The Weather Forecasting Forum
http://www.weathergraphics.com/forum/

Double 300 mb station plots
http://www.weathergraphics.com/forum/viewtopic.php?f=5&t=1531
Page 1 of 1

Author:  Fred Capello [ Sun Jul 26, 2009 10:35 pm ]
Post subject:  Double 300 mb station plots

I don't know whether other users of DA out there have experienced this problem or not. For at least a couple of years now, whenever I plot upper air (radio sonde) data, some stations are plotted twice (over each other) at the 300 mb level with different data (of that same station). Every time I have to eliminate the wrong data, using the QC panel. It only occurs at the 300 mb level, so I suspect a bug. In the QC panel, the wrong data usually appears as -9999.0 in the temperature, dew point and height boxes.
Every time I find those "double plots", I have to delete the wrong ones and accept the valid entries in the QC panel, to make my 300 mb level chart appear OK. Any chance this "problem" can be corrected in the future?

Author:  Peter [ Mon Jul 27, 2009 5:41 am ]
Post subject:  Double 300 mb station plots

Fred,

This has nothing to do with DA programme.

If you view the station details using the Quality Control function as you have done then if the station is showing more than once then it is in the downloaded data more than once.

I have looked at 4 stations that exhibit this problem just north of your area and they are
WMO 91348, 91334, 91413 and 91376
When viewed using the QC function they all show several reports from the same station at 300mb level but the Wind Speed and Wind Direction on each report is different but the timing remains the same. Therefore this shows corrupted data from those stations.
Why this shows only at the 300mb level I have no idea.

So, DA can only process the data as it is and if there's more than one data entry from the station, it cannot do anything other than process that station data as it stands. DA does not decide which data is correct or not and processes all data.

That is the reason for the Quality Control function, to be able to manually exclude or delete station data that we can see is incorrect or corrupted.

Therefore DA is working correctly and the problem lies in the data downloaded and is caused by corrupted upload of the station data to the server.

Until that situation is corrected wherever the data is assembled, you will have to continue manually deleting any rouge or doubled up station data.

Corrupted data does occur often, as in the UK sometimes I see we have maximum temperatures of 67 deg C and minimum of -87 deg C sometimes!!!!

Regards

Author:  Peter [ Mon Jul 27, 2009 7:13 am ]
Post subject:  Double 300 mb station plots

Fred,

After I had posted the note below I realised I was looking at the wrong stations for your area but have looked at WMO 78988 and 80035 which I think are close to you and they show the problem as do many north of you.
But some stations don't exhibit that problem and therefore what I said below stands.

The fact that many stations worldwide do not show the problem indicates the problem is not in DA but in corrupted data in the download.

The Upper Air data data I have downloaded, looking at it in raw I can find several entries for the same station, all reported at the same time but with different temperatures and wind speeds.

Regards

Author:  Uwe Bergholter [ Sat Aug 01, 2009 3:13 pm ]
Post subject:  Double stations

Hi Fred and Peter,
double station reports are familar to me. Some (few) stations are notoriously transmitting two radiosonde reports in a web file, as Dar-el-Beida (WMO 60390). Funny enough the bulletins are different, as one is the A-part of the sounding (TTAA...) and the other one is the B-part. They have a different ascent times as well, which is a fake of course to separate them.
My solution is as follows: The data are filtered first. Then they are evaluated by RAOB and sent back to DA as custom data. This is to produce charts for my HP which DA cannot produce. It is done automatically by a homemade Visual Basic program several times a day (still under construction).
As double data are in a constant sequence (e.g. A-part first) a selection can be made during filtering. At any rate it can be assured that only one report makes it into the charts.
Regards, Uwe

Author:  Fred Capello [ Sat Aug 01, 2009 5:42 pm ]
Post subject:  Double 300 mb station plots

Hi Peter and Uwe,

Thanks for your reactions! I didn't find the time to react sooner to your replies. Indeed, Peter, you're right on the money! I'm from CuraƧao in the southern Caribbean, which actually has the WMO code 78988. Undoubtedly, the double plot wouldn't occur if there had been only one report from each station. As you stated in your first post, Peter, the mystery lies in the fact that this double plotting occurs only at the 300 mb level. So, I guess that these duplicate messages seem to be filtered out well at the other levels, but a possible bug in DA allows this double plotting to occur time and again at the 300 mb level. I think Tim has to investigate why this is occurring.
I normally plot the upper air data in the Caribbean Area and I extend this area into northern South America and the southeastern United States. I noticed that several (American) upper-air stations are plotted twice over Florida as well, so it's not restricted to Caribbean stations only. I also noticed that, when this double plotting occurs, the wind barbs in the data from stations without temperature, dew point and height differ substantially from the ones which do that data.

Kind regards.

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/