There are two main consumer options for Internet-connected, "citizen science" PM2.5 air quality monitors: PurpleAir and IQair.
The US EPA has studied some models of the PurpleAir and developed sensor adjustments for wildfire smoke during the 2020 California wildfires.
The PurpleAir is the only consumer sensor that appears on the US government AirNow FIre and Smoke Map (but only the outdoor sensors, which makes sense).
It is also important for Canadians to be aware that the US AirNow map only shows PurpleAir sensors within the US and near the US border, not all available sensors.
PurpleAir does have its own map though, with a lot of detail about individual sensors when you click them on the map.
https://map.purpleair.com/
The most important thing to know is sensors with a black ring around them are indoor sensors.
This is only indicated on the completely separate map startup guide.
PurpleAir should:
- Indicate this as a legend on the map itself
- Indicate this in the popup information box about the sensor
- Indicate this in the popup chart from the sensor(s)
If there are enough sensors around, you can compare and see if there are outlying values that may indicate a malfunctioning sensor.
The popup box may also give you additional information to determine if the device's sensors are malfunctioning.
AirNow Map
For example at the original time of writing in and around Ottawa all outdoor sensors are showing red unhealthy levels, except for a single device in Almonte reporting a much higher (ten times higher) hazardous level. On the AirNow map, circles are official government sensors, and squares are PurpleAir sensors.
PurpleAir Map
You can't get deep details about the sensor on the AirNow map, but you can on the PurpleAir Map.
Note that all sensors on the PurpleAir map have a circle; the meaning of a circle is different on the AirNow map (official sensor) and the PurpleAir map (all sensors). You'll also notice you see more sensors than on the AirNow map, as a reminder that's because AirNow shows only outdoor sensors, and circles with a black outline on the PurpleAir map are indoor sensors.
Also note the small legend in the upper left: by default PurpleAir is displaying US Air Quality Index (AQI) numbers (on a scale from 0 to 500) and colours.
These are not PM2.5 sensor measurement values.
You can visually see that an outlier sensor has a reading ten times that of the other outdoor sensors near it.
In the above section, I highlighted:
- How to go from the AirNow map to the PurpleAir map to get more detail about a sensor
- That the PurpleAir map shows indoor sensors with a black ring around them
- That the PurpleAir map is showing US AQI by default, not raw PM2.5 particle measurements
PurpleAir Map Data Layers
Default Data Layer: US AQI
As I indicated above, the default data display for the PurpleAir map is called US EPA PM2.5 AQI.
You might naturally think this means it is showing you raw PM2.5 sensor values, but it is not. It's the US Air Quality Index (AQI) scale from 0 to 500, based on PM2.5 measurements.
Just to re-emphasize, this is not PM2.5 raw measurements. I'll explain below how to get raw measurements.
SIDEBAR: The display in the IQair map is also US AQI. IQair doesn't have any way to change this, but you can see the underlying PM2.5 sensor reading by clicking on a sensor. END SIDEBAR
Data Layer: Canadian AQHI
The Canadian Air Quality Health Index (AQHI) has a different numeric range, from 1 to 10+, and a different color range than the US, going from blue to a kind of dark brownish-red, with a different colour for each of the values.
To change the data layer on the PurpleAir map, click on the cogwheel ⚙️ in the upper left (or anywhere on the entire cogwheel line).
Select Canadian AQHI.
In the PurpleAir URL, you will notice it changes from mAQI (US AQI) to mAQHI (Canadian AQHI).
Note that this is unadjusted PurpleAir readings, computed into Canadian mAQHI.
The scale at the bottom left of the screen will also change, and it is clickable, with explanations.
Data Layer Enhancement for Wildfire Smoke
If you want to see the readings using the US EPA sensor conversion for wildfire smoke, select Apply conversion: US EPA.
I realise it's super confusing to have two different US EPA items, and to think about applying a US conversion to Canadian data.
Think of "US EPA" conversion instead as "US scientific wildfire smoke conversion factor".
In the PurpleAir URL, you will notice it changes from cC0 (no conversion) to cC5 (US EPA conversion).
In practice what it means is values will be adjusted lower based on the results of the US analysis of the PurpleAir sensors versus official US sensors.
Data Layer Filter for Outdoor Sensors Only
You probably only want to see outdoor sensors, as indoor sensors (sensors with a black ring around them) will vary widely depending on the interior conditions and whatever air filters they may have running.
To see outdoor sensors only, uncheck the "show inside" item in the data layer popup.
