Author Topic: Purpleair Range  (Read 1216 times)

0 Members and 1 Guest are viewing this topic.

Offline azchrisf

  • Cobra Weather Dominator Operator
  • Senior Contributor
  • ****
  • Posts: 239
    • Copperwood WX
Purpleair Range
« on: February 01, 2018, 09:36:15 AM »
Does anyone know or have a guess as to how big of a circumfrence this sensors measurement would be valid for?
I'm assuming the results are good for at least 1/2 mile.

IIRC the AirNow stations does miles in each direction.

Offline Toxic

  • Senior Contributor
  • ****
  • Posts: 275
  • #conligwx
    • conligwx.org
Re: Purpleair Range
« Reply #1 on: February 01, 2018, 12:27:54 PM »
not sure you can really range the sensor, but some more pdf's here:

http://www.aqmd.gov/aq-spec/product/purpleair-pa-ii


Offline Eno

  • Senior Member
  • **
  • Posts: 75
Re: Purpleair Range
« Reply #2 on: February 11, 2018, 11:14:55 AM »
Depends on what kind of site the sensor is located in.  High density urban?  Heavy industrial?  Rural, open country?

Probably either what EPA terms "Middle Scale", where sensors would show show significant differences between locations that are ~0.1 to 0.5 km apart, or I'm guessing more common would be "Neighborhood Scale" 500 m to 4 km.


Offline azchrisf

  • Cobra Weather Dominator Operator
  • Senior Contributor
  • ****
  • Posts: 239
    • Copperwood WX
Re: Purpleair Range
« Reply #3 on: February 11, 2018, 11:39:14 AM »
Suburban.

Offline Forever

  • Contributor
  • ***
  • Posts: 127
Re: Purpleair Range
« Reply #4 on: February 11, 2018, 12:08:10 PM »
Like with our weather stations it is only really accurate for the location it's installed.

You may have a source of pollutant that my not exist only a few hundred feed away or vice versa.

Looking at the PurpleAir map I tend to have readings close to the sensors around me but as the distance grows so can the difference in the readings.
« Last Edit: February 11, 2018, 12:19:27 PM by Forever »

Offline Eno

  • Senior Member
  • **
  • Posts: 75
Re: Purpleair Range
« Reply #5 on: February 11, 2018, 12:37:48 PM »
It's a sample point; n = 1.  The more sample points the better to estimate a true average.  I think that's the advantage of the PurpleAir network, because it is better at showing the spatial variability of PM concentrations.  The EPA monitors are supposed to represent air quality for large populations of people, but the nearest one to me is 7 miles away and I noticed, way back when my PA sensor was online  :sad:, that the EPA monitor didn't always represent the air quality around my neighborhood.

Some details about air quality monitor network design:

https://www3.epa.gov/ttn/amtic/files/ambient/pm25/network/r-99-022.pdf

Offline azchrisf

  • Cobra Weather Dominator Operator
  • Senior Contributor
  • ****
  • Posts: 239
    • Copperwood WX
Re: Purpleair Range
« Reply #6 on: February 11, 2018, 12:44:58 PM »
Well, hmm...I've marked my sensor as being semi-accurate for about 1/2 mile around me for my neighborhood.
*Technically* that may not be correct, but I think it gives a pretty good indicator of the overall pollution level in the neighborhood.

Offline Eno

  • Senior Member
  • **
  • Posts: 75
Re: Purpleair Range
« Reply #7 on: February 11, 2018, 01:09:28 PM »
Well, hmm...I've marked my sensor as being semi-accurate for about 1/2 mile around me for my neighborhood.
*Technically* that may not be correct, but I think it gives a pretty good indicator of the overall pollution level in the neighborhood.

It looks pretty homogeneous from the imagery. Seven other sensors in Phx and at the moment they're all showing 1 or 2!  Probably more variability on bad air days?  Valley Surface of the Sun. I used to live in Globe; would've been interesting to have a PA sensor there, downwind from the copper smelter.

The PA map shows 2 sensors at your house. :?:

Offline azchrisf

  • Cobra Weather Dominator Operator
  • Senior Contributor
  • ****
  • Posts: 239
    • Copperwood WX
Re: Purpleair Range
« Reply #8 on: February 11, 2018, 01:12:51 PM »
Well, hmm...I've marked my sensor as being semi-accurate for about 1/2 mile around me for my neighborhood.
*Technically* that may not be correct, but I think it gives a pretty good indicator of the overall pollution level in the neighborhood.

It looks pretty homogeneous from the imagery. Seven other sensors in Phx and at the moment they're all showing 1 or 2!  Probably more variability on bad air days?  Valley Surface of the Sun. I used to live in Globe; would've been interesting to have a PA sensor there, downwind from the copper smelter.

The PA map shows 2 sensors at your house. :?:
Oh yes it's variable on other days, a big bought of wind came through over night and cleared out the air. The other one is inside, I need to have them remove it from the map.

Sent from my Pixel using Tapatalk


Offline yan

  • Member
  • *
  • Posts: 8
    • personal website
Re: Purpleair Range
« Reply #9 on: July 09, 2018, 10:30:59 PM »
The device does not have a range.
if you barbecue in your backyard for example you can have a spike that is ultra local. PM results typically double from sidewalk to the center of a busy street. The result you get in terms of range will depend on distance to sources, diffusion rate, bulk transport and long range average. As a result, an instantaneous reading will probably reflect your immediate local environment, while a daily average will reflect the weighted average of a large range upwind. for example, let say there is a constant south wind at 1mile/h, PM diffusion rate 0.1mile/h, and the long range PM value is 5ug/m3. your 1h average PM reading area will look at the long range value plus sources in a 1milex0.2mile ellipse due south. your 24h reading will look at 24milesx4.8miles area + long range PM. The weight of the area in term of range will be proportional to the distance: local sources will weight more than far away sources. Evidently, wind direction and diffusion rate will greatly change based on a lot of factors, and you need powerful atmospheric models to account for that. You will also need a correct model to know where the sources are.

The value of the device is in its ultra local result and in the grid

 

anything