Breathing Easy — Scientists Praise a Grassroots Movement as Air Quality Sensors Arrive At Our Doorstep
Spurred by wildfires, climate change, and the Internet of Things, a citizen science movement takes root as thousands of low-cost air quality sensors are installed in homes and backyards around the world —and scientists are starting to take it seriously
Adrian Dybwad vividly recalls the day in 2015 when his chance experiment gave birth to his start-up — PurpleAir, a company that would quickly go on to launch a community science movement for air quality monitoring around the world.
The computer scientist and hardware engineer had just purchased his first air quality sensors. Motivated by his own curiosity about the state of air quality near his home in Draper, Utah, Dybwad sat down at his workstation, placed the two identical sensors he’d just purchased next to each other, and locked them into an air-tight calibration chamber. That’s when he noticed something remarkable: they both agreed with the exact same reading.
This surprised Dybwad, because the devices he purchased were hardly professional-grade. In fact, they were two low-cost Plantower sensors, normally designed for use in home air purifiers, containing tiny laser instruments the size of matchboxes. The ones Dybwad bought were available off-the-shelf for about $25.00 each.
It was the readings themselves that took Dybwad most by surprise. From his simple, low-cost sensors, Dybwad was seeing measurements of airborne particulate matter comparable to professional-grade instruments — those used by government organizations like the E.P.A., and relied upon by scientists and air quality experts around the world; hardware vastly more accurate, stable and reliable that can cost more than $150,000.
Then, Dybwad did something even more remarkable: he wired his sensors into the Internet via Wi-Fi. This began a quest to freely share his air quality data for anyone in the world with an internet connection to view.
From these humble beginnings, and just five years later, PurpleAir, which Dybwad has led as CEO, has grown to include nearly 10,000 sensors installed in over a dozen countries around the world. During a two-day period at the height of the wildfires in Northern California, more than one million visitors navigated to the PurpleAir website to view up-to-the-minute air quality information.
Scientists Take Notice
It didn’t take long before PurpleAir caught the attention of scientists and experts in the field of air quality.
“I would put PurpleAir in a class of companies that are doing good — even great,” said Anthony Wexler, professor of mechanical and environmental engineering and director of the Air Quality Research Center at the University of California at Davis.
In the eyes of Wexler and many air quality experts, PurpleAir has offered serious advantages since its inception. The sensors Dybwad had chosen were low cost but accurate, enough so that users could get good, reliable data from them. And with a low price and ease-of-use, these sensors were designed to be affordable and more accessible for a mass audience. A PurpleAir sensor costs consumers about $200.
Driven by growing community adoption, the popularity of PurpleAir means that more sensors can be installed in vastly more locations, giving both scientists and the public a truer picture of air quality in more places. Experts refer to this as spatial distribution and stress its importance as a key factor to ensure accuracy in air quality monitoring.
What spatial distribution comes down to is this: because air pollution varies so widely from place to place, only by having more instruments in more places over a wider range of geographical terrain can scientists obtain a more accurate picture of air quality.
Having too few instruments spread out over too wide an area has long plagued experts in the field of air quality monitoring. By spreading more sensors to more places, PurpleAir helps overcome this fundamental problem posed by the spatial distribution of air pollution.
“Wherever you place that instrument, you’re only going to measure the air quality in that specific location at that specific time. One instrument cannot measure air quality in a whole city,” explains Wexler.
Motivating Citizens Through Community Science
The kind of movement that experts believe PurpleAir has spawned stems from the sensors’ low cost, ease of use, and flexible installation (sensors can work both for indoor or outdoor measurement).
Sensors are finding their way into schools and STEM classes, where students are able to see first-hand the state of air quality wherever they live. According to the American Lung Association’s annual “State of the Air” report, nearly half of all Americans live where air pollution reaches unhealthful levels — with 34.2 million children under the age of 18 being among those affected, many suffering from asthma.
