Improving Tools for Predicting Wildfires

October 27, 2017

By Rebecca Fowler

The Whittier fire in Santa Barbara County, California forced thousands of people to evacuate and consumed more than 18,000 acres in July 2017. Photo: U.S. Forest Service

The complex interactions among fire, climate, weather, vegetation, and humans make it difficult to predict where wildfires might start and how they may spread. To discuss gaps in current knowledge and identify areas where advances in fire prediction can be made over the next decade, the Columbia University Initiative on Extreme Weather and Climate, with support from the Center for Climate and Life, hosted the Fire Prediction Across Scales conference from October 23 – 25, 2017 in New York City.

Warmer and drier conditions — a result of climate change — are increasing the likelihood of wildfires occurring, along with their size, intensity, and duration, making fires more destructive and difficult to fight. Research by Park Williams, a Center for Climate and Life Fellow, shows that human-induced climate change doubled the area affected by forest fires in the American West over the last 30 years. And, as evidenced by the devastating fires in California this summer and fall, these events lead to loss of life and property and come at a steep cost to taxpayers.

The fire prediction conference, held at Columbia University, brought together scientists, fire modelers, public land managers, insurance industry representatives, and many others who share the goal of advancing the science of fire prediction. Their work and knowledge exchange will help create more accurate computer models, improve fire management practices, and mitigate the threats fires pose to human lives and livelihoods.

Williams, a bioclimatologist at Lamont-Doherty Earth Observatory, was one of the seven organizers of the conference. He explained his motivation for participating in an email: “Given that fire modeling is a relatively new science, there are many methods emerging among many groups, and this meeting served the purpose of putting these groups in touch.”

Speakers included Center for Climate and Life Fellow Andrew Robertson, a senior research scientist at the International Research Institute for Climate and Society (IRI), who presented his research in a session on seasonal fire prediction.

“The IRI and other centers have been issuing seasonal climate predictions routinely since the late 1990s, but prediction on the subseasonal range — which fills the gap between weather forecasts and seasonal forecasts — is an emerging area of research,” Robertson told the IRI. “This timescale could be particularly relevant to fire risk early warning and prevention, giving managers forecasts on a range of timescales in a ‘seamless’ way, from several seasons ahead, down to a few days ahead.”

Williams wrote that he was especially pleased by the conversations and new relationships the conference sparked among fire managers and scientists who are attempting to model fire behavior using computers.

“The discussions exposed a lot of opportunities for fire modeling to be used for pro-active land-management measures,” he said. “I think fire modelers can help fire managers get the most bang for their buck by helping identify areas where a specific fuels treatment would be very effective in terms of reducing fire danger for an area or community of interest.”

Visit the conference website to check out the full program; follow #CUFire2017 on social media; and read IRI’s post, “The Science of Fire Prediction.”

Related: In the wake of this fall’s wildfires in Oregon and California, Williams has helped make sense of the rise in fire activity in the American West.