Topic: CWOP Barometric QC

[dasman]

I have read the excellent post about QC and not fretting over it, however I am looking at my most recent QC email telling me that my Barometer readings for yesterday had 143 fails out of 143 data sets.  That's right 0% correct.  The Barometer QC issues started 3-4 weeks ago and have gotten progressively worse.  I believe my 28 day average is down to 43% correct.  This is nuts!  Over the last 2 weeks I have tracked perfectly with the 2 airports by me, Lansing 20 miles North and Kankakee 20 miles south.  In fact the CWOP page says the difference I have with each is 0.01mb.  Yet if you look at Lansing airports CWOP page, it is at 99 or 100%.  I do see that Kankakee airport is getting a 28 day 61%.


So I looked at another airport 22 miles NW of me and my difference is again 0.01mb.  Finally I looked at the NWS office in Romeoville (KLOT) and again the difference is 0.01mb.

I emailed Philip Gladstone last weekend and he said something is definitely not right.  But I have not heard back since last Sunday.  And the QC of the data just keeps getting worse.

This goes beyond the bell curve talked about in the article.  It is obvious to me the data is solid and that CWOP's analysis in this area is... Questionable (Im trying to be nice).

[Faceng Ret]

CWOP analysis of my reported pressures has always been suspect to me. I have an VP 2, a Vue , an Envoy and an Envoy 8X that all are always in close agreement in pressure. My reported pressures are also in close agreement with two airports that are within 10 miles of my location. I believe that my reported pressures are very close and just ignore what CWOP says they should be. I opted out of getting the emails a year or two ago as I saw no value in the email and was not interested in chasing their numbers.

[ValentineWeather]

CWOP is an awesome tool for delivering our data but the QC, most noticeable pressure has issues.  The QC worked fairly well for years, but reading on their own CWOP thread I found they included weatherlink.com stations and ever since issues with QC have become more prevalent. It’s not just pressure but pressure is the one most noticeable because the bad data your station is being compared with is so far off.
I've decided for myself not to fret over it anymore.
 

[miraculon]

The State of Michigan DOT (MDOT) has set up a number of camera/weather towers across the state. The one closest to me, MC053, shows a "stuck" value of 29.88 inHg (1012mb). This was wreaking havoc on my QC analysis and "thumbs" score. As my own station departed further from this "stuck" value, the absolute error increased as did the "standard deviation" on the Analyis plot.

I complained to MDOT about it, but they said that they checked it and it is varying and is OK per their contractor. Curiously, the Wunderground plot shows exactly the same issue. Yet, when I go to the state's website the pressure shows what appears to be uncorrected pressure. If I compensate for altitude it is correct and varying with prevailing conditions. The state guys claim that they do not push the data to MADIS, they must be pulling it. (not sure where these databases are or how they can be accessed)

I also contacted MADIS-support at NWS. After some investigation, it appears that they have removed the offending station from my QC Analysis and my numbers have improved quite a bit. (two thumbs up, 100%) Kudos to NWS for fixing this, all I got from the State was "it's not our problem" followed by "crickets". The CWOP guys didn't seem too interested either...

These MDOT stations are all over Michigan, and any CWOP participants close to one would have been affected by this. (a fellow CWOPper did contact me directly about it and was seeing the same issue)

Greg H.