Pandemic air quality due to weather, not just lockdowns, study finds

"PM_2.5 is the leading global environmental determinant of longevity. It is a key pollutant of concern for health," said Randall Martin, the Raymond R. Tucker Distinguished Professor in the Department of Energy, Environmental & Chemical Engineering in the McKelvey School of Engineering at Washington University in St. Louis.

New research from Martin's lab, in collaboration with the Goddard Space Flight Center, California Institute of Technology and Dalhousie University in Nova Scotia, mapped PM_2.5 levels across China, Europe and North America. Using satellite data, ground-based monitoring and an innovative computer modeling system, researchers found mostly slight changes in PM_2.5 -- with one exception.

The majority of changes they found were not driven by lockdown, but by the natural variability of meteorology. Their results were published June 23 in the journal Science Advances.

The meteorological effects that we experience everyday also affect PM_2.5 variability, Martin said.

"The shorter time period, the more susceptible PM_2.5 is to meteorology," he said.

During the pandemic, among the images of overflowing ICUs and empty grocery store shelves, there were some photographic bright spots: before and after pictures accompanied articles proclaiming air quality improved because people were staying home.

The visuals were striking -- both on the ground, where blue skies shone over LA highways, and from space -- data from NASA satellites made clear atmospheric reduction in nitrogen dioxide.

"People automatically started wondering, 'What's the picture for PM_2.5?'" said Melanie Hammer, a visiting research associate in Martin's lab. That was the obvious question not just because particulate matter often comes from the same sources as NO₂, but because NO₂ can form PM_2.5.

"NO₂ is considered a secondary source of PM_2.5," Hammer said. When it's emitted, NO₂ interacts with other chemicals in the atmosphere and can form PM_2.5. A few early studies looked at data gathered from ground monitoring sites, which test the surrounding air, but those ground sites are few and far between and incapable of piecing together a bigger picture.

Only about a fraction of the world's population live in countries that have more than three PM_2.5 monitors per million people. Most of the population live in areas with no monitoring.

"We decided to look again, using a more complete picture from satellite images," Hammer said. The images, provided by NASA, contain data for columns of atmosphere spanning the ground to the edge of space. These data, referred to as aerosol optical depth, related to surface PM_2.5 concentrations using the chemical transport model GEOS-Chem, which simulates the composition of the atmosphere; the reactions and relationships of its different constituents; and the way they move through the air.

Researchers focused on three regions that do have extensive ground monitoring systems in place: North America, Europe and China, and compared monthly estimates of PM_2.5 from January to April in 2018, 2019 and 2020.

When they compared PM_2.5 levels over the three years during the months that coincided with each region's lockdown phases, there weren't many clear signals over North America or Europe.

"We found the most clearly detectable signal was a significant reduction over the North China Plain, where the most strict lockdowns were concentrated," she said.

To figure out whether lockdown was responsible for that signal, and several smaller ones dotted around the areas surveyed, the team ran "sensitivity simulations" using GEOS-Chem, changing parameters to see which scenario most closely matched reality.

They simulated a scenario where emissions were held constant and meteorology was solely responsible for year over year changes in PM_2.5.

"We found that that explained a large part of the differences we were seeing," Hammer said. They also ran a simulation in which they reduced transportation-related emissions and other man-made sources of NO₂, mirroring lockdown, when fewer people were driving and fewer industrial sites were operational.

"On its own, that actually didn't really explain much at all," Hammer said. But combining the two, "That's when the signal over the North China Plain stood out."

Hammer suspects that the change in PM_2.5 levels over the North China Plain was so striking because of how polluted it tends to be in "normal" times. "You're probably more likely to see a larger reduction in a region that has higher concentrations to begin with."

In a way, that highlights a relevant point that may not at first be intuitive: Average PM_2.5 levels have been dropping steadily in North America and Europe. "It's just harder to perturb really low concentrations," Hammer said.

But it also underscores the complex relationship between NO₂ and PM_2.5. Although NO₂ does interact with other atmospheric chemicals to form PM_2.5, the two do not have a linear relationship; twice as much NO₂ in the atmosphere does not necessarily lead to twice as much PM_2.5.

Hammer said, intuitively, she did expect to see more of a reduction in PM_2.5 levels. "It was kind of a surprise that meteorology played such a dominant role.

"Turns out, it's a pretty complex relationship and it doesn't always behave how you would expect."