How scientists use artificial intelligence to predict forest fires

Forecasting fires is not easy: there are so many different types of challenges that forecasters face, both before a wildfire starts and even while it is active. These fires can start anytime and anywhere. So many different factors need to be considered before we can estimate how quickly this could spread and how deeply it could impact our lives and our communities.

“The biggest challenges involve forecasting in complex terrain, ‘fuel assessment’ and communicating the message of risk,” Heather Hockenberry, National Fire Weather Program Manager at the National Oceanic and Atmospheric Association’s National Weather Service said last week. “Complex terrain includes steep slopes, intersecting valleys and other geographic factors that cause small-scale changes in weather. Meteorologists still need to interpret and improve the highly complex weather models to actually make them work when the mountains and valleys change winds and rainfall. “

Weather plays a major role in the development and longevity of wildfires, especially when lightning is involved.

Related: The U.S. government awards NOAA millions for wildfire response research

The millions of strikes we have every year in the United States are not the only contributors to this hundreds injured and just under twenty dead, but they can also cause forest fires within seconds. This dangerous lightning poses a threat, especially if it is not accompanied by rain, where a single strike can ignite a full-fledged fire. Under the right conditions, these fires can often spread quickly, threatening communities without warning. And that’s if there’s any warning at all.

So, to help forecasters, artificial intelligence has been incorporated into predicting these disasters, which continue to be exacerbated by man-made disasters. climate change.

“Probably the best improvements AI can bring are helping people distinguish the unusual from the exceptional,” says Hockenberry. “There are more than 80,000 wildfires each year, which is significantly more than the number of tornadoes or hurricanes in a year. AI and other machine learning will most likely continue to narrow these thousands upon thousands of fires to those that pose the greatest risk to wildlife. our nation.”

AI is not uncommon in the organization; Such techniques have been used in the past to help predict the development of severe weather and hurricanes, detect volcanic eruptions, and even help the aviation community monitor cloud conditions. An example of this is a mechanism called Probably serious, used by National Weather Service (NWS) forecasters to provide more lead time ahead of severe weather as they monitor storm development and issue warnings for both severe thunderstorms and tornadoes. With the success of such applications the LightningCast AI model was developed by John Cintineo of the University of Wisconsin/Cooperative Institute for Meteorological Satellite Studies (CIMSS), and tested in 2021 to advance fire forecasting by providing information that is easy to use and continuously accurate.

Hotspots and smoke plumes from the Mosquito Fire in California as seen from NOAA's GOES-18 satellite on September 13, 2022.

Hotspots and smoke plumes from the Mosquito Fire in California as seen from NOAA’s GOES-18 satellite on September 13, 2022.

“AI is the automation of intellectual tasks normally performed by humans. While human experts excel at extracting information from satellite images, they can only analyze a small portion of the firehose of environmental data, so automation that approaches the skill of human experts is imperative to fully utilize environmental data sources such as satellites,” says Mike Pavolonis, physical scientist at the NOAA/NESDIS Center for Satellite Applications and Research. “LightningCast has continued to evolve and is now routinely used by NWS forecasters for decision support, aviation forecasting and most recently for forecasting thunderstorms in wildfire incidents, as thunderstorms pose a major hazard to firefighters and such storms can be challenging to predict.”

How does this AI model work? It works in conjunction with two of NOAA’s GOES-R satellites, processing data every day from more than 6,600 images generated by their two instruments: the Geostationary Lightning Mapper and the Advanced Baseline Imager. The machine’s trained algorithm can recognize similar, complex patterns to determine where the lightning strike is most likely to occur in the next hour. This is done in part by generating maps in seconds.

This has saved scientists time and resources to make more accurate predictions, as seen here an earlier example from Washington, DC on July 7, 2021.

“AI tools sift through mountains of data, allowing human decision makers to make more informed and timely decisions while freeing up more space time for communicating and coordinating with stakeholders and partners – something AI cannot do well,” Pavolonis said. ‘AI has the potential to be a game changer in a number of ways, including early fire detection, lightning forecasting, predicting fire spread and behavior, fire mapping. fire boundaries and assessing wildfire risk prior to ignition.”

Related stories:

— Satellite images capture forest fires breaking out across Greece (photo)

— Satellites see wildfires raging in northwestern Canada (photos)

—Climate change could change the color of Earth’s oceans

But this is just the beginning for integrating AI into research and forecasting. With continued support from the Bipartisan Infrastructure Law and collaboration with NESDIS Cooperative Institutes, forecasters look forward to continued testing combining AI with satellite and environmental data sources to create a new algorithm for detecting fires early and even predicting behavior and distribution.

“NESDIS is testing a new AI algorithm that is part of the Next Generation Fire System (NGFS) that can be applied in many different ways satellites and is specifically designed to detect fires earlier than existing satellite methods,” Pavolonis said. “The NGFS also automatically tracks fires, allowing for near-continuous monitoring of intensity and smoke production and will be evaluated by operational users during upcoming NOAA Fire Weather Testbed experiments. “

Leave a Comment