Author Topic: CWOP Data Analysis Inconsistencies...  (Read 2873 times)

0 Members and 1 Guest are viewing this topic.

Offline W3DRM

  • Forecaster
  • *****
  • Posts: 3360
    • Emmett Weather
CWOP Data Analysis Inconsistencies...
« on: August 01, 2017, 01:58:33 PM »
Yes, there are many posts on this forum regarding CWOP data, especially their barometric reports and yes, I have sent emails to the CWOP folks requesting an explanation for the barometric spikes I see in the analysis data that is used to validate "my data". I see the same issues with that of others who live close to me. Today, I get yet another CWOP report that says I am out of tolerance. Interestingly, looking at my data and that of another wx station 0.9 miles away from me and our local airport (KMEV - Minden Tahoe Airport). In both cases, I see a major and sudden increase in barometric pressure (over 8 milliBars in a few minutes) in the "analysis data" yet that of my data and that of others around me, show consistent values that, in my opinion are the correct values. I've attached a few screen captures showing what I am seeing.

The answer I always get back is that this discrepancy is due to faulty data being received in their "secret" data sampling. It would be nice if the CWOP algorithms could weed-out sample data that is grossly different than other data in the given geographical area. Doing so would certainly help with all of the spikes we see in barometric pressures across the board.

Sorry for the rant but it just seems strange that only the barometric data exhibits this behavior. See my screen images below:
Don - W3DRM - Emmett, Idaho --- Blitzortung ID: 808 --- FlightRadar24 ID: F-KBOI7
Davis Wireless VP2, WD 10.37s150,
StartWatch, VirtualVP, VPLive, Win10 Pro
--- Logitech HD Pro C920 webcam (off-line)
--- RIPE Atlas Probe - 32849

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
Re: CWOP Data Analysis Inconsistencies...
« Reply #1 on: August 01, 2017, 04:20:15 PM »
Agree wish it would work like that but it doesn't. The system is flawed and shows especially with pressure.
 
You have couple nearby running high pressure.
http://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?stn=AT077&unit=0&timetype=LOCAL
http://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?stn=E9139&unit=0&timetype=LOCAL
Randy

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: CWOP Data Analysis Inconsistencies...
« Reply #2 on: August 01, 2017, 06:31:24 PM »
Well, when half the stations in your area submit mean sea level pressure (MSLP) and the other half altimeter (QNH), the analysis algorithm is going to bounce around quite a bit. Garbage in, garbage out. That's the situation we have because many software developers are either unaware or don't care that altimeter has been the CWOP requirement since at least 2005 according to multiple CWOP documents.

Then there's this little nugget from the NWS regarding PWS selection for CWOP (https://www.weather.gov/media/epz/mesonet/CWOP-PWSintro.pdf):

Quote
Due to constant problems with calibration and confusion over what type of Barometric Pressure is reported by a PWS, this weather element is of lesser importance when considering a PWS.

Nice!

So... even when CWOP users' weather station software correctly submits QNH pressure, oftentimes their sensor is completely uncalibrated. I had to put in a 2mb (0.06") offset into my 1 yr old Davis VP2 console to get it to sync up properly with the airport (but it stayed sync'd after that). Unfortunately, my software only submits MSLP to CWOP so I choose to omit that from my feed to CWOP until such a time as QNH is available (the developer considers compliance a "feature request").
« Last Edit: August 01, 2017, 06:43:49 PM by openvista »
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: CWOP Data Analysis Inconsistencies...
« Reply #3 on: August 01, 2017, 07:19:01 PM »

 I had to put in a 2mb (0.06") offset into my 1 yr old Davis VP2 console to get it to sync up properly with the airport (but it stayed sync'd after that). Unfortunately, my software only submits MSLP to CWOP so I choose to omit that from my feed to CWOP until such a time as QNH is available (the developer considers compliance a "feature request").
The VP2 console with a "zero" elevation preset will mimic the altimeter to an extent, but will increasingly show it's inaccuracy as the upper and lower spectrum of pressure is being observed. The Vue console has a true altimeter and is well worth the consideration if that's important to you. Your software is obviously another issue, pressure is obviously not considered "important" enough to be consistent as the other obs.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: CWOP Data Analysis Inconsistencies...
« Reply #4 on: August 01, 2017, 08:01:24 PM »
If I suddenly switch my hardware settings to altimeter, then it destroys the continuity of my database (which also appears on my website). How would I know in the future if I've set a pressure record for a month, season, year or all time? I wouldn't.

Asking a user to choose one or the other (MSLP vs QNH) is totally unnecessary when there are publicly available formulas for converting VP2 MSLP to altimeter. This can and has been done correctly for years by certain apps like Cumulus and WeeWX.

Bottom line: it's incumbent upon developers to comply with the specification if they are going to transmit data to CWOP.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: CWOP Data Analysis Inconsistencies...
« Reply #5 on: August 01, 2017, 08:40:45 PM »
If I suddenly switch my hardware settings to altimeter, then it destroys the continuity of my database (which also appears on my website). How would I know in the future if I've set a pressure record for a month, season, year or all time? I wouldn't.

Asking a user to choose one or the other (MSLP vs QNH) is totally unnecessary when there are publicly available formulas for converting VP2 MSLP to altimeter. This can and has been done correctly for years by certain apps like Cumulus and WeeWX.

