fhzero configurations in GEFSv13 prototype run

[junwang-noaa]

Description

The atmosphere accumulation bucket was updated in FV3TAM PR #835. The bucket in GEFSv13 restart run looks not correct, further investigation is needed to make the feature work for GEFSv13.

To Reproduce:

Additional context

Output

[junwang-noaa]

GEFSv13 run directory: /lfs/h2/emc/gefstemp/Bing.Fu/fcst.75320

[lisa-bengtsson]

Is this a new issue in all applications or just the ensemble configuration?

[junwang-noaa]

@lisa-bengtsson No, we are looking at the test case when GEFSv13 restarts at fh=03 (not a multiple of fhzeros) and has bucket of 3hr for fh=3-6hr and 6hr for the rest of forecast time. I don't think other tests use his configuration.

[junwang-noaa]

@bingfu-NOAA, I checked the run output in /lfs/h2/emc/gefstemp/Bing.Fu/fcst.75320. In the run output file, I saw:

fhzero : 3.00000000000000
fhzero_array : 3.00000000000000 6.00000000000000
fhzero_fhour : 6.00000000000000 0.000000000000000E+000

Would you please change the 2nd fhzero_fhour to a large number, such as:
&gfs_physics_nml
fhzero = 6
fhzero_array = 3,6
fhzero_fhour = 6, 2400

The fhzero will keep having value 3 if the second element of fhzero_fhour is zero. With the change, we have:

at the first step:

 ---isec,seconds         300       11100
   gfs diags time since last bucket empty:   8.333333333333333E-002 hrs
 in atmos_model update, fhzero=   3.00000000000000      fhour=   3.083333
   3.000000

at fh=06:

 ---isec,seconds       10800       21600
  gfs diags time since last bucket empty:    3.00000000000000      hrs
 in atmos_model update, fhzero=   3.00000000000000      fhour=   6.000000
   6.000000

at fh=09:

 ---isec,seconds       10800       32400
  gfs diags time since last bucket empty:    3.00000000000000      hrs
 in atmos_model update, fhzero=   6.00000000000000      fhour=   9.000000
   6.000000

at fh=12

---isec,seconds       21600       43200
  gfs diags time since last bucket empty:    6.00000000000000      hrs
 in atmos_