How Caltech's Real-Time Air Quality Monitor Unraveled Multiple Mysteries
The monitor atop Caltech Hall. Courtesy: John Crounse.
My phone says air quality is good, so I’m headed out for a long run. Caltech Scientists say ‘not so fast’!
By Jon Kelvey
Southern California residents, accustomed to living with smoggy skies and the threat of wildfires like those that devastated the area in January, tend to pay particular attention to the air they breathe. They are not alone, especially as wildfires grow more common across the country. At Caltech, researchers have expanded upon a legacy of studying air quality by building a new on-campus monitoring station that provides real-time information about local air quality. Their longstanding work also illustrates limitations in how many people learn about their local air quality—a question that came into sharp focus during the Eaton fire and other recent blazes.
The research team, led by John Crounse, a member of the professional staff and a lead researcher in the laboratory of Paul Wennberg, Caltech’s R. Stanton Avery Professor of Atmospheric Chemistry and Environmental Science and Engineering, has collaborated with Tom Morrell, Caltech Research Data Specialist at Caltech Library, to create a hyper-local air quality resource at breathe.caltech.edu. Updated by the minute with measurements from a monitoring station installed on the roof of Caltech Hall, the website averages the station’s data to create a new real-time Air Quality Index (AQI) score every half hour. An AQI of 50 or less on the Environmental Protection Agency’s AQI scale means the air is healthy for everyone, while scores higher than 50 represent increasingly hazardous conditions — scores of 101 to are considered definitively unhealthy, 200 to 300 very unhealthy, and scores of 301 up to 500 are considered hazardous. The air quality station was installed in 2022; the corresponding website that shares real-time data went live in September 2024.
”We want the members of our local Pasadena community to have access to trustworthy air quality data,” says Nathan Dalleska, director of the Caltech Resnick Water and Environment Lab and a member of the team. “We want people to be able to make informed decisions about their own and their kids’ outdoor activities when the air quality reaches an unhealthy level.”
The idea of building and installing an air quality monitoring station on campus had its origins long before the January wildfires. Because of LA’s long-standing reputation for dangerous levels of air pollution, Caltech scientists beginning with Arie Haagen-Smit dedicated themselves to understanding various environmental impacts. Haagen-Smit helped define the nature of smog in the late 1940s; because of this work, as well as efforts by the Los Angeles County Air Pollution Control District and later the EPA and California Air Resources Board, the region’s air quality actually improved precipitously from the 1970s through the early 2000s.
And then, something changed. “Around 2010, according to regulatory agencies, we plateaued,” Wennberg says. “Air quality improvements in the Los Angeles basin stalled.”
That puzzling data is why the team decided to launch their own air quality study, and when they realized they would need air quality monitoring equipment on campus. They also realized that such a station “would also provide a resource for our students here on campus, so they can know if it’s a good idea to go for a run this afternoon, or if they should work out in the gym instead,” Crounse notes.
Not long after the launch of the monitoring station, however, the team stumbled on a second puzzle: They noticed discrepancies between their measurements and the air quality data provided to most people via their smartphones. For example, one smartphone app displayed a relatively low number at the same time the Caltech monitoring station recorded an AQI that warned against going outside.
To figure out this two-pronged problem—the air-quality plateau and the data discrepancy—the research team zeroed in ozone. Ozone (O3) consists of three oxygen atoms bonded together, as opposed to the two oxygen atoms that make up the O2 we breathe. Ozone is highly reactive, and can irritate biological tissues and exacerbate conditions like asthma when breathed in. “High up in the atmosphere we need ozone as it absorbs harmful UV photons from the sun,” says John Crounse. “But ozone at ground level is something that’s toxic to us, our pets, and plants.” Indoors, ozone reacts on the walls, floors, and pretty much any surface so concentrations tend to be quite low.
Outside, there is also a daily cycle in the levels of ozone near the ground. “Due to the atmospheric chemistry driven by sunlight, ozone generally peaks in the early afternoon here, and it's often highly elevated for several hours around the afternoon peak”, Dalleska says.
Soon after the Caltech air quality website launched, the research team began to hear from colleagues asking why their phone apps said air quality was poor in the evening while the report from the Caltech air quality site showed that the worst of the day’s air pollution had receded. Crounse was able to tie the reported air quality to (mis)reporting of ozone.
After digging into the problem, Crounse realized that the issue has its origin in one of the methods the U.S. Environmental Protection Agency (EPA) uses to assign ozone pollution. “The EPA’s primary responsibility is to manage the public health of Americans through regulatory action” notes Wennberg. “Only recently has the agency taken on the added responsibility for providing more real-time data to the public for use in personal decision-making — for example, when to go outdoors for exercise.”
It turns out that the EPA’s ozone reporting assumes that its long-term toxicity is related to exposure averaged over an 8-hour period. As a result, the air quality number reported using EPA method is an average of the preceding 8 hours. This explains the odd (and clearly not correct) communication to the public that ozone air quality is fine at a time when ozone is peaking in the early afternoon (because ozone levels were, on average, low for the preceding 8 hours) but awful in the late afternoon and early evening, long after the peak of pollution has passed.
Dalleska notes: “You might look at your phone weather app and say, ‘Hey, it's one o'clock. It's beautiful in Pasadena and it says the air quality index is good. I'm going to go for a run.’ Except that, in fact, at that moment in time, the ozone levels may be very unhealthy.”
Crounse says that he reached out to a couple of companies that provide AQI data on smartphones. None responded. “At some point, we hope. . ., whoever else is doing this will figure out how to better report real-time air quality, ” Crounse says. “For now, people should look carefully at their air quality data.”
It also turns out that ozone matters for the bigger-picture problem of why air quality improvements have stagnated. “Reducing ozone is one of the most challenging aspects right now for improving air quality. Ozone has, if anything, actually gone up since 2010,” Dalleska says, adding that although air quality in Los Angeles has gotten much better overall, “it’s still terrible.”
Understanding what is going on in the Los Angeles atmosphere to prevent further gains in air quality requires an in-depth investigation of air chemistry.
That work has already begun. From 2020 to 2022, Wennberg and colleagues including Caltech’s Richard Flagan the Irma and Ross McCollum-William H. Corcoran Professor of Chemical Engineering and Environmental Science and Engineering, and John Seinfeld, Louis E. Nohl Professor of Chemical Engineering, undertook comprehensive measurements of air chemistry on the Caltech campus, replicating observations that were made more than a decade earlier during an NSF and NOAA funded field campaign called CalNex. These measurements made in the past few years illustrate that the relationship between sources of air pollution and atmospheric levels of those pollutants is complicated. As emissions from cars continued to decrease over the past decade, the atmospheric chemical reactions involving emissions that create ozone sped up.
Just why those reactions sped up is an open question for researchers, but new data suggests the stability of ozone levels is potentially only temporary. For example: In the the spring of 2020, when car traffic was greatly reduced because of COVID restrictions, the researchers found ozone levels began to fall again, suggesting that further reducing car and truck emissions past a certain threshold will lead to ozone levels also finally decreasing again in the coming decade.
As for aerosol pollution, the Caltech team is on the case here, too, measuring atmospheric aerosols around the Los Angeles region as part the Atmospheric Science and Chemistry mEasurement NeTwork (ASCENT), a National Science Foundation-funded data-collection project.
“The network has sites in both rural and urban areas across the country, and we think this is an excellent way to try unpack what’s going on and learn how we're going to further improve air quality,” Flagan says. “By looking at lots of locations, we may be able to get at the dynamics leading to poor air quality.”