Armed with more accurate data, these citizen scientists are helping to motivate interest across the country, inspiring not only learning but political action. Many are motivated to do something about the growing problem of air pollution where they live, work or play.
Measuring Particles Twenty Times Smaller Than a Human Hair
PurpleAir is in a class of air quality monitor that measures airborne particulate matter (PM). Larger particles (referred to as PM10) include dust, pollen and mold. Smaller particles (PM2.5) can result from wildfires, refining fossil fuels, and driving our cars.
Air quality scientists often stress two important facts about particulate matter and air pollution: 1), particulate matter is usually so small that we cannot see it or smell it; and, 2), we are inhaling particulate matter from the air all the time.
Particulate matter becomes unhealthful when it interacts and bonds with chemicals in the air, including sulfur dioxide and nitrogen oxides. Exposure to sunlight can cause particulate matter to form Ozone, a harmful and all too common air pollutant.
The smallest particulate matter (PM2.5) is considered more dangerous to humans because these lighter particles can remain suspended in the air for a long time — longer compared to larger and heavier particulate matter. And with more time in the air, PM2.5 has a greater likelihood of interacting with sunlight and other chemicals, exposing humans to more harmful air with each breath we take.
Normally, particulate matter remains low but can flare up due to a wide range of natural elements, such as wind, barometric pressure or high temperatures, or man-made events like refinery flares, diesel burning cars and trucks, mining operations, or in times of forest fires. Even fireworks can trigger elevated levels of particulate matter.
What scientists like most about PurpleAir is its ability to gain a more representative picture of air quality in areas where official sensors are lacking, especially in areas where air quality may be worse than normal, such as near refineries, heavily used roadways, shipping terminals and mines.
Sensors Prompt Praise from Scientific Scrutiny
In a recent study, scientists at the Lawrence Berkeley National Laboratory (LBL) used the wildfires that raged across western states to put air quality monitoring to the test.
They wanted to see how well low-cost sensors like PurpleAir perform compared to regulation-grade instruments during incidents of especially high amounts of airborne particulate matter. Their findings — encouraging for PurpleAir — reveal strongly consistent correlations in readings when changes in air quality occur over time.
“We compared the low-cost monitors to one that is used by regulatory agencies in air monitoring stations. It turns out their correlations are phenomenally good. When one goes up, the other goes up at the same time, and it is proportional. That gives us a lot of hope for being able to use them for real information,” said Woody Delp, a senior mechanical engineer who studies air quality instrumentation and one of the lead authors of the LBL study.
Feeding The Web With Citizen Science and the Internet of Things
PurpleAir has achieved for air quality what Waze, the GPS app, has done for navigating through traffic jams.
With an open design, PurpleAir’s biggest achievement may lie in its ability to share reliable, up-to-the-minute data on air quality and push that data over the internet for display on websites for all the world to see.
Through a simple Wi-Fi interface, measurements from any sensor can be streamed live over the internet for display on any number of websites, provided they are set up to receive the data in its standardized format.
On its website, PurpleAir sensors display air quality measurements using the Air Quality Index (AQI), a standard measure used by official government agencies like the E.P.A. The data is then stored and archived as part of PurpleAir’s internet service offering.
Other websites, such as AirNow, an official source that is working in partnership with many air resources government agencies, including the E.P.A, NASA, NOAA, the National Park Service, and the U.S. Forest Service, has begun overlaying data from PurpleAir sensors alongside official data from government instruments. This further reflects PurpleAir’s growing acceptance within scientific circles.
Changing How We Think about Field Science Monitoring
For decades, air quality experts believed that having a single, expensive, high-quality, regulation-grade instrument located in one city or one area was enough to provide an accurate picture of air quality.
Today, the PurpleAir sensor is turning this traditional assumption upside down, by radically reshaping what’s possible. And with the success of PurpleAir for air quality, experts are re-thinking how low-cost instruments might soon find their way into other avenues of science where field monitoring is critically important.
In the future, environmental monitoring will look different, thanks to PurpleAir. Already, sensors are being developed to measure Ozone and other harmful pollution — using the same low-cost sensor design that has made PurpleAir a movement.
