Why Air Quality Index Readings Vary So Much During Wildfire Smoke Events

Millions of people on the east coast are getting familiar with the air quality index, or AQI, due to the abhorrent air quality floating down from the Canadian wildfires. One thing they may be noticing is a huge range in AQI readings even within a relatively small distance. For example, looking at the popular website PurpleAir, I can see readings from a low of 43—which is considered normal—to a high of 349—hazardous—right now in Brooklyn. Meanwhile, the government-run AirNow website is giving readings of 159 to 188 in the New York City area. What gives?

“The explanation for this is simple,” said Faye McNeill, professor in earth and environmental sciences at Columbia University. AirNow data comes from a small number of monitors that are “carefully selected to avoid being influenced by any one local pollution source,” McNeill said. PurpleAir data comes from low-cost sensors sold to the general public. They’re less reliable than the AirNow monitors, McNeill said, and can be installed anywhere, including indoors, or next to a pollution source like a highway or high-emitting building. As a result, PurpleAir data will naturally be more variable.

When we talk about air quality, we’re usually referring to PM2.5 readings, or particles smaller than 2.5 micron, said Suresh Dhaniyala, a professor at Clarkson University and co-director of the Center for Air Resources Engineering and Science. There are federal standards for how to measure those particles and scale them to get an AQI reading. But the PurpleAir sensors use a low-cost light scattering sensor that does not use the federal standards. As a result, Dhaniyala says, they typically run “around 60 percent higher than EPA values” for urban air pollution, although it may be different during smoke fire events. Plus, the PurpleAir numbers are updated much more frequently, usually every minute to ten minutes, whereas the EPA numbers through AirNow are updated every hour. McNeill says her research has shown the discrepancy between Purple Air and AirNow monitors is higher when pollution levels are highest.

As a result, there are trade-offs between the more accurate and reliable sensors from AirNow with the more ubiquitous sensors from PurpleAir and other apps that use PurpleAir’s data such as IQAir. AirNow’s numbers are the gold standard in terms of scientific accuracy but may not capture rapid changes and don’t provide real-time readings. They are also more spread out. For example, there are only two in all of Brooklyn, so someone living far from those sensors may not be getting a reading that accurately reflects their conditions.

So there’s no universal answer as to which site or data source provides the “best” information. McNeill says she always uses AirNow for U.S. air quality information, although she also uses Purple Air sensors for research purposes after taking care to calibrate the data before sharing it. But Dhaniyala says if you don’t have an AirNow site near you, Purple Air “might be the most relevant,” especially when looking at changes over time rather than the absolute AQI number, which you can see by clicking on the number on the map. But regardless of the specific number, they’re all telling a similar story right now: Shit’s bad out there.