As you might have already figured out, I’m a bit of a geek. Tinkering is a part of my being. And connect that to air pollution, it’s a happy day.
Nerds rejoice when we talk about Arduino, Raspberry Pi, or Thingspeak - terms probably unfamiliar to many. But these tools are just tiny computers and data platforms, cheap and easy to program with thousands of YouTube help videos, blog posts, and educational engagement manuals for even the newest of newbies.
Nerds to Air Quality Action
About 10 years ago, fellow geeks had an idea to couple these simple-to-use computers with readily available air pollution sensors that had been around for ages. These sensors have been in heating and ventilation systems, industrial emissions controls and even cars and trucks for decades. But when connected to one of these simple computers - voilà - a low cost air pollution sensor for the average person. In 2012, the Air Quality Egg was born. And it was good.
Okay, maybe the original Egg was good as an innovation, but as an air quality monitor, it left something to be desired.
Since that time, dozens of manufacturers have raced to market with clones of low cost sensors, usually available for a few hundred US dollars. Most market their sensor design in the most wonderful terms, but often overpromise and under deliver. Sure, it spits out numbers once they are plugged in, but are these numbers meaningful? It turns out that some are great. And some are junk.
Worldwide, air quality regulators have been watching. These sensors, which measure the pollutants that harm health, like ozone, particulate matter, carbon monoxide, and nitrogen dioxide, could replace much more expensive equipment. A $250 USD solar-powered particle sensor that can be zip-tied to a tree might someday replace a $25,000 reference monitor that has to be housed in an air conditioned shelter and maintained weekly by a technician. State and local government environmental agencies, who bear the costs of this monitors, are surely salivating at the idea of saving a tidy sum of money if they could replace their expensive equipment with much lower cost instrumentation. It’s a pricey business, this air quality stuff.
Piques of reluctant interest.
In 2019, the US EPA began a campaign to use a specific low cost sensor in conjunction with its existing reference monitoring network in order to observe forest fire emissions - mainly from the Western US - and when they might reach the ground where people live. This data, published on their Fire and Smoke map updates in near real time, and includes information on where models think fire emissions are travelling across the US.
The challenge for regulators is that the have to produce the highest quality data because they have a critical public health mission. This means they have to have a close understanding of how the sensor is working, evaluate its performance continuously against gold-standard calibration comparisons, and use a system to collect, track, and analyze all of the data. None of these things come in the low-cost sensor you just purchased for a few hundred dollars.
No one expects the data produced by these low cost sensors will carry any regulatory weight any time soon. Those pricier reference monitors, show on the map as tiny triangles lost in the cloud of more numerous low cost sensors, measure air pollution for regulatory purposes. In other words, those monitors are the ‘official’ ones and the low cost ones are not. But we can still get some useful information from these cheaper alternatives, even if it isn’t official. In the image above, it’s a pretty clear day across the entire United States on this day, despite a handful of fires burning in the Pacific Northwest.
A Cheaper Future?
Low cost sensors are catching up. Every week brings a new device to the market, and they are getting better and better with time. Some groups, like AQ-SPEC in California offer a test platform so that manufacturers can make improvements on their sensors by comparing their device against regulatory instruments in real world conditions. Low cost sensors are in pursuit of their much higher priced cousins.
But they are not there yet.
For now, there are likely too many flaws with low cost approaches to justify replacing more expensive regulatory networks. We have to make sure they are compared with accepted standards - a calibration - to make sure they are accurate. We have to compare them with something that gives us some confidence that they are accurate - otherwise, they are just science fiction. Usually a vendor provides a ‘factory calibration’ with the device, but this is pretty much the same thing as buying a watch that has been pre-set at the factory - they almost always need to be adjusted to the local environment where it is going to be used, and we use calibration to make that adjustment.
And we need a plan to handle the tsunami of data they produce. Not because a single sensor produces tons of data. But it is likely there will be tons of sensors producing modest amounts of data. Either way, we need a plan to manage this flow of information.
But they are quite effective in teaching the public about air pollution science. There are lots of educational lesson plans for all ages.
And while they might not be ‘official’, they most certainly can be used to provide important information about pollution in locations where reference monitors are absent. This includes large swaths of low and middle income countries where no monitors are present, or in areas in the United States and Europe where air pollution hazards persist and are not detected by traditional monitoring. Low cost sensors are a really useful tool to reduce environmental injustices and inequities across the world because they shine a light, even a imperfect one, where no light has previously existed.
Looks a bit hazy here today. I think I’ll go check my sensor. Maybe you should, too.
Like what I have to say? I’d love for you to subscribe. It’s free!
Love the clock metaphor to explain why lcs calibration is flawed!