Other areas of sensors and monitoring may include noise levels and water quality where sharable data could help inform scientists about noise and water pollution. Low-cost sensors may soon find their way to lakes, rivers and streams near oil refineries or pipelines, locations typically not served by many official instruments.
Professor Wexler summed up the phenomenon: “Sensors like PurpleAir have opened up so many opportunities here: for education, for people to have greater appreciation for the quality of their air — which may not be clean even though it looks clean. And for scientists and engineers like me, those who really want to understand the spatial distribution of air pollution in a city or a region, wow, what a resource these things are.”
Interview with Professor Anthony Wexler, director of the Air Quality Research Center, University of California Davis, November 19, 2020.
Bio: Dr. Anthony Wexler is Distinguished Professor of Mechanical, Civil and Environmental Engineering at U.C. Davis and is currently director of the Air Quality Research Center there. His lifetime work has focused on measuring and understanding the human health affects of air pollution. He received his PhD from California Institute of Technology, M.S. in Mechanical Engineering from MIT, and a B.S. in Engineering Physics from U.C. Berkeley. His work in air quality has helped inform experts at large government agencies like the E.P.A. as will as city governments in Pittsburgh, Houston, and Fresno, California. He is frequently cited for his studies of air quality and the consequences for human health posed by rising temperatures and global climate change.
Question: As a scientist and expert in the field of air quality, how do you see PurpleAir as a good thing? Should we even call what PurpleAir is doing with low-cost air quality monitoring as “a citizen science movement?”
Well, there are many companies getting involved — I’d say not all of them are good. But I would put PurpleAir in a class of companies that are doing good — even great, from many different points of view.
One point that’s good is how PurpleAir is helping to make people aware that the air they breathe may not be healthful, even though they cannot see it or smell it. So that helps from a political point of view: politicians just can’t ignore what people are seeing from the data — it’s like, wait a second, my air sucks.
Secondly, these sensors are great from a STEM point of view. It’s making it possible that schools and classrooms can have these sensors; they can install them indoors or outdoors. So students can get to see the actual data, and say, hey, look, this is very interesting, ten kids in our class have asthma. And they might see when these kids with asthma start using their inhalers more often when particulate matter goes up and air quality goes down.
So, air quality becomes something that’s no longer abstract for these STEM students. Kids in class can say, gee, air quality is affecting my life, my friend’s life — it becomes real for them. And this is incredibly important here.
From an air pollution measurement point of view, PurpleAir is great, but in some very different ways. In the conventional world of air quality, typical instruments for particulate monitoring, including PM10 and PM2.5, are pretty expensive. So government organizations like the E.P.A. can only afford to put a few of them around. So, especially within bigger regions, a city might have only one instrument. But air pollution concentrations vary a lot, especially in larger cities where you’ve got big differences in air quality from one location to another.
This touches on the environmental justice issues that come up a lot with air quality. If you happen to live near a factory, or a busy roadway, you may be experiencing very high levels of air pollution. Whereas, people living just a half mile or even a quarter of a mile away, they may be breathing very different air, with significantly less air pollution.
So this ability to get more data on air quality becomes great motivation for regulation. People start saying, hey look, I thought I was living in a clean neighborhood, and here it is that my air quality is bad. Yet my friend who lives a quarter mile away, their air quality is pretty good. And that’s when people start calling their elected officials asking that they start addressing air quality as an issue.
So, yes, I think we can accurately call this a movement.
What are the challenges you see with PurpleAir — are there data accuracy issues? And how might we improve the reliability of these sensors?
Sure, there are data accuracy issues, but consider that you’re spending $200 for a PurpleAir sensor vs. $20,000 for a regulatory-grade instrument. And even if you have the money to invest in that $20,000 instrument, you’re going to run into the problem of spatial distribution: wherever you place that instrument, you’re only going to measure the air quality in that specific location at that specific time. One instrument cannot measure air quality in a whole city.
