Welcome, Guest
Username: Password: Remember me

Here we can describe more what should be posted here

TOPIC: missing German obs in cy40

missing German obs in cy40 1 week 2 days ago #2448

  • Bert van Ulft
  • Bert van Ulft's Avatar
  • Expert Boarder
  • Posts: 102
  • Thank you received: 22
Dear all,

a reanalysis experiment at the ECMWF, based on harmonie-40h1.2.tg2 shows a drop in the number of synop observations over Germany on the 23rd of September 2019. To try to solve this I copied the MARS requests from cy43 (develop) Prepare_ob to Oulan by replacing the OBSTYPE for synop/buoy with:


Running this manually the German synop observations are available in the ob file again when viewed with codes_ui and also in the Harmonie run the Oulan task seems to retrieve the observations well. However, the observations are then filtered out again and in the end the run (from Bator) gives exactly the same output as before.
Does anybody know what would need to be changed to make Harmonie use these new observations?
We are stuck to cy40 as it is a continuation of a previous project.
Concatenated log files are on ecgate in /scratch/ms/nl/nkl/hm_home/WINS50_40h12tg2_fERA5_CTL_ptA/archive/log. The files with runtime in the file name are with the original cy40 MARS retrieval, the ones without have the cy43-style retrieval (e.g. HM_Date_2019092409_20200602-092710.html vs. HM_Date_2019092409.html.

best wishes,


ps. the problem seems similar to Erik's post www.hirlam.org/index.php/forum/2-harmoni...ri-canaco-f90-at-cca, but I did not find a way forward there.

missing German obs in cy40 1 week 2 days ago #2449

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • Gold Boarder
  • Posts: 209
  • Thank you received: 41
Hi Bert,

I will begin with "I wouldn't start from here" as many Irish men say when they are giving directions!

The handling of BUFR and TAC data is not straightforward. With Oulan you can tell it to use "BUFR" reports only with LNEWSYNOPBUFR, LNEWBUOYBUFR, ... set in scr/include.ass. Unfortunately these switches only provide processing for "TAC" or "BUFR" not both.

I would suggest using Bator to process conventional BUFR and let CARRA/ECMWF blacklisting decide on which data to use.

No quick fix, just more work I'm afraid!

The following user(s) said Thank You: Bert van Ulft
Time to create page: 0.079 seconds