Bottom line: it's incumbent upon developers to comply with the specification if they are going to transmit data to CWOP.
Continuity is a personal choice for your PWS, if I change something of relevance, I make note for future reference and have done so numerous times but understand your records are now on a different page. I choose to find what is MY happy place is for MY consumption and let the rest fall as it may with the inconsistencies. I gave up months ago trying to get WU to leave my altimeter alone, they insist on MSLP, so be it, not worth the time and BS until greater heads prevail.  :roll:

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: CWOP Data Analysis Inconsistencies...
« Reply #6 on: August 01, 2017, 10:34:53 PM »
Continuity is a personal choice for your PWS....

I suppose.

Perhaps you mean that you value altimeter above MSLP, and when you discovered the workaround (zeroing out elevation), you made the leap. If so, fair enough. That's something different than CWOP compliance vs. database integrity.

When I look on a weather map, however, I don't see altimeter. I see SLP. That's what meteorologists use to make forecasts. Given a properly calibrated sensor, I should be able to plot my SLP reading on a current weather map with a fair degree of accuracy (assuming isobars aren't too compressed and temperature assumptions are correct). Indeed, I can do that and that means the data I provide to the public via my website and other providers can be used for accurate forecasting, modeling and ground truth verification.

In any event, and not to put too fine a point on it, no one should have to choose a pressure type at the hardware level just to accommodate CWOP given how long open source software solutions have been available. Suppose, for example, that CWOP expected temp in Celsius. Should the user have to change their console to Celsius just to accommodate a lazy developer who can't be troubled to convert from Fahrenheit before sending to CWOP? I suppose one could argue that preserving past temperatures is a personal choice, especially given how universal Celsius measurement is globally. I would counter argue that's a tangential issue in this particular discussion.
« Last Edit: August 01, 2017, 10:52:44 PM by openvista »
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: CWOP Data Analysis Inconsistencies...
« Reply #7 on: August 01, 2017, 11:02:53 PM »
Continuity is a personal choice for your PWS....
Perhaps you mean that you value altimeter above MSLP, and when you discovered the workaround (zeroing out elevation), you made the leap.
Yes, that was my workaround with the VP2. The reason I got the Vue is for the exact reason of continuity. Official observations and records, like it or not, are kept at airports, and they use the altimeter. I like an even playing field when I can get it. With that, I certainly do get your gripes.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: CWOP Data Analysis Inconsistencies...
« Reply #8 on: August 02, 2017, 12:29:23 AM »
Not all official records are kept at airports. They certainly aren't at my local WFO.

Pressure records are reported rather inconsistently varying according to the custom of the local WFO. While I assume all offices measure pressure, they don't always report those values in their record summaries. When they are reported, even if the official measurement is taken at the airport, the observations are ultimately converted to SLP before being officially recorded. If you're a forecasting office charged with warning the public of dangerous approaching weather, it would make NO sense to record altimeter which always assumes standard atmospheric conditions -- not a useful assumption when tracking cold and warm fronts let alone tornadoes or hurricanes.

If you want to see your local airport's pressure converted to SLP by the NWS go here: http://www.wrh.noaa.gov/mesowest/getobext.php?table=1&wfo=mqt&sid=KTUS

References supporting SLP in official and preliminary records:
http://iwin.nws.noaa.gov/view/prodsByState.php?state=MI&prodtype=climate (do a CTRL-F text search for "SLP")
http://www.crh.noaa.gov/images/dlh/StormSummaries/2010/october26/pressurerecord.pdf
https://twitter.com/NWSGrandForks/status/839498750443864065
« Last Edit: August 02, 2017, 12:38:44 AM by openvista »
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Re: CWOP Data Analysis Inconsistencies...
« Reply #9 on: August 02, 2017, 12:49:22 AM »
Not all official records are kept at airports. They certainly aren't at my local WFO.

Pressure records are reported rather inconsistently varying according to the custom of the local WFO. While I assume all offices measure pressure, they don't always report those values in their record summaries. When they are reported, even if the official measurement is taken at the airport, the observations are ultimately converted to SLP in millibars (the worldwide standard for weather maps) before being officially recorded. If you're a forecasting office charged with warning the public of dangerous approaching weather, it would make NO sense to record altimeter which always assumes standard atmospheric conditions -- not a useful assumption when tracking cold and warm fronts let alone tornadoes or hurricanes.

If you want to see your local airport's pressure converted to SLP by the NWS go here: http://www.wrh.noaa.gov/mesowest/getobext.php?table=1&wfo=mqt&sid=KTUS

References supporting SLP in official and preliminary records:
http://w2.weather.gov/climate/f6.php (do a CTRL-F text search for "SLP")
http://www.crh.noaa.gov/images/dlh/StormSummaries/2010/october26/pressurerecord.pdf
https://twitter.com/NWSGrandForks/status/839498750443864065
All I'm saying is I prefer to use the altimeter because I have a direct, close proximity, quick and accurate comparison with the ASOS's in my area, as I'm familiar with the KTUS link and use it daily for said comparisons. One stop shopping. http://www.wrh.noaa.gov/total_forecast/other_obs.php?wfo=twc&zone=AZZ504