Scientists have developed a groundbreaking artificial intelligence model that significantly improves the ability to predict when and where wildfires will ignite, Financial Times reports. The new Probability of Fire (PoF) model, developed by researchers at the European Centre for Medium-Range Weather Forecasts (ECMWF), integrates a broader range of data sources than traditional wildfire forecasting tools, making it a potential game-changer in disaster mitigation.
The PoF model analyzes not only weather conditions—such as heat and dryness, which have long been used to assess wildfire risk—but also factors like the presence of flammable vegetation and human activity, which are often key triggers. By leveraging AI, it can identify ignition hotspots with far greater accuracy than conventional fire danger indices.
“Traditional fire danger indices often fail to pinpoint areas at risk of ignition with enough specificity,” said Francesca Di Giuseppe, the scientist leading the project at ECMWF. “This is where machine learning can help.”
The AI model has already demonstrated its potential. In January, wildfires in Los Angeles killed at least 29 people and destroyed 16,000 homes. While conventional forecasts identified large areas as highly flammable, they failed to accurately locate ignition points. In contrast, the PoF model—trained using 19 different datasets—was able to predict with much greater precision where the fires would likely start.
The findings, published in Nature Communications, come as wildfires become an increasingly frequent threat due to climate change, human activity, and land use changes. A University of Tasmania study last year found that extreme wildfires worldwide have doubled in frequency since 2003, and projections from Californian researchers suggest a further 30% increase by mid-century.
Florence Rabier, director-general of ECMWF, emphasized the importance of improving fire forecasting. “Although fire prediction is a challenging subject, as ignition remains an unpredictable process, agencies in charge of providing information now have access to improved tools to help better protect lives, livelihoods, and ecosystems,” she said.
A key advantage of the AI-powered model is its accessibility. Unlike traditional wildfire models that require massive supercomputing power, the PoF model can be implemented by smaller agencies as long as they have access to high-quality data. “Almost anyone could do this locally because it doesn’t require massive supercomputing facilities,” Di Giuseppe explained.
Experts in the field have praised the model’s approach. Theo Keeping, a wildfire specialist at the University of Reading, noted that the PoF model fills a crucial gap in operational wildfire forecasting. “An important shift here is the integration of the predicted amount and dryness of vegetation into the forecast product,” he said. “These factors are absolutely essential to predicting wildfire but have hitherto been missing from purely meteorological forecast products.”
As wildfires continue to intensify globally, AI-driven predictive models like PoF offer a promising new line of defense, equipping authorities with the tools needed to respond more effectively to one of the most devastating consequences of climate change.
