KATHMANDU, April 8: Israeli researchers have developed an artificial intelligence (AI) model that predicts wildfires caused by lightning. As climate change triggers extreme weather patterns, lightning incidents have reportedly increased in recent years.
The research team at Bar-Ilan University created the AI model to forecast with approximately 90 percent accuracy when and where lightning is likely to spark a wildfire. They say the machine learning model can be applied globally to predict lightning-induced wildfires and serve as a powerful tool for wildfire prevention and response.
The Israeli Embassy in Nepal stated in a press release on Monday that Dr Oren Glickman and Dr Asaf Samuel from the Department of Computer Science at Bar-Ilan University developed the AI model in collaboration with experts from Ariel and Tel Aviv universities. While building the model, the researchers used satellite-based data from the past seven years along with detailed environmental factors such as local vegetation, weather patterns and topography, to map and predict the global risk of lightning-induced wildfires.
New wildfires in California bring total to over 100 active fire...
They recently published their study in the scientific journal Scientific Reports. The researchers emphasized that the model performs significantly better than traditional fire risk indicators. They integrated satellite data, weather systems, and environmental variables to enhance the model’s accuracy. When they tested it using wildfire data from 2021, the model matched over 90 percent of the incidents. Experts believe the model could revolutionize global emergency response and disaster management.
As climate change accelerates across the globe, it triggers extreme weather events such as lightning, storms, heatwaves, and droughts with increasing frequency and intensity. While human activities cause many wildfires, lightning strikes in remote areas often unleash unexpected and devastating consequences. These lightning-induced wildfires can smolder undetected for days, then erupt into uncontrollable and explosive blazes that challenge firefighting efforts.
In August 2020, lightning sparked a destructive wildfire in Northern California that scorched more than 1.5 million acres of land and claimed dozens of lives. Researchers believe improved forecasting of lightning-triggered wildfires will enable meteorological agencies, fire departments, and emergency planners to respond more intelligently and effectively. These advancements could save lives, minimize large-scale economic losses, and protect vital ecosystems.
Although previous models predicted wildfires caused by human activity with some success, they struggled to assess lightning-induced wildfires, particularly in remote regions.
Although researchers have not yet integrated this AI model into real-time forecasting systems, Dr Samuel considers its development a milestone in wildfire prediction.
“We have reached a crucial point in understanding the complexities of wildfire ignition,” said Dr Glickman from the Department of Computer Science at Bar-Ilan University. “As climate change intensifies, we need new modeling approaches to better understand and predict its effects. Machine learning offers significant potential to advance these efforts.”
