AI models can predict lightning-guided wildfires with unprecedented accuracy

The new AI (AI) model developed by Israeli researchers promises to revolutionize wildfire prediction, with a special focus on lightning-induced flames that are becoming increasingly common due to climate change. The new AI model can predict where and when a lightning strike is most likely to cause wildfires, achieving accuracy of over 90%.

Dr. Oren Glickman and Dr. Assaf Shmuel of the Department of Computer Science at Bar-Ilan University collaborated with Ariel and experts at Tel Aviv University to map and predict the risk of wildfires that trigger global lightning, using seven years of high-resolution global satellite data, along with detailed environmental factors such as vegetation, weather patterns, and topography. Their research was recently published in a science report.

What makes the research by Dr. Glickman, Schmuel and their colleagues so important is their ability to predict lightning-induced wildfires with incredible accuracy. AI models outperform traditional fire hazard indexes by taking a global, data-driven approach. It integrates data from satellites, weather systems and environmental factors to assess the likelihood of lightning-induced fires and overcome limitations in local and data-limiting models.

This model was rigorously tested using wildfire data for 2021 and showed unprecedented accuracy of over 90%.

As climate change accelerates, extreme weather events have contributed to more frequent and intense wildfires, including lightning, hot and dry conditions, and ecosystem changes. Human activity is often responsible for setting fires of many fires, but lightning continues to be one of the most unpredictable and deadly causes, especially in remote areas.

These fires could only spurt into hell out of control before firefighters respond, smoldering people who aren’t for days. The devastating wildfire that destroyed Northern California in August 2020 burned more than 1.5 million acres, claiming dozens of lives, and was caused by lightning strikes.

With improved ability to predict lightning fires, weather services, fire departments and emergency planners can respond faster, smarter, more effectively, save lives and protect ecosystems. The model also addresses key gaps in existing wildfire prediction models. Many models are effective against fires caused by human activity, but they struggle to predict fires caused by lightning.

Although AI models are not yet integrated into real-time forecasting systems, their development takes an important step in wildfire prediction. As Dr. Schmuel points out, “As the impacts of climate change are increasing, new modeling tools are needed to better understand and predict their impact. Machine learning has significant potential to enhance these efforts.”

New machine learning models developed by the team can predict wildfires listed in lightning bolts around the world, providing powerful tools for fire mitigation and response. With the increased risk of climate-driven wildfires, early detection and prediction are essential to protect forests, wildlife and human communities from the devastating effects of these fires.

“We are at a critical moment in understanding the complexities of wildfire ignition,” said Dr. Glickman, Department of Computer Science at Bah Iran University. “Machine learning offers the potential to revolutionize the way we predict and respond to lightning-described wildfires, providing insights that can save lives and sustain ecosystems.”