Karen Short, a Forest Service research ecologist who contributed to the study and maintains a historical database of national fire reports, says understanding why they start is essential to preventing them and educating the public. Strategic prevention seems to be working: according to the National Fire Protection Association, house fires in the United States have decreased by almost half since the 1980s.
In 2024, Short expanded her wildfire archive to include more information useful to investigators, such as weather, elevation, population density, and the timing of fire. “We need to have those things captured in the data to track them over time. We’re still tracking things from the 1900s,” she said.
According to Short, wildfire trends across the western United States have changed with human activity. In recent decades, ignition of power lines, fireworks, and firearms have become more common, in contrast to the railroad- and sawmill-caused fires that were once again common.
The study found that vehicles and equipment are likely to be the number one culprit, possibly causing 21 percent of unexplained wildfires since 1992. Last fall, the Airport Fire in California there was just such an event, burning more than 23,000 acres. And a growing number of fires are the result of burning and accidental ignition – whether from smoking, gunfire or campfires – which make up another 18 percent. In 2017, Arizona’s selection of a pair of blue smoke-spewing fireworks for a baby gender reveal party lit the Sawmill Fireburning close to 47,000 acres.
But these results are not definitive. Machine learning models such as those used for the study are trained to predict the likelihood of a given fire cause, rather than proving that a particular ignition occurred. Although the study’s model showed 90 percent accuracy in choosing between lightning or human activity as the ignition source when tested on fires with known causes, it had more difficulty determining exactly which of 11 possible human behaviors was to blame, only achieving it exactly half of the time. time
Yavar Pourmohamad, a data science PhD researcher at Boise State University who led the study, says knowing the likely causes of a fire could help authorities warn people in high-risk areas before a fire actually starts. “It could give people a hint of what’s most important to take care of,” he said. “Perhaps in the future, AI can become a reliable tool for real-world performance.”
Synolakis, the USC professor, says Pourmohamad and Short’s research is important to understanding how risks are changing. He advocates for proactive actions such as burying power lines underground where they cannot be buffeted by winds.
A 2018 study found that fires started by downed power lines—such as the Camp Fire in Paradise, California, that same year—were increasing. Although the authors note that while power lines are not responsible for many fires, they are associated with larger areas of burned land.
“We need to really make sure our communities are more resilient to climate change,” Synolakis said. “As we’re seeing with the extreme conditions in Los Angeles, fire suppression alone doesn’t do it.”
