LSU researchers are testing a new AI-powered flood prediction system in Ascension Parish, creating what officials describe as a real-world laboratory for the next generation of flood forecasting.
Led by LSU professor Z. George Xue of the Department of Oceanography & Coastal Sciences, the project combines real-time drainage data, weather forecasts and artificial intelligence modeling to develop a more predictive flood response system. According to Xue, the work is made possible through a close partnership with Ascension Parish government, which not only provides extensive operational data but also contributes its own sensor network and local expertise.
Ascension Parish operates its own monitoring and sensor system and has agreed to share that data with LSU, enabling Xue and his team to train and refine the machine learning model that forecasts water levels. The collaboration pairs parish-level knowledge and infrastructure with LSU’s scientific and computational capabilities. Xue has emphasized that this exchange of trust and expertise is central to the project’s value.
“I think the most interesting is the collaboration itself,” Xue says. “This kind of mutual trust is more interesting than the technology itself.”
The AI tool generates hourly forecasts by analyzing information from more than 30 parish facilities, including flood gates, tilting weirs, rainfall gauges, stream-level monitoring stations, pumps and the Marvin Braud Pump Station, which can move up to 1.35 million gallons of water per minute at full capacity. The system also incorporates data from the U.S. Geological Survey and National Oceanic and Atmospheric Administration to predict local stream levels up to 24 hours in advance.
Ascension Parish Deputy Chief Administrative Officer Pamela Matassa says the project builds on the parish’s investments in modernizing its drainage operations through SCADA and telemetry systems, which centralize critical infrastructure data. Says Matassa: “Our teams have worked to bring critical drainage data into a centralized platform, creating the foundation needed for advanced forecasting tools like this one.”
Currently, the model is undergoing testing and calibration as researchers compare its forecasts with actual drainage conditions. Xue’s team retrains the AI daily using both recent observations and historical operational data. “We are continuously improving the model’s performance,” Xue says, noting that iteration is key to improving forecast reliability.
Officials emphasize that the technology is designed to support, not replace, human decision-making.
“The goal is not to replace human decision-making, but to provide our drainage professionals with better information,” Matassa says. “By combining real-time conditions, weather forecasts and historical data, the AI model has the potential to help us better anticipate how our drainage system will respond to future rainfall events and support more informed operational decisions.”
Looking ahead, the parish hopes to develop a public dashboard that would allow residents to view real-time AI-generated water-level predictions and flood forecasts. Xue has also outlined a broader ambition for the system: expanding it across Louisiana, beginning with parishes in the river basin where flood risk and drainage complexity are highest.
“We are excited to see how this partnership can advance flood forecasting and drainage management not only in Ascension Parish but also potentially in other communities across Louisiana,” Matassa says.
