Brand-spankin’ new sensor owner here, so please pardon if these are remedial questions. I dug around the forum but couldn’t find ready answers, so I’m hoping someone out there might be able to provide some insight:
My sensor is consistently reporting a number 4-5 higher than other local and national sources for my immediate area. I’m not sure how to validate/troubleshoot or even rationalize this discrepancy. For example, as I write this:
I took a quick look at your sensor on the PurpleAir map.
You have it set to 10 minute average EPA AQI. (The AQI is an air quality index that is calculated from the raw data.) Also your sensor had no conversion factor applied.
PurpleAirs are known to report high in relation to regulatory PM2.5 monitors, so applying a correction helps get the number closer to a regulatory monitor. EPA has a national conversion factor (or correction algorithm).
If you look at your PA sensor with data layer as ‘raw PM2.5’, averaging period ‘real time’ and US EPA conversion factor, your PM2.5 value is 18 and if you look at it as a 10-min average w/ US EPA conversion the value is 20. So the value seen is dependent on those settings.
I do not know how Weather Underground or IQAir report their values but they likely do averaging of some type and apply some type of conversion factor.
The EPA fire and smoke map: they do apply a correction to the PurpleAir data and I believe the averaging time is 10 minute averages. They also ‘translate’ the data into an AQI.
This might explain the EPA fire and smoke map better:
I hope that helps a bit explain the discrepancies.
(I am not sure about the real time clock is missing so would be interested to also hear what that means).
Thank you very much for this very helpful reply. I have a follow-on question or three if you don’t mind:
When you say “You have it set to…” and “your sensor had no conversion factor applied”, it sounds to me like you might be suggesting these are things I can configure on my particular sensor rather than for the entirety of the map view (i.e., all sensors visible to me on the PA map). Am I understanding that correctly? Or is it just a way to alter the view of the entire map?
When I apply the settings you’ve outlined (which make a ton of sense, thanks again!) it’s true that my AQI number drops considerably, but I’m still quite high relative to other sensors in my immediate area. Seems weird (?).
With those settings applied. I took at look at an area nearby with higher sensor density and reporting Redmond, WA. Now the discrepancy seems to have swung entirely the other direction, with IQAir reporting 19, wunderground 35, and AirNow at 50. As you can see from the linked screenshot, PA is reporting all low single digits for Redmond given these settings. This holds true for 10 min and 1-hour averaging periods.
Totally appreciate your patience here. I feel like I’m just being dense and missing something obvious.
For 1) yes, changing the setting for your sensor will change it for the others on the PA map - so those same ‘settings’ will apply to other PAs on the PurpleAir map. But I would check in case the settings get ‘reset’ if you leave the page. I haven’t experimented with this, I just do quick looks. The setting I described was ‘raw’ PM2.5 not AQI. It a bit confusing. AQI or air quality index is based on an algorithm applied to ‘raw’ data. (that web link I had shared had a good explanation). If you want to match the AQI on EPA’s Fire and Smoke map, then I suggest the ‘US EPA PM2.5 AQI’ for data layer, with ‘US EPA’ conversion, and ‘10-minute’ averaging period or 1-hr.
Your higher values could be represent local conditions or the sensor could be reporting high; it is hard to say. Is it near a location where car’s idle, near a grill, outside smoking location, potential close ‘sources’ for PM2.5? Some people co-locate 2 low-cost sensors to help ‘verify’ the readings of the other. If you had 2 PurpleAir sensors and their readings were very similar, it would give one higher confidence in the readings. Just using your individual sensor - it does have two laser counters: Channel A and Channel B (or in your case Westport Coast A and B). If the two channels are close, that can give confidence in performance. Note siting (location) of the sensor plays a big part on what type of air is being sampled, ambient or very local situation.
I haven’t looked into IQAir and wunderground so I can not comment on their values.
“AirNow at 50” - I am guessing that is the NowCast AQI not the PM2.5 concentration. The neat thing about the Fire and Smoke Map is that you can click on your PurpleAir sensor and see how EPA presents the data with their correction algorithm applied. Fire and Smoke Map
Your sensor was currently yellow when I clicked on it (it is identified by the sensor index # which I think yours is 165493, so it would be PA_165493 on the fire and smoke map). If you click on it, then click on ‘History’ or scroll down to 'Recent History". It will show your sensor’s values as 'NowCast AQI", “Hourly Concentration”, and “10 minute averages”. When I changed the settings on the PurpleAir map to match one those on the fire and smoke map, your sensor matched the Fire and Smoke map fairly well. I have found the Fire and Smoke map is a little delayed compared to PurpleAir map, so you might have to go back an hour or 2 to compare.
It is my understanding that for a PurpleAir to report to the EPA fire & smoke map, it needs to be public reporting, not designated private.
Here is a FAQ and an excerpt for more info:
There are several possible reasons why no sensors appear on the map for your area, including:
• There may be no publicly available sensors in your area;
• The sensors in your area may have failed to meet the quality control screening criteria
developed for the Fire and Smoke Map, or
• The sensor data is temporarily unavailable or offline.
Newly registered sensors also may not appear on the map for up to 48 hours.