Welcome, Guest
Username: Password: Remember me
HELPDESK

Here we can describe more what should be posted here

TOPIC: 10m wind fullpos vs history

10m wind fullpos vs history 4 weeks 1 day ago #2335

  • Bert van Ulft
  • Bert van Ulft's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 96
  • Thank you received: 22
Dear all,

in cycle 40 HARMONIE runs the 10m wind components CLSVENT.ZONAL and CLSVENT.MERIDIEN differ between the fullpos (PFHARM*) and history (ICMSHHARM*) files. Other fields I compared are either identical (accumulated rain + low clouds) or there is just some random noise (surface and 2m temperature). The image below shows the difference in U10m (fullpos - history) at +01 into the forecast:

diff_u10_fp-his.png

The difference in files I looked at is at most 0.18m/s. The values change during the forecast, but the range remains similar. In general it looks like the difference is around 0 in a band in the middle of the domain and that the difference increases towards the top (North) and bottom (South) of the domain.
I ran into this problem while setting up a run with sub-hourly output, using fields from the history files at full hours and from fullpos files with a small subset of fields at 10 minute intervals. But knowing this difference exists mixing fields from history and fullpos files seems a bad idea.
Does anybody know what could be causing this difference and which one can best be used?


best wishes,

Bert
Last Edit: 4 weeks 11 hours ago by Bert van Ulft. Reason: Added N-S effect

10m wind fullpos vs history 3 weeks 1 day ago #2336

  • Ulf Andrae
  • Ulf Andrae's Avatar
  • OFFLINE
  • Administrator
  • Posts: 273
  • Thank you received: 29
Bert,

My guess is that 10m wind is re-diagnosed via the call to FPACHMT. Don't know if there is a possibility not to do it.

Ulf

10m wind fullpos vs history 3 weeks 1 day ago #2337

  • Bert van Ulft
  • Bert van Ulft's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 96
  • Thank you received: 22
Thanks Ulf, seems like a good guess.
There is a switch LLHMT (LDHMT) in the code which avoids (re-)calculation of the fields, but I don't know if it breaks other parts if I simply switch it off. Also, calculations work fine for e.g. t2m, so the calculation of u10m and v10m might be incorrect for some reason. But at least now I have an idea where to start looking.

best wishes,

Bert
Time to create page: 0.269 seconds