Air sensors like PurpleAir can work in concert with the regulatory-grade instruments — to help give us a more accurate picture of air quality. You’ve got the regulatory-grade instrument that’s giving us a “gold standard” reading. And then you’ve got the PurpleAir sensors, and they are not a “gold standard,” but they give you this spatial distribution which you couldn’t get otherwise.
Now, in terms of the data coming from these low-cost sensors, there are techniques to improve their accuracy. First off, in the PurpleAir sensor, you’ve got two redundant sensors inside. So this design provides a kind of built-in accuracy check on the data: if one sensor doesn’t agree with the other, you can say the data isn’t very reliable. But if they agree with each other, chances are pretty good that the result is decent.
There are further ways to improve accuracy of sensors like PurpleAir — and I’ve been talking to others in the field about this: since we already have these regulatory-grade monitors, people could just bring their PurpleAir sensor to within, say, 10 feet of that monitor, where we know the air quality is the same. They could then re-calibrate their PurpleAir sensor to the reading of the “official” monitor. So, there could be a service, where local air districts or states or the federal government could help facilitate this process where people could re-calibrate their instruments, which might involve leaving the sensor in close proximity to the instrument for a few hours or a day, to make sure the reading is accurate. Then, once you take that sensor home again, you’d have a better idea that you’ve improved the accuracy because you have this calibration. It becomes a quality assurance check.
A big part of what PurpleAir is doing is getting many more people interested in air quality. Just one example of what we’re seeing as a consequence: we’re seeing much greater interest in our annual air sensors conferences. Last year, we ran a conference session all about low-cost airs sensors and crowdsourcing. And we saw 650 people from around the world at that conference, which was a significant increase.
Spatial distribution of air pollution monitoring will always be key. Where the regulatory sensor happens to be may measure quite different air pollution from your home or your office or wherever you might be.
Sensors like PurpleAir have opened up so many opportunities here: for education, for people to have greater appreciation for the quality of their air, which may not be clean even though it looks clean. And for scientists and engineers like me, those who really want to understand the spatial distribution of air pollution in a city or a region, wow, what a resource these things are.
What about the future? Given how PurpleAir has become this phenomenon that’s likely here to stay — and now that the sensors are performing fairly well — what would you suggest as a scientist to improve what PurpleAir is doing? What about climate change? — shouldn’t we be measuring more greenhouse gasses? How might PurpleAir be made better going forward, in your opinion?
One important fact about greenhouse gases is that they don’t affect your health directly — not in the way that particulate matter and toxins affect your health when they are factors for air quality. But greenhouse gases do affect health indirectly, through climate change.
Damaging greenhouse gases — like CO2 and nitrous oxide — last for very long periods of time in our atmosphere — hundreds of years. So, for greenhouse gases like these, their spatial distribution is not so varied. Therefore, having local measurements is not so important for these kinds of gases.
The things that are most important are the things that are affecting peoples’ health. For example, Ozone. Ozone is something that has a great deal of variation in term of spatial distribution, so it’s important to measure Ozone.
Some of the more popular sensors are terrible at measuring Ozone. But a few others are proving to be better than most. We could improve upon this trend.
Most of what PurpleAir is doing is focused on measuring the particles. What we need most is measuring the gases — particularly the harmful gases that affect health. And, unfortunately, there are a lot of gases out there that damage health. Ozone is a common one, but there are tons of others: carcinogenic compounds, heavy metals, carbon monoxide.
PurpleAir is focused on physical measurement — you measure at the particle level, the physical size. Where we need to be heading is toward the chemical measures — carbon monoxide, ozone, and especially PAH’s (polycyclic aromatic hydrocarbons, a class of chemicals that occur in coal, crude oil, and gasoline known to cause cancer) — including benzene, pyrene, and xylene.
These are carcinogenic compounds that are released in vast quantities into the atmosphere, but people in neighborhoods don’t know it because they can’t see them or smell them —these compounds don’t give off a smell, but they are causing cancer.
