ISREAL: Researchers at Bar-Ilan University have unveiled a groundbreaking artificial intelligence (AI) model that could transform the future of wildfire prediction, particularly in predicting lightning-induced fires. The new model, which boasts over 90% accuracy, represents a significant leap in wildfire forecasting, offering hope for better disaster management and early response.

Led by Dr. Oren Glickman and Dr. Assaf Shmuel from Bar-Ilan University’s Department of Computer Science, in collaboration with experts from Ariel and Tel Aviv Universities, the team utilized seven years of high-resolution satellite data alongside environmental factors such as weather patterns, vegetation, and topography. This comprehensive approach enables the team to predict where and when lightning strikes are most likely to spark wildfires, creating a global model for wildfire risk forecasting.

The new AI model outperforms traditional fire danger indices by integrating data from multiple sources, including satellites and weather systems, to provide a more accurate assessment of wildfire risks. In rigorous testing using wildfire data from 2021, the model achieved an accuracy rate exceeding 90%, setting a new standard for forecasting lightning-induced fires.

As climate change continues to fuel more frequent and intense wildfires, lightning has become an increasingly significant and unpredictable cause of these disasters. The new model’s ability to predict lightning-induced fires, which often begin in remote, hard-to-reach areas, could prove invaluable for fire departments and emergency planners, enabling earlier intervention and potentially saving lives.

While the model is not yet integrated into real-time forecasting systems, its development marks a crucial step forward in wildfire prediction. As Dr. Shmuel points out, “Machine learning holds significant potential to enhance efforts to predict the impacts of climate change.”

The AI model offers hope for mitigating the catastrophic effects of lightning-induced wildfires, a critical issue in the age of climate change. Dr. Glickman adds, “Machine learning provides the tools to revolutionize how we predict and respond to lightning-ignited wildfires, offering insights that could save lives and preserve ecosystems.”

The research, recently published in Scientific Reports, holds promise for global wildfire risk management, providing a powerful tool to help safeguard forests, wildlife, and human communities against the increasing threat of climate-driven fires.