WRFx; An Open-Source Wildfire Forecasting System

As contending with wildfires becomes necessary for many organizations across the West, an open-source and interpretable forecasting system for wildfires becomes an integral part of daily operations. The Wildfire Interdisciplinary Research Center (WIRC) has developed, evaluated, and validated a forecasting system built on the coupled fire-atmosphere model WRF-SFIRE (Mandel et al., 2014) and an assortment of Python code that can produce forecasts in real time for anything from smaller prescribed burns to >100,000 Acre wildfires. The system is fundamentally built on assimilating real-time fire perimeters from NIFC and satellite hotspot detections from various satellites that, through a spin-up algorithm (Kochanski et al., 2023), initialize a WRF-SFIRE simulation with the most current fire information available to the public. These forecasts are then run through the wrfxpy Python system that automates all WPS and WRF processes and simplifies options through a single job file. As this system is all open source, the wrfxpy system can be intuitively and easily modified for each user’s needs. Finally, through our wrfxctrl and wrfxweb systems, users can have forecast outputs automatically and incrementally sent to a visualization server where many atmospheric and fire-related variables are immediately available to be plotted. This forecasting system has been well-tested by the team at San José State University and, while still under development, is ready to be used operationally by those without the need or means for paid forecasting systems.

Previous forecasts can be found at http://demo.openwfm.org under the SJSU user.

Links to GitHub repositories:

WRF-SFIRE: https://github.com/openwfm/WRF-SFIRE

wrfxpy: https://github.com/openwfm/wrfxpy

wrfxweb: https://github.com/openwfm/wrfxweb

wrfxctrl: https://github.com/openwfm/wrfxctrl