Author Topic: ASOS/AWOS Accuracy  (Read 3468 times)

0 Members and 1 Guest are viewing this topic.

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
ASOS/AWOS Accuracy
« on: December 13, 2023, 11:28:45 AM »
My weather station (VP2) is located just about exactly 2 miles from the weather station at KBLI and, before this fall, the readings between the two were generally identical. Since October the airport station has consistently been reporting daily minimum temperatures up to 3 degrees colder than my station. When I check other CWOP stations in the general area for comparisons it seems as though what the airport is reporting is unlikely to be accurate. Two questions: 1) Is the weather station at KBLI an ASOS or AWOS? 2) How likely is it that the KBLI weather station is reading too cold?
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6741
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: ASOS/AWOS Accuracy
« Reply #1 on: December 13, 2023, 04:30:59 PM »
It's an ASOS. As far as temp anomalies, if you think there's a problem, call your WFO and make your case. They can fail like anything else.

https://www.faa.gov/air_traffic/weather/asos


Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #2 on: December 13, 2023, 04:58:43 PM »
Thanks for the info. In your experience, how common is it for ASOS instruments to fail?
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6741
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: ASOS/AWOS Accuracy
« Reply #3 on: December 13, 2023, 05:05:33 PM »
They are on a PMS like all FAA equipment, so I'm sure it's checked at least once a year. If it is off, this could potentially be critical with pilots now that icing conditions are becoming more prevalent.

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #4 on: December 13, 2023, 06:25:11 PM »
PMS?
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6741
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: ASOS/AWOS Accuracy
« Reply #5 on: December 13, 2023, 06:33:31 PM »
In my Navy days, it was referred to as the "preventative maintenance schedule". Routine checking of equipment based on a schedule, not a failure. Kinda like Disneyland replacing light bulbs before they burn out. Applies in the FAA as well.

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #6 on: December 13, 2023, 07:04:30 PM »
Got it. I would expect that ASOS equipment would be subject to regular inspection, calibration and/or replacement. I emailed the Seattle WFO so we'll see if and how long it takes for them to respond.
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6741
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: ASOS/AWOS Accuracy
« Reply #7 on: December 13, 2023, 07:28:20 PM »
Good. In mine and others experience, usually, the NWS appreciates being fed potentially pertinent info, as they don't "see" everything.

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #8 on: December 13, 2023, 07:52:19 PM »
I highly recommend contacting Cliff Mass. He runs a blog and has influence with the NWS as a professor in the Department of Atmospheric Sciences at UW.

https://cliffmass.blogspot.com/2018/05/time-to-fix-yakima-airport-temperature.html?m=1

https://cliffmass.blogspot.com/2014/04/the-great-sea-tac-temperature-mystery.html?m=1

https://sites.uw.edu/cmass/cliff-mass-info/
« Last Edit: December 13, 2023, 08:02:09 PM by CW7491 »

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #9 on: December 13, 2023, 09:02:01 PM »
Long time Cliff Mass blog reader here. Hes hit or miss as to whether hell respond and since this a recent phenomenon rather than a longstanding issue I think the WFO is my best bet. If they dont respond to my email in a reasonable amount of (bureaucratic) time Ill give them a call.
« Last Edit: December 13, 2023, 09:19:29 PM by Buick »
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #10 on: December 13, 2023, 09:10:37 PM »
It’s not like the temperatures are just obviously wrong but there’s definitely a noticeable shift in how my station compares. I’ve been running my setup for 8 years and I’ve never seen such consistent cooler daily minimum temps at KBLI compared to my site. The ASOS is only about 25’ higher and 500 meters further inland than mine. I just find it a little hard to believe that the two sites would regularly have 2-3F difference in daily minimum temp in the absence of some kind of sensible weather that would promote such a difference. Especially since what’s occurring now is not typical.
« Last Edit: December 13, 2023, 09:27:43 PM by Buick »
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #11 on: December 14, 2023, 03:04:51 PM »
I received this response from NWS Seattle. I was a bit surprised at the Celsius to Fahrenheit conversion error in the text of the response and also that the ASOS temp sensor is nominally less accurate than the SHT31. I guess if the KBLI sensor is calibrated then it is what it is but I still find it strange that there now appears (to me) to be a cold bias as compared with previous years of observations.
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CNYWeather

  • Forecaster
  • *****
  • Posts: 2296
    • CNYWeather
Re: ASOS/AWOS Accuracy
« Reply #12 on: December 14, 2023, 03:41:59 PM »
Thanks for the info. In your experience, how common is it for ASOS instruments to fail?

