Researchers at San Jose State University say they used artificial intelligence in a recently published study to determine the link between rising temperatures due to climate change and the day-to-day behavior of wildfires in California.

The study, published Aug. 30 in the journal Nature, departs from earlier research on California wildfires that examined the impact of climate change on broad measures of fire activity, such as the total area burned across large regions and entire years.

It revealed that the impact of climate change on historical fires is not uniform. Instead, its influence varies based on individual fires and specific days, hinging on whether rising temperatures and arid conditions cross critical thresholds.

In some instances, climate change has minimal effect, while in others, it raises the risk of extreme daily fire growth — more than 10,000 acres in one day — by more than five times.

The research also showed that stringent emissions reductions, reaching net zero in the 2070s, would result in a 21 percent increase in extreme wildfire days by the end of the century compared to today.

Conversely, if conditions continue to rise until mid-century, not reaching net zero by 2100, the increase would more than double at 45 percent.

What the study revealed

Researchers said these findings highlight the importance of emission reductions and the need for strategies such as improved forest management to control risk.

They added that the study’s implications regarding critical thresholds that trigger extreme wildfire behavior could guide practical decisions to mitigate fire risk, while insights into optimal times and locations for fuel management can inform wildfire prevention strategies.

“With this method, we have developed an operational wildfire intensity forecast that should help inform decisions for fire management agencies, as well as any group trying to minimize exposure to wildfire risk,” study co-author Craig Clements said in a statement.

The research authors said using AI allowed them to determine the relationship between temperature and wildfire behavior directly from the data, eliminating any preconceived notions and allowing for a fresh understanding of how temperature influences fire activity.

A screenshot of the AI model researchers used to determine the relationship between temperature and wildfire behavior. The study provides a fresh understanding of how temperature influences fire activity. (SJSU Wildfire Interdisciplinary Research Center)

The AI model used in the study is now operational and available on the SJSU website.

“We were able to assess climate change’s impact on the day-to-day behavior of individual historical fires without coming at the problem with built-in assumptions about what we think that relationship should be,” study lead author Patrick Brown said in a statement.

Brown is co-director of the Climate and Energy Team at the Oakland-based Breakthrough Institute, a visiting research professor at SJSU, and a member of the university’s Wildfire Interdisciplinary Research Center.

Brown worked alongside Clements, who is the director of the Wildfire Interdisciplinary Research Center, as well as with member Adam Kochanski and SJSU alumnus Holt Hanley, with additional co-authors from UC Berkeley, UC Irvine and the PG&E meteorology and fire science group.