Data Layer: Raw PM2.5 values
If you prefer to see the actual PM2.5 values in micrograms per cubic metre (µg/m³) then select Raw PM2.5 (µg/m³) from far down in the Data layer menu.
You will notice the URL changes to mPM25. Note that you can apply the US EPA conversion factor (explained above) even to the "raw" values, if you're correcting for wildfire smoke.
The scale at the bottom left of the screen will also change, and it is clickable, with explanations.
Below you can see all of the information available in the PM2.5 view, as well as an individual sensor popup.
At 8:45am Eastern on Friday July 21, 2023 it is showing an extremely hazardous level of PM2.5 in Yellowknife, reading 520 micrograms per cubic metre µg/m³ with US EPA wildfire smoke adjustment, and showing the hover text in the lower left indicating ">250µg/m³: Health warning of emergency conditions".
Looking at an individual sensor (defective sensor example)
Please note this example is from the original Ottawa area map at the top of this post, not from the Yellowknife map directly above.
It's an example of how to tell whether or not a sensor is providing correct values. If you see unusual (outlier) values e.g. much higher readings than surrounding sensors, that may indicate a defective sensor.
If you click on the sensor in the original Ottawa area map at the top of this post, you get a popup with lots of details. Note this is the US AQI value, not a PM2.5 sensor reading.
There's lots to parse out here, but the main item is the single line near the bottom of the popup:
💾 🌡 A B ✓60% PA-II-FLEX 7.04
This is telling you a lot about the device:
- 💾 = it has an SD card
- 🌡 = it has a temperature, humidity and pressure sensor
- A B = it has A and B sensors (i.e. it has two sensors)
- ✓60% = the sensors only agree with each other by 60%
- PA-II-FLEX = it is the PurpleAir Flex model
- 7.04 = it has firmware version 7.04
The sensors are right next to each other, they should agree very closely, as in nearly 100% closely. This is clearly a defective unit.
SIDEBAR: Note that the Flex and the Zen outdoor models have PMS-6003 Laser Particle Counters (which are user-replaceable).
The PurpleAir Classic (PA-II and PA-II-SD) outdoor models have PMS-5003 sensors. END SIDEBAR
In the popup chart you can also see both sensors clearly diverge to an extreme degree, as well as options to download the data.
In this section I have highlighted:
- How to parse some of the information in the PurpleAir map popup about a device
I highlighted the clearly defective sensor not as a criticism of the device but simply as a useful way to illustrate how to read the map's information.
If you click on any other PurpleAir sensor in Ottawa you'll see all the other units with dual sensors are showing 100% correlation between their two laser sensors.
Annex: PurpleAir History
Annex: US Government Analysis of PurpleAir Sensors for Wildfire Smoke
Where does the sensor data on the fire map come from?
At this time, the sensor data on the Fire and Smoke Map comes from PurpleAir. AirNow is using data from PurpleAir for the pilot, because EPA researchers have used a scientific approach to develop a correction equation and quality assurance steps for PurpleAir sensors. Researchers assessed these sensors, because their use has grown exponentially, resulting in an extensive network of publicly reporting sensors worldwide. EPA is in the process of collecting and analyzing data for other sensor models; however, the Agency does not have as much data for other sensor technologies since there are fewer deployed across the U.S.
Why do you need a correction equation for sensor data?
EPA scientists have found that PurpleAir sensors are biased -- they consistently overpredict fine particle concentrations in most locations and under higher humidity compared to the regulatory-grade monitors that are operated in the same location. While EPA does not use the sensor data for regulatory purposes, the correction equation our scientists have developed correct the PurpleAir bias, so you can easily compare readings from regulatory-grade monitors and PurpleAir sensors when you see them on the same map.
Additional context:
Things may have changed in the three years since then, but at the time, The Bold Italic said:
[For the PurpleAir, we] recommend using the EPA correction factor (available as of 10/3/2020). To do so, click the word “None” [for correction factor] and change it to “US EPA”.
If there are updates to this information for 2023 please do let me know.
![[Slide: Podcasts and audiobooks. Many sample Siri phrases listed on the slide, including Play the audiobook Tiny Beautiful Things on appName. Caption text reads: Maybe your app is about listening to Podcasts or Audiobooks.]](https://live.staticflickr.com/65535/53219382756_5d14c3ec78_z.jpg)
![[Slide: Many types of media content, listing Music, Audiobooks, Podcasts, Radio and Meditation. Caption text reads: Any app adding new support for SiriKit Media Intents...]](https://live.staticflickr.com/65535/53219688643_3753034826_z.jpg)
Recent Comments