So, that’s where I hope we’re heading — measuring the chemicals and not just the particulate matter.
Right now, if you want to measure toxic metals in the atmosphere in real-time, the kinds of instruments that measure these carcinogenic compounds cost $150,000, and there’s only one company that makes them.
Here at our research center, we are developing an instrument that could cost about $8,000 to $10,000. Obviously, this is not like a low-cost sensor, but it’s not like the $150,000 instrument either. We hope to see chemical sensors that can detect multiple chemicals that are priced in this range. They won’t be in the low-cost range, but we think many more people will be able to afford them.
For the instruments of the future, the ones that measure these toxic metals, I’m talking to companies that can commercialize it.
Main Sources for my writing — covering PurpleAir and the science of air quality monitoring:
Video: PurpleAir Discussion — Interview with Adrian Dybwed, Founder, Owner and CEO of PurpleAir (Links to an external site.) — UC Davis Air Quality Research Center, May 15, 2020
Scientific Article: Wildfire Smoke Adjustment Factors for Low-Cost and Professional PM2.5 Monitors with Optical Sensors, by William W. Delp and Brett C. Singer, Indoor Environment Group and Residential Building Systems Group, Lawrence Berkeley National Laboratory, Berkeley, California. Published in Sensors — Special Issue 2020: Sensors for Air Quality Monitoring.
Link to EurekAlert article: “Low-cost home air quality monitors prove useful for wildfire smoke — Berkeley Lab air quality scientists assess performance of four consumer devices,” New Release, 18 August 2020
Link to full report: https://www.mdpi.com/1424-8220/20/13/3683/htm
“Using AirNow During Wildfires,” AirNow.gov, October 2020. Link: https://www.airnow.gov/fires/using-airnow-during-wildfires/
Video: “Particulate matter air pollution,” Harvard Professor David Keith explains how particulate matter (PM) impacts air quality and highlights the threats PM has on human health. Cites major studies that link particulate air pollution to declines in human health around the world. https://www.youtube.com/watch?v=wl3BkTE4TEc
“Making Sense of PurpleAir vs. AirNow, and a New Map to Rule Them All,” by Kevin Stark, KQED Science, September 4, 2020 — https://www.kqed.org/science/1969271/making-sense-of-purple-air-vs-airnow-and-a-new-map-to-rule-them-all (Links to an external site.)
“Californians Turn to Low-Cost Sensors for Highly-Local Air Quality Data,” by Kevin Stark, Peter Arcuni, and David Brooks, KQED Science, November 5, 2019 — https://www.kqed.org/science/1950648/californians-turn-to-low-cost-sensors-for-highly-local-air-quality-info (Links to an external site.)
“Special Report: U.S. air monitors routinely miss pollution — even refinery explosions,” by Tim McLaughlin, Laila Kearney, and Laura Sanicola, Reuters, December 1, 2020. Highlights the need for increased spatial distribution. Insights into how government monitoring programs are too often underfunded. Also stresses the need for greater monitoring of toxic chemicals in the air, not just particulate matter. https://www.reuters.com/article/usa-pollution-airmonitors-specialreport/u-s-air-monitors-routinely-miss-pollution-even-refinery-explosions
American Lung Association “State of the Air” Report 2020 — https://www.stateoftheair.org/key-findings/
Bio’s of Professor Anthony Wexler — From the Center for Health Journalism, https://centerforhealthjournalism.org/resources/sources/anthony-wexler , and, the University of California at Davis, https://mae.ucdavis.edu/directory/anthony-wexler
Sources on PurpleAir sensor installation, set-up and other resources for the general public
Video: “PurpleAir PA-II-SD Unboxing and Set-up,” July 14, 2019 (Links to an external site.) — Demonstrates in-door mounting and set-up of PurpleAir device
EPA Air Sensor Toolbox, https://www.epa.gov/air-sensor-toolbox
PurpleAir User Group on Facebook — https://www.facebook.com/groups/purpleair