Heng Huang Receives NSF Funding for Wildland Fire AI Research

Each year, wildland fires burn across millions of acres in the United States, threatening communities, wildlife and natural resources. Forecasting where and when those fires will spread remains one of the most difficult challenges for scientists and emergency responders. At the University of Maryland, researchers are turning to artificial intelligence to close that gap.

Leading that effort is Heng Huang, a professor in the Department of Computer Science, who has received $1.86 million in funding from the National Science Foundation (NSF) to lead a project aimed at advancing AI-driven methods for detecting and predicting wildland fires.

The initiative, titled "Collaborative Research: Advanced AI Framework to Improve Understanding and Prediction of Wildland Fire," is a collaboration between the University of Maryland and American University, with the University of Maryland serving as the lead institution.

The research aims to address the complexity of fire dynamics, which limits traditional modeling approaches in their ability to forecast events. Huang and his team will develop large-scale AI and machine learning algorithms capable of analyzing diverse sources of geoscientific data, including satellite observations, atmospheric records, fuel information and historical fire data. The goal is to develop a framework that enhances detection accuracy and forecasting capabilities while maintaining robustness in real-world conditions, where data availability is often incomplete.

“I am excited for this project to develop and apply advanced AI techniques to address the challenging wildland fire prediction and prevention problems, which could make a large impact on human communities, wildlife and the environment,” said Huang, the Brendan Iribe Endowed Professor.

The project will integrate multiple streams of information to provide a more comprehensive view of fire conditions. Data will be collected from geostationary and low-Earth orbit satellites, as well as reanalysis of atmospheric trends and ground-level fire and surface characteristics. The framework will feature an interpretable multimodal transformer to integrate these diverse sources, a time-series deep learning model to forecast fire events and a federated learning platform to support collaboration across research groups.

By combining multimodal and longitudinal datasets, researchers aim to capture not only current fire activity but also patterns that could inform predictions of future events. This approach, Huang noted, is designed to leverage both large-scale computational power and innovative algorithm design.

Another aspect of the project is the development of an open-source, integrated dataset that can be shared with the broader research community. By releasing these resources, the team aims to assist other researchers in benchmarking their methods and accelerate advancements in wildland fire science.

Partnership with federal, state and local agencies will also play a central role. Organizations such as the National Aeronautics and Space Administration (NASA), the U.S. Forest Service and the National Park Service will provide input to ensure that the tools align with operational needs for fire tracking and management. The collaboration aims to translate academic research into systems that can be used on the ground during fire seasons.

“The emerging AI techniques offer a promising avenue for providing accurate wildland fire understanding and forecasting,” Huang said. “By analyzing real-time satellite signals and reanalysis data, our advanced AI technologies can improve forecasting accuracy. With accurately forecasting wildland fire events and implementing preparedness and mitigation strategies in advance, societies can better shield themselves against the devastating impacts of such disasters.”

—Story by Samuel Malede Zewdu, CS Communications