Collaborative Research: FIRE-MODEL: Developing the scientific basis for reducing Santa Ana wildfire risk

Fast-moving, wind-driven wildfires are among the most destructive natural hazards facing the United States, threatening lives, communities, critical infrastructure, water resources, ecosystems, and the stability of insurance markets. Although weather forecasts can accurately predict many extreme wind events, they cannot reliably determine which will lead to catastrophic wildfires because the interaction between extreme weather and critically dry vegetation remains poorly understood. This project addresses that fundamental scientific challenge by combining advances in atmospheric science, ecology, machine learning, and wildfire modeling to establish the scientific foundation for a new generation of wildfire prediction systems. Southern California provides an ideal natural laboratory because its Santa Ana winds produce some of the Nation's fastest and most destructive wildfires while sharing key characteristics with other wind-driven fires across the western United States. The research will produce new tools to improve wildfire preparedness, evaluate strategies to reduce ignitions, and support more coordinated decision-making among fire agencies, utilities, land managers, local governments, and emergency management organizations. Datasets, modeling tools, and other project products will be made publicly available, and the project will provide interdisciplinary training for graduate students in atmospheric science, ecology, artificial intelligence, and wildfire science. By advancing the Nation's ability to anticipate catastrophic wildfires, protect lives and critical infrastructure, strengthen economic resilience, and sustain U.S. leadership in wildfire science and innovation, this project serves the national interest by promoting scientific progress and improving the Nation's safety, prosperity, and resilience. This project develops a new scientific framework for predicting and managing Santa Ana wind-driven wildfires by integrating high-resolution atmospheric modeling, ecosystem modeling, machine learning, and physics-based fire simulations. The research will (1) create a dynamically downscaled 0.6-km meteorological reanalysis of Southern California spanning 1950–2028 that resolves terrain-driven wind variability during Santa Ana events; (2) develop machine learning models that combine high-resolution meteorology with spatially explicit estimates of live fuel moisture to predict wildfire occurrence and burned area; (3) generate 30-m burn probability maps by coupling ignition models with ensemble fire spread simulations driven by historical Santa Ana events; (4) quantify the effectiveness of alternative wildfire mitigation strategies through counterfactual simulations that isolate the contributions of different ignition sources; and (5) co-develop operational decision-support products with regional stakeholders. The project will produce the first long-term, large-eddy-resolving climatology of Santa Ana winds, new understanding of how weather and live fuel moisture govern fast-moving, wind-driven wildfires, and a predictive framework that can be extended to other regions affected by downslope wind-driven fires. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.