Using AI to Improve Wildfire Emergency Response

Due to droughts, dry forests, and environmental conditions, many wildfires have been destroying communities, leaving hundreds displaced. There have been reports of fires in California, British Columbia, Canada, Louisiana and throughout North America this year alone. Most recently, wildfires in Maui caused almost 100 fatalities due to the lack of maintenance of forestry, downed utility power lines and strong winds, according to AP News. The U.S. National Institute of Standards and Technology (NIST) awarded a grant to Xilei Zhao, Ph.D., an assistant professor, and her team to continue building a platform that will use real-time sensing data and state-of-the-art artificial intelligence to facilitate real-time decision-making to mitigate the impacts of wildfires.  

Dr. Zhao, her lab, and colleagues from other universities built the Wildfire Emergency Management (WEM) Platform and used it to monitor a county in California to test its capabilities.  

“The WEM was conceptualized to provide a one-stop platform to aggregate information from multiple sources to assist pre-event planning and real-time decision-making in wildfire emergency management. Communities like Sonoma County in California can directly adopt the platform to better prepare for, respond to, and recover from wildfire emergencies,” Dr. Zhao said.  

She and her team hope the research from this award will advance the methods of analyzing and measuring emergency response to wildfires with the use of artificial intelligence. Most importantly, researchers hope their work will improve wildfire emergency management technologies and support decision-making for wildland-urban interface communities in wildfire events. 

Dr. Zhao and her team are embarking on a solution that hasn’t been previously explored by implementing multi-sourced real-time data to inform wildfire emergency management. Dr. Zhao explains, “Our methodology will produce highly accurate predictions for disaster impacts, population movement, traffic congestion, rescue demand, and other important information that can be directly used to save lives, protect properties, and reduce losses.”  

The funding will help the team enhance features on the WEM Platform. The team will incorporate social media data and large language models to gather real-time information from residents to facilitate emergency response and resource allocation. It will also support new AI technologies that leverage remote sensing data (optical imagery and InSAR) to significantly improve emergency response capabilities and promptly conduct rapid damage assessment after the fire ceased. 

The original WEM Platform was created in 2022 by Dr. Zhao, Ruggiero Lovreglio, Ph.D., (Massey University, New Zealand), Erica Kuligowski, Ph.D., (RMIT University, Melbourne), Thomas Cova (University of Utah), Daniel Nilsson, Ph.D., (University of Canterbury, New Zealand), and Susu Xu, Ph.D., (Stony Brook University) with support from NIST.  

“Through this platform, we will improve wildfire response with emerging multi-modal data and harness deep learning to advance methods of estimating and monitoring wildfire response progress,” Dr. Zhao said.  


Reba Liddy 

ESSIE Marketing and Communications