What I want to do or know
There are a lot of really neat optical monitors coming out around particulate matter sensing. Specifically I've heard about people using the Spec and the Air Beam. There's a new one (to me) that @MD2020 and @guolivar recently referenced on the air-quality google group called "Purple Air." It seems really interesting, but I am wondering about what environmental factors these sensors take into account in producing their readings. Particularly, I've heard humidity can affect the results of these optical monitors. Looking into the Purple Air page, it looks like only some of the Purple Air monitors have humidity sensing add-ons, and it's unclear if or how this data relates to the results. Looking for more information about this question if anyone has ideas or resources!
Yes is the simple answer: high humidity and/or moisture can effect particulate counts by raising them. Once you accept this the next step is how to decide if that is the case for a particular set of readings. An obvious thing to do is check the weather history around the relevant time period. Since most weather history will come from airports which are most likely not near to the monitoring device one has to work in generalities. If the weather shows potential for problems ( in addition to the data being unusually "high", distinctly different from before and after), label the relevant data as problematic and I would suggest you do not use it in your analysis and be sure and state that in your commentary. There is no one size fits all adjustment that can be made, especially for the newer, cheaper sensors whether they record humidity or not. As an aside, high moisture content will in general make it more difficult for particulate close to or on the ground to become or remain airborne. This is why it is a best practice for construction or mining outfits to hose down roadways, tires, and sand piles in exceedingly dry times. So in high humidity/moisture times a hefty rise in particulate matter without a conspicuous cause is questionable. I realize Stevie posted this as a question, but it is relevant to the plots air quality group also and I don't know how to get this reply to all the relevant groups. There was a recent post for help in the air quality group regarding the Dylos particulate monitor and a graph of some monitoring results was provided. There were many one hour average readings of over 300 µg/m3 concentrations for pm2.5. I replied with some weather data that definitely showed potential for problematic readings. But there are a few more questions pertinent to optical sensors that are important. Optical sensors count particles. They do not monitor concentrations. There needs to be some way to translate from a count to a concentration. The poster has not responded to a question about how he/she does this. The question is relevant even for those optical monitors that give an output in concentrations for somehow the manufacturer has had to come up with a method to translate. If this was done using some experimental apparatus a question exists about the range of known concentrations used for the calibration. In the above case if the range did not include concentrations over 300µg/m3 then the graphed concentrations are questionable even without high humidity or moisture. I'd welcome comments, questions, or disagreements.
Hi Jeff, thanks! This is great. I'll copy the question link into that email chain as well. Think some of the readers there would be interested, it's the same list, but it directly relates to that other thread!
Reply to this comment...
Log in to comment
Yes, environmental conditions do affect the ability of an optical sensor, not too much to "see" a particle but rather to "size" it.
All of these optical sensors work by measuring the intensity of the light scattering from the particles and that intensity is related to their size which is how they can be "assigned" to PM10 or PM2.5 or any other size bin they have. So, if you have hygroscopic particles (they absorb water), then they will grow in high humidity environments. That's not a problem as they will grow for any sensor that measures them but the issue is that all "regulatory" measurement has a preconditioning of their sample to comply with whatever standard they are supposed to follow. In New Zealand the guidance is to have a heated inlet that guarantees that you're always sampling particles in the same conditions, even if that means that you may be losing volatiles compounds in the process (water being one of the volatiles).
I have been working with these sensors for a while now. Here is a presentation I have last year on what we've learned so far: https://figshare.com/articles/From_Sharp_to_Plantower-What_we_learned_about_cheap_dust_sensors/4213815 [EDIT: I saw that the link didn't work so here is a shorter one that does work] https://doi.org/10.6084/m9.figshare.4213815
It can be summarised as "After using the Sharp sensors now we've moved to the Plantower PMS series and are quite happy".
A more involved answer could be that, even though the Sharp sensors served us well and we got a lot of really good information from them, their response to temperature (much more than to humidity .... see at the bottom the PDF of a paper we submitted to AMT but didn't follow up to publication) is annoying so when we tested the Plantower PMS3003 and found no significant response to temperature we moved all our low-cost sensors to that technology.
You can check our project's blog here: https://cona-rangiora.blogspot.co.nz
Our "calibration" exercise for PM2.5 and PM10 (showing a much better agreement with regulatory PM2.5 than PM10) here: https://rpubs.com/guolivar/odin-sd-pm25-cal https://rpubs.com/guolivar/odin-sd-pm10-cal
Wow. These are awesome resources!! Thanks for clarifying the seeing/sizing distinction.
My quick perusal of the attached pdf didn't find anything about effects of humidity. What you say about humidity effecting size makes sense but I would suggest the question for people using optical sensors who are not part of some organized scientific project is what effect within a bin, 2.5µg/m3 for example, would high humidity have? My experience is that counts increase; which I would speculate is due to smaller particles which might not be counted under low humidity become visually large enough to be counted.
Is this a question? Click here to post it to the Questions page.
You're right, the effect of humidity would be to change the "apparent" size but it's not just increasing the number of particles that now fall in that bin but also growing some particles out of that same bin and onto the larger one. How much it's "small particles becoming visible" against "large particles moving up a level" is not easy to estimate (it would be related to the available surface area that comes in contact with the water vapour).
That work in fact doesn't say anything about humidity because we didn't find anything relating the concentrations with relative humidity. That is not to say that effect doesn't exist, it just means that for the conditions we observed and the sensor we used there (Sharp gp2y1010au0f) the effect was not significant. I do have some data in higher humidity conditions but the main effect we always saw with the Sharp sensors was temperature not humidity. The Plantower units we're using at the moment don't seem to show any impact of humidity or temperature ... BUT we've been using them in woodsmoke dominated environments which have their own response to humidity and for other more hygroscopic particles (like sea salt) that may be different. My group is preparing some publications with the data we got last year so later this year we'll have something out and I'll make sure to include a discussion on humidity.
Reply to this comment...
Log in to comment