In July 2026, the world was looking forward to World Cup soccer. Yet with wildfire smoke from Canada spreading as far south as New Jersey, which recorded some of the worst air quality across the globe, it was not the games that everyone was talking about. Wildfires are costly economically, socially and environmentally. According to the United Nations Office for Disaster Risk Reduction, the US total cost of wildfire damages in 2025 alone was over 200 billion.
Understanding, predicting and managing wildfires is therefore critical. The Canadian Fire Weather Index (CFWI) system was developed to do just that in the Canadian boreal and temperate forests. It has become one of the primary tools for wildfire management in many areas around the world. However, not all environmental factors that are used in the index transfer effectively to other geographical areas. Scientists from the US's Argonne National Laboratory have been studying how to recalibrate the CFWI focusing on Southern California, an area prone to destructive wildfires. This area is dominated by a chaparral landscape, an area where broad leafed evergreen shrubs, bushes and small trees predominate. It typically has hot dry summers and during the mild winters is often affected by the offshore Santa Ana winds. In two of Southern California's worst fire years, 2003 and 2007, they experienced strong Santa Ana winds which significantly increased the spread and destructive nature of the fires. Neither this type of wind nor vegetation is accounted for in the CFWI.
When describing their research, Argonne scientist and paper's first author Brandi Gamelin said their "study sought to examine whether the CFWI could be adapted to better represent Southern California's unique fire-weather conditions. Specifically, we evaluate its ability to identify periods of elevated wildfire risks during Santa Ana wind events in San Diego County." Their results on recalibrating the CFWI for Southern California were published as a cover article in Advances in Atmospheric Sciences . The paper is part of a topical collection on the 2025 LA wildfires .
The researchers focused on including metrics that would account for conditions in San Diego County that may affect wildfire risk including, but not limited to, the strong Santa Ana winds, the effect of winter dry chaparral vegetation that fuels fire, low humidity and low soil moisture. Specifically, they incorporated relative measurements, maximums and minimums instead of the static noon time inputs of the CFWI. These included daily maximum temperature, daily maximum wind speeds, and several more specific humidity measurements, daily minimum relative humidity and relative humidity thresholds.
The new metrics were added to different models to test their utility. Data from 2000-2009 was added into the models and the predictions were then compared to the actual wildfires that occurred over this period. One metric they used in their recalibrated CWFI was wind speed. When threshold is applied only to wind speed, the fire danger was overestimated by the model when compared to the actual fires. This was likely due to the fact it did not differentiate between the Santa Ana winds and other winds. "To focus on the Santa Ana winds, they added a modifier when zonal wind are from the east and relative humidity are below 20th percentile ", that is the wind speed for local winds that flow parallel to lines of latitude. In this area that would be the offshore Santa Anas. Incorporating the more specific measure for this meteorological driver decreased the probability of over-estimation of fire danger during non-critical wind events.
The recalibrated CFWI was less robust in areas that were urban. "This reduced sensitivity likely reflects the influence of complex fire ecology and wildland–urban interface dynamics, suggesting that meteorological recalibration alone may be insufficient. These results highlight the importance of integrating meteorological, ecological, and anthropogenic factors in regional fire-danger assessment," said Rao Kotamarthi, Argonne scientist and co-author.
Finally, the researchers found that there was strong seasonal and spatial variability further strengthening the utility of recalibrating the CFWI to those specific local conditions.
"This work demonstrates that physically motivated regional recalibration improves the use of the CFWI to wind-dominated fire regimes, providing a more reliable tool for operational fire danger assessment and wildfire preparedness in locations prone to extreme wind-driven wildfires," Brandi Gamelin said.
The recalibrated index still requires more testing by comparing the model to longer term observational data sets. One exciting avenue to work towards is extending the investigation to similar regions such as Northern California with their Diablo winds or the Rocky Mountains with their Chinook winds. Opening further opportunities to work on recalibrating the CFWI in areas that share some similarities but would still require regional specific metrics.
Other contributors include Connor Aghili and Jiali Wang from the Environmental Science Division, Argonne National Laboratory, USA; and Corrine Decham, Jordan Branham, and Rao Kotamarthi from Decision and Infrastructure Sciences Division, Argonne National Laboratory, USA.
The San Diego Gas & Electric to Argonne National Laboratory through the U.S. Department of Energy contract DE-AC02-06CH11357, Argonne's Laboratory Computing Resource Center (LCRC), Argonne's Leadership Computing Facility (ALCF), supported this research.