The next time you’re planning on going for a run, check more than just the weather — the air quality might affect your performance too.
At least, this is the advice given in a recently published study by Brown researchers, which aimed to determine whether air quality impacts marathon finish times, according to first author Elvira Fleury ’22 MPH ’23.
“There were some studies out there on (air quality) done in marathons in shorter races, but there had been nothing really comprehensively done,” said Joseph Braun, a professor of epidemiology. A marathon and triathlon enthusiast, Braun originated the idea for the study.
Fleury and the rest of the researchers discovered that a higher concentration of minuscule pollutant particles — less than two-millionths of a meter wide — is associated with an increase in marathon finish times.
“You have to be pretty dang healthy to be able to run (across) a marathon finish line,” Braun said. “If air pollution is having an effect on (marathon runners), it shows that really, there's no limit to who is susceptible to the effects of air pollution.”
According to Allan Just, an associate professor of epidemiology and environment and society who worked on the study, any amount of air pollution negatively impacts human health, even within limits deemed safe.
“There’s a really robust body of literature showing that air pollution is associated with increased mortality (and) readmissions to hospitals,” Fleury said.
Fleury and the rest of the researchers studied nine United States marathons between 2003 and 2019, analyzing over one and a half million male and one million female finish times. The researchers then compared air quality monitor data, when it was available, with the runners’ times.
The researchers found the data through web scraping — creating a code to download data — a task they were not familiar with until Fleury found Gray Bittker ’27, the paper’s second author, on Sidechat.
“I was on Sidechat, actually, and someone needed help with web scraping … and I messaged them,” Bittker said. “I was like, ‘Hey, I know this stuff.’”
After Bittker answered some of Fleury’s questions about web scraping, she asked him if he would like to help with the project.
But sometimes air quality monitor data was not always available to pair with the run time data found by Bittker. So Just created a “spatiotemporal machine learning model” for reconstructing air pollution numbers.
To produce a “best estimate” of air pollution concentrations on a given day, Just entered mile-by-mile markers of marathon routes, looking at longitude, latitude, date, time and weather, he said.
The model reconstructs the air pollution levels using satellite aerosol optical depth — a photo taken from space that measures the density of pollutants by determining how much sunlight does not reach Earth’s surface, according to Fleury.
While Just was creating the model, he compared predicted pollution levels with measured values where he did have the data to ensure accuracy.
While the study focused on a narrow subset of people, it has implications for the broader population, Braun said.
“We often think about air pollution as impacting people who already have a chronic condition,” Just said. “We don’t think about how air pollution impacts people who are not experiencing bad health. This (study) is novel in that sense.”
He noted that the last few decades have seen a decrease in air pollution in the United States, largely due to the Clean Air Act, a 1970 law that regulates U.S. air pollution.
“Our model goes back to 2003, and air pollution has been improving since then, which is a success story,” Just said.
While air quality has been improving, global temperatures are projected to rise, Braun pointed out. Going forward, he hopes to study how temperature increase will impact marathon race times.
Leah Koritz is a senior staff writer covering science & research. Leah is from Dover, Massachusetts and studies Public Health and Judaic Studies. In her free time, Leah enjoys hiking, watching the Red Sox and playing with her dog, Boba.
