A team of UC Riverside researchers is working with other analysts to expand and improve an online platform that uses data from past wildfires to provide scenarios of how similar ongoing fires may behave, potentially aiding first responders, campus officials said Thursday.

UC Riverside Assistant Professor Ahmed Eldawy, who teaches computer science, and other UCR researchers are collaborating with analysts from Stanford and Vanderbilt universities to refine wildfire modeling using open-source datasets available via a platform called WildfireDB.

The researchers have compiled millions of data points covering the period 2012-17, examining the patterns of numerous wildfires in the continental United States.

“WildfireDB is the first comprehensive and open-source dataset that relates historical fire data with relevant covariates, such as weather, vegetation and topography,” Eldawy said.

He and co-researchers Sam Singla and Vinayak Gajjewar used a UCR-developed system called Raptor to input relevant information captured from satellites, including terrain and weather activity, into WildfireDB, along with details from past wildfires, enabling them to construct datasets that could predict how new wildfires may behave, according to UCR.

“Predicting the spread of wildfire in real time will allow firefighters to allocate resources accordingly and minimize loss of life and property,” Singla said.

Firefighters anywhere in the country will be able to access WildfireDB and plug in data connected to their particular event to develop models that offer insights as to how fires may spread, according to the researchers.

The platform is a work in progress, and the success of turning it into a potential life- and property-saving tool is detailed in a paper authored by the researchers titled, “WildfireDB: An Open-Source Dataset Connecting Wildfire Spread with Relevant Determinants.”

Leave a comment

Your email address will not be published.