Side by Side: Improving Air Quality Science

Jun 18, 2019

by Calvin Cupini


I grew up in St. Louis, a city in the middle of the country right on the eastern edge of Tornado Alley. My summers were always filled with heavy dangerous storms, power outages, and regular interruptions on television for important weather updates. Where I’m from, knowing the difference between a tornado watch and a tornado warning can mean life or death. After all, it’s not good enough to only know about a region-wide band of storms: people need to know “Will it touch down, and if it does, where?.”


For severe weather, the “where” can matter as much as the “what.” It’s not alone in that regard.

Everyday, people are exposed to air pollution that affects their health. Local levels of air pollution can differ based on proximity to its source. Studies have even shown that air pollution can vary by as much as eight times within just one city block.

Thanks to the Clean Air Act, air quality has to be maintained at  standards established to protect public health. Measuring whether or not these have been met is a giant challenge. Expensive monitoring equipment has been installed all over the country to provide data about air pollution concentrations and relate them to health impacts via the Air Quality Index.

There’s a challenge, though. While the monitors are fantastic, they are expensive and resource intensive to maintain. They are also few and far between in most of the country.  There are currently just 23 monitors measuring Particulate Matter (PM2.5) throughout all 100 of North Carolina’s counties. It’s simply not possible to measure air pollution with these stations at a neighborhood-by-neighborhood scale. Even worse, monitors are being shut down more often than installed due to budget constraints.

Communities near busy roads and highways often breathe in much more air pollution than people living just a few blocks away. There are currently too few official monitoring stations to capture these disparities.

Overall air pollution has been improving, but it is still responsible for 70-100 thousand deaths each year in the US. That’s more than breast cancer, and even more than traffic accidents and gun shootings combined. Location plays a major factor in understanding the cause and impacts of weather, crime, and traffic accidents.  It’s a huge part of allocating resources to address and improve these issues as well.

We need to know now, more than ever, exactly where air pollution is and who it most affects.  


Clean Air Carolina is providing small, low-cost air sensors to Citizen Scientists around the state, enabling us to capture hyperlocal disparities in air quality.

That’s where Citizen Science comes in, scientific work undertaken by the general public in collaboration with professionals. Advances in technology have resulted in a diverse market for small, portable, low-cost air pollution monitoring devices that can measure the same pollutants as the instruments used by official regulatory agencies, including PM2.5. Citizen Scientists can deploy sensors at their homes, places of work, churches, etc. in a network that focuses on spatial scales ranging from county to street level data collection. This means collecting air pollution information at hyperlocal levels that have never been captured before. Impacts from construction, local industry, prescribed wildland burning, neighborhood rush hour traffic, and other local sources can be seen in high resolution, real-time, and made freely available to the public.  

This brings the power of data to the people who need it most.


It isn’t that easy, though. Lower-cost sensors present a quality challenge for citizen scientists and researchers. The devices measure the same pollutants, but in very different ways. For instance, the official PM2.5 monitors in Mecklenburg County have special dehumidifiers to remove water from the air before it’s sampled; after all, water is a particle in the atmosphere but not one we want to measure.  Low-cost sensors don’t have this ability, often resulting in PM2.5 measurements that are high compared to the scientific reference. This would mean all of the data collected by citizen scientists is skewed, inaccurate, and possibly unusable.

That wasn’t going to stop us.  

The data coming out of low-cost devices was impressively consistent with data from other low-cost devices. The issue stemmed from the low-cost sensors measuring the temperature and humidity in the air that the official monitors control for. We were able to find out how much this effect was by setting one up in the exact same location as an official, regulatory sensor and comparing the results. Then, once we knew the effect of the other variables, we were then able to go back and adjust the readings to make them more accurate.  

In fact, we made them a lot more accurate.

After a multi-year partnership with Mecklenburg County Air Quality, UNC Charlotte’s Dr. Brian Magi and Clean Air Carolina located a low-cost sensor at the Remount Road air quality monitor for sampling side-by-side in March of 2017.  We collected over a year’s worth of data (several million individual measurements), and began to write code that would make sense of the data. Once quality controls were completed, the data was compared, corrected, and compared again. The results showed up to 57% improvement in the accuracy of the low-cost sensors. More importantly, the correction can be applied to thousands of sensors in use today, improving the data for citizen scientists and produced by low-costs tools everywhere.     

A low-cost sensor was located at the Remount Road monitoring station, providing side-by-side data comparisons.


Now. Clean Air Carolina and its Citizen Science AirKeepers program are establishing a statewide monitoring network in North Carolina designed to serve both the gathering of scientifically-relevant particle pollution (PM2.5) data and connecting communities with information needed to protect their health and advocate for change. Clean Air Carolina currently has low-cost sensors installed in 73 North Carolina counties and aims to have a sensor installed in all 100 counties by the end of 2019.  

A big part of that mission is the use of reliable, accurate, and affordable sensors affixed to homes, businesses, and places of interest in communities all across the state. That is only possible with the help of our AirKeepers, government partners, researchers, and supporters like you.  

Today, we make sensors just a bit better, tomorrow we make communities healthier. The forecast for the future of air quality has never looked clearer.   

Dr. Brian Magi of UNC Charlotte sets up a low-cost air sensor.


For more information about this project, read the recent Aerosol Science and Technology publication by Dr. Brian Magi, Calvin Cupini, and others.