Earlier this year, wildfires in southern California killed 30 people, destroyed more than 18,000 homes and burned more than 57,000 acres. The fires were a stark reminders of the threat of worsening climate change, and the increased likelihood of future devastating fires. With these fires comes smoke, which has long-term health effects for the people exposed to it – whether they are close to the source, or many miles away.
A Harvard atmospheric modeling team has created an online platform that could help communities identify areas in need of controlled burns or other fire management strategies, with the goal of preventing future uncontrolled fires and, more specifically, reducing smoke exposures.
The research is published in Environmental Science & Technology and is led by Loretta Mickley , senior research fellow in chemistry-climate interactions at the John A. Paulson School of Engineering and Applied Sciences (SEAS) and leader of the Atmospheric Chemistry Modeling Group . Two alumni from Mickley's group co-led the effort: Tianjia (Tina) Liu, now an assistant professor at University of British Columbia, and Makoto Kelp, now a postdoctoral fellow at Stanford University.
"We want to know not just where catastrophic fires are likely to occur, but which fires will generate the most smoke, and affect the most people downwind," Mickley said. This is important given the large numbers of people who die prematurely from breathing wildfire smoke, particularly in the West, Mickley continued, even as decades of progress have been made in reducing urban air pollution on the coasts.
Targeted land management can reduce smoke exposures
Focusing on the wildfire-prone Northern California region, the researchers estimated that in 2020, targeted land management in the 15 highest-risk areas (about 3.5% of the region) could have reduced total smoke exposures by as much as 18%. They also estimated that, following the 2020 western fire season, 36,400 people died from complications associated with breathing the very fine particles found in smoke. These particles, known as "PM 2.5," have diameters of less than 2.5 microns and are a potentially toxic form of air pollution that disproportionately impacts vulnerable people, including those with asthma, heart conditions, and the elderly.
The researchers' new user-friendly application, SMRT-Flames , allows fire managers and policymakers to assess potential wildfire-related smoke exposure across regions and to target land management practices accordingly. Most wildfire tools attempt to predict fire risk, but SMRT-Flames also explicitly considers smoke exposure across populations. The app currently focuses on wildfire management in Northern California, but it can be expanded to other areas.
"With our methodology, you can consider hypothetical scenarios and plan prescribed fires to reduce smoke exposure over an entire region, not just the immediate area where that prescribed fire is happening," Liu said.
The research leverages a Harvard-developed computer model called GEOS-Chem . The model combines meteorological, chemical, and geophysical data to predict how fires would behave across various regions and how smoke would pool and disperse.
Prescribed burns are efficacious
SMRT-Flames underscores the efficacy of prescribed burning, a land management strategy that involves intentionally setting smaller fires to prevent larger, catastrophic events in the future. Prescribed burning has not been widely practiced in the West, but a growing body of research points to the benefits of controlled burns for culling accumulated underbrush.
"This idea of wildfires being out of control is due to a combination of factors, including climate, but also this legacy of fire suppression where we've actively prevented fires for the last 100 years, which has led to this huge buildup of fuels," explained Kelp.
Previous research by the team found that controlled burns in key zones in northern California, western Oregon, and eastern Washington could have an outsized effect on reducing wildfire smoke exposure throughout the entire western U.S.
The new research paper required methodological innovations and decisions about what factors should account for smoke risk, which is difficult to estimate due to confounding factors like meteorological conditions and differing landscapes. "We had decisions to make during the research process for integrating the knowledge of wildfires from various disciplines, from an atmospheric perspective as well as from a land cover perspective," explained first author and undergraduate research assistant Karina Chung.
Among the team's most surprising insights was showing that people who live at the borders between urban and rural areas, known as the wildland-urban interface, are especially vulnerable to smoke exposure. "That really popped out of our results," said Mickley. "This finding underscores the need to really think about fuel treatments close to or in those areas. We hope that our results provide some justification for considering prescribed fires as a prevention strategy, even close to where people live."
The paper was co-authored by Karn Vohra, Dana Skelly, Matthew Carroll, and Joel Schwartz. It was supported by the NOAA Climate Program Office's Modeling, Analysis, Predictions, and Projections Program under Grant No. NA22OAR4310140 and also by NOAA Climate and Global Change Postdoctoral Fellowships to Liu and Kelp.
