped refine the platform and integrate it into CIROH’s cloud infrastructure, where it is now being scaled for broader use.
Looking ahead, the team is focused on expanding coverage to narrower rivers, enhancing radar-based texture detection and applying the same technology to monitor lake ice. Long term, they aim to build a nationwide operational system with real-time alerts and predictive modeling capabilities based on historical ice records and machine learning.
“River ice can have a significant impact on flood potential in the Northeast United States. River forecasters use satellite information to verify the presence of ice in remote locations to inform the river forecasts we issue,” said Paula Miano, hydrologist-in-charge Northeast River Forecast Center. “The tools created by Dr. Temimi and his team have helped forecasters analyze this data for the formation of river ice and to help track the movement during the breakup season. We look forward to the continued collaboration with the researchers at the Stevens Institute.”
At its core, the system is about protecting people. For communities vulnerable to ice-related flooding, these tools can mean the difference between scrambling in a crisis and having time to prepare.
“This is why we do the work,” Temimi said. “We’re turning satellite images into life-saving information.”
About CIROH
The Cooperative Institute for Research to Operations in Hydrology (CIROH) is a NOAA Cooperative Institute hosted at The University of Alabama and supported by the National Weather Service. As a unique national consortium with 28 partner institutions, CIROH unites academic, government and private sector partners to co-produce research that enhances operational water prediction capabilities and informs critical decision-making for issues like floods, droughts and water quality. Through its interdisciplinary research, educational initiatives and outreach programs, CIROH advances NOAA’s water prediction services with new techniques such as artificial intelligence and remote sensing, equips the next generation of water-resource professionals with essential skills and fosters public and stakeholder engagement to support federal agency and private sector needs. CIROH transforms scientific innovations into practical solutions and operational tools that boosts the delivery of actionable water intelligence, increases effectiveness of water resources management, promotes community resilience and informs decision-makers across the United States.