.png)
With artificial intelligence’s recent uprising in the past couple of years, it has been used for many things including practical, daily things such as helping write an email, or more technical uses such as helping detect cancer cells and other biomedical applications. But recently AI has been implemented in something new and extremely important, wildfire detection. One place where these new AI’s have started being used is California. California’s main firefighting agency started testing AI with over a thousand mountain-top camera feeds. The AI is trained to detect “abnormalities” and alert emergency services where they can check if it is indeed smoke or something else. The cameras used by the AI were originally monitored by humans but the billions of bites of data that they were required to go over caused a lot of fatigue in the workers. Humans are still required to confirm any smoke sightings but the workload is greatly reduced. Previously workers would be watching multiple screens and cameras at once but now they only have to look at what the AI has flagged as a possible fire. This new implementation of AI has already been helpful. In the middle of the night a battalion chief was alerted by the AI, he confirmed the smoke sighting from his cellphone and was able to have first responders get to the fire. The dispatcher who took his call said that without him being alerted the fire could have very easily become much more of a problem, and likely wouldn’t have been noticed until morning. Not only is the AI reducing workload, it is also reducing the response time for most of the fires. Without it, detecting wildfires relied on 911 calls to report them. Saffers were then needed to confirm the sighting, before finally deploying firefighters and water-dropping planes. Generally, only 1 in 20 911 calls turns out to actually be a wildfire. Often it's just a cloud of fog or someone cooking in their backyard. Using AI to spot fires isn't the only use for AI either, scientists have started using it to help predict fire weather or determine whether conditions are favorable for fire growth using relative humidity, wind speed and direction, mixing heights, and soil moisture. One of the most difficult parts of predicting fire weather is an assessment of fuel in a complex terrain, such as steep slopes, intersecting valleys, or other terrain that could affect the weather on a smaller scale. Landscape isn't the only thing that affects fire weather either, as predicting fire weather also includes things like predicting lightning strikes, as lightning is one of the leading causes of wildfires. Lighting is even more dangerous when no rain is present as a single strike could quickly and easily create a fire without anything restricting it. So to help forecasters, an artificial intelligence called LightningCast, has been implemented to help predict where and when these disasters could happen. The lightningCast AI model was created by John Cintineo at the University of Wisconsin/Cooperative Institute for Meteorological Satellite Studies (CIMSS) in 2021. The lightning-predicting artificial intelligence has continued to evolve and change and is even being used by the National Weather Service. LightningCast uses two of NOAA's GOES-R satellites and 6,600 pictures it gets from them daily to identify the most likely places for lightning to strike in the next hour. Mike Pavolonis, a physical scientist with the NOAA/NESDIS Center for Satellite Applications and Research said, "AI has the potential to be a game changer in a number of ways, including early fire detection, lightning prediction, forecasting fire spread and behavior, mapping fire perimeters and assessing wildfire risk prior to ignition." LightningCast is only the beginning of implementing artificial intelligence into predicting weather and fire forecasting. The National Environmental Satellite, Data, and Information Service (NESDIS) is currently testing a new AI part of the Next Generation Fire System (NGFS) which is designed to detect fires earlier than any current method of fire detection. The NGFS system is planned to automatically track fires allowing monitoring of fire intensity and smoke production. Forecasters are looking forward to implementing more AI into satellite data for detecting fires earlier and even predicting behavior and spread.
Chan, K. (2023, September 24). The threat of wildfires is rising. so are new artificial intelligence solutions to fight them. AP News. https://apnews.com/article/artificial-intelligence-fighting-wildfires-climate-change-62002360f2e3be32ccc7df020239efea
Garofalo, M. (2024, February 6). How scientists are using artificial intelligence to predict wildfires. Space.com. https://www.space.com/how-scientists-are-using-artificial-intelligence-to-predict-wildfires
Owen Dustin is a sophomore at Poudre High School. This is his first year writing for the Poudre Press. At school, he is on the Alpine Robotics team and the unified flag football team. Outside of school he likes to mountain bike, play video games, and hang out with friends.