A few times in the past 10 years, KRME has had their ASOS fail in either temp or precip.
NWS Binghamton is a hike for them to diagnose and repair, so they use my station data as "official".
The temp sensor went out for a few days this past year I think. It happens I guess.
They don't replace sensors/equipment as often as I have the past 5 years  :-(
Tony




Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #13 on: December 14, 2023, 04:17:06 PM »
It's interesting that the ASOS temp sensor has a nominal accuracy of +/-1.06F and KBLI has been running 1.1F cooler on average than my station...
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #14 on: December 14, 2023, 08:23:29 PM »
While not really relevant to your particular issue since youve been comparing like systems over time, the accuracy of the ASOS sensor is only part of the problem when you consider rounding and C to F conversion errors. Ive mentioned this before in other discussions if you search ASOS rounding, so I apologize Im repeating myself. It might make your eyes glaze over too.

To start, I wouldnt even bother comparing to an AWOS station. Someone may find an exception and make a liar out of me, but as I understand it, all AWOS systems measure temperature and dewpoint in C, round to the nearest C, then convert to F and round again to the nearest F when it reports temperature and dewpoint for non-aviation purposes. For an extreme but common example, the AWOS measures 9.4C. This is rounded to 9C converted to 48.2F and reported as 48F even though 9.4C converts to 48.9F or 49F. If the temperature rises to just 9.5C AWOS rounds to 10C converts to 50.0F and reports 50F even though 9.5C converts to 49.1F or 49. So even if both your station and the AWOS are reading exactly the same raw temperature data, the AWOS can read 1F high or low on rounding and conversion to say nothing of sensor accuracy.

ASOS does exactly the same thing except for hourly and special observations. For hourly and special observations, it works as youd expect. It takes a raw C temperature, converts it to a raw F temperature and rounds it to the nearest F. So for our above examples ASOS will convert 9.4C to 48.9F and report 49F and convert 9.5C to 49.1F and round to 49F. For KBLI the hourly observations are at :53.

Long way of saying I wouldnt compare my data to an AWOS and I would only look at hourly and specials for ASOS to compare with my data for not only temperature but also dewpoint. Use caution if you look at the 5min ASOS temp and dewpoint.
« Last Edit: December 14, 2023, 08:26:02 PM by CW7491 »

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #15 on: December 14, 2023, 08:38:54 PM »
I was absolutely not aware of this! Ive also noticed recently that rh and dew point seem off at KBLI as well. For example, at the 17:10 observation KBLI reported rh of 87%. There are several nearby stations on WU (including my own) all of which report current rh of >95%. I look at KBLI data using MesoWest which shows data at 5-min intervals. It appears to show temp and dew point to the nearest 0.1F. Is this actually some kind of rounded C-F conversion? Whats strange to me is that, for years, the average temp at my site was generally indistinguishable from what KBLI reported and now suddenly theres this consistent difference: (my monthly mean temp - KBLI monthly mean temp) = 1.1F, (my monthly mean daily minimum temp - KBLI monthly mean daily minimum temp) = 1.3F or 1.4F. Prior to this fall, these values were typically within ~0.1F. When I compare my data with KBLI on a monthly basis, I use the NowData feature and I round all daily min and max temps to the nearest degree F and all averages to the nearest 0.1F as thats what the NowData does so that Im getting (what I thought) was an apples-apples comp. I would definitely like to learn more about ASOS rounding, etc and would really appreciate if you could point me in the direction of any threads that discuss this issue.
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6741
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: ASOS/AWOS Accuracy
« Reply #16 on: December 14, 2023, 08:42:03 PM »
Excellent info. That said, there's the averaging algorithm to consider.

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #17 on: December 14, 2023, 08:47:57 PM »
Im really only concerned with temp. I struggle with believing that the daily minimum temps that KBLI has reported recently are as much colder than what Ive measured when this has never previously occurred on such a consistent basis. For what its worth, I replace my temp/hum sensors annually.
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #18 on: December 14, 2023, 09:02:49 PM »
CW2274 brings up another good point in that ASOS uses a running 5min avg for temperature to further complicate the issue.

