Abstract:
The 2025 wildfires in Southern California were amongst the most destructive wildfires on record to impact the region. Given the significant damage caused by these wildfires, there have been renewed calls for improvements to fire weather prediction and monitoring.
These wildfires, like many of the most destructive wildfires in this region, were fueled by Santa Ana Wind (SAW) events, which are episodic downslope wind events that significantly heighten wildfire risk across Southern California. And while the link between SAW events and wildfires is well established, a comprehensive, event-based climatology based on observed meteorological conditions remains underdeveloped. In this talk, I will present results from a database of 1,442 SAW events spanning 1950-2021, with a focus on their meteorology, their relation to wildfire occurrence, and our ability to predict wildfire risk given particularly dangerous SAW conditions using a machine learning model.
This catalogue of SAW events offers a robust and event-level foundation for understanding these events in the region and their relationship with wildfires. Ongoing work will seek to develop spatially-refined risk information with our model and explore our model’s performance for the 2025 Los Angeles wildfires.