There have been many historical examples of ASOS temp problems you can google. SEA, HNL, TUS, TPA, YKM, and DCA just off the top of my head. Here is a link to the Washington Post article explaining the problem at DCA

https://www.washingtonpost.com/news/capital-weather-gang/wp/2015/08/11/weather-service-installs-new-temperature-sensor-at-reagan-national-airport/

See the attached screenshot of a portion of the article if you want to see how +/-5F is within spec for the NWS when they do their validation with a reference thermometer.

 [ You are not allowed to view attachments ]

A few other threads where Ive commented on this:
https://www.wxforum.net/index.php?topic=34658.msg395628#msg395628
https://www.wxforum.net/index.php?topic=34151.msg358075#msg358075

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #19 on: December 14, 2023, 09:12:42 PM »
Thank you! I think I knew about the running 5-min avg. In any case, it would be a shame to lose what Ive previously considered to be a valuable reference (if only because the data corresponded almost perfectly with mine! :-|).
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #20 on: December 14, 2023, 09:16:49 PM »
It seems like from what Im gathering here that its not atypical for ASOS measurements to be less than accurate (depending upon the standard) though it doesnt appear that KBLI is operating out of spec. I guess the upside is that now my station may be the most accurate in the neighborhood!
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #21 on: December 14, 2023, 09:37:10 PM »
If you want the reference document, its FAA ASOS Site Technical Manual S100. It use to be easy to download online, but seems tougher to find now.

Use the US Government Documents hotlink to download it if youre interested.
https://catalog.library.vanderbilt.edu/discovery/fulldisplay/alma991011864019703276/01VAN_INST:vanui

Chapter 5 is Temperature and Dewpoint Sensors chapter.
Specifically 5.5.2.4 explains their temp/dewpoint verification procedure.
Table 5.5.5 Step 10 gives the temp tolerances to reference instrument
Table 5.5.6 gives dewpoint tolerances.

Ive been told that they no longer use the Belfort Psychron 566 or the Psychro-Dyne for reference anymore but the Fluke 971.

https://www.amazon.com/Fluke-971-Temperature-Humidity-Meter/dp/B005T5JW2S/ref=asc_df_B005T5JW2S/?tag=hyprod-20&linkCode=df0&hvadid=312152817250&hvpos=&hvnetw=g&hvrand=7428861331032207506&hvpone=&hvptwo=&hvqmt=&hvdev=m&hvdvcmdl=&hvlocint=&hvlocphy=9027607&hvtargid=pla-458962573377&psc=1&mcid=0d63bb60ad9533e38802e6f1b02421f6&tag=&ref=&adgrpid=61525452389&hvpone=&hvptwo=&hvadid=312152817250&hvpos=&hvnetw=g&hvrand=7428861331032207506&hvqmt=&hvdev=m&hvdvcmdl=&hvlocint=&hvlocphy=9027607&hvtargid=pla-458962573377&gclid=Cj0KCQiA7OqrBhD9ARIsAK3UXh044f43RRYBEr-9Y-4viFS1j9IV1h8kRohOb7i-GPC0jYSWjSWapdYaApymEALw_wcB
« Last Edit: December 14, 2023, 09:45:58 PM by CW7491 »

Offline Buick

  • Senior Member
  • **
  • Posts: 99
    • KWABELLI466
Re: ASOS/AWOS Accuracy
« Reply #22 on: December 14, 2023, 11:18:26 PM »
Thank you! How typical do you think a situation similar to mine occurs?
VP2 (6163) w/ 7345.425 & 24hr FARS
6410 @ 26' AGL, 12' above rooftop
6415 @ 12' AGL
7714
Rainwise 111
Stratus
WiFiLogger2
WeatherLink Live

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #23 on: December 14, 2023, 11:36:43 PM »
It doesnt seem all that typical especially because what youre seeing is the ASOS running cooler. I think almost all the examples Ive heard of its the other way around

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 205
Re: ASOS/AWOS Accuracy
« Reply #24 on: December 15, 2023, 10:06:31 AM »
The other thing I forgot to mention is you are seeing KBLIs minimum temp being up to 3F lower than your station. In the example I gave, if the temperature minimum was 9.4C and it sat there for hours but got no lower, the ASOS would report 48F for all that time during the 5 min obs, but would still record the correct minimum temperature of 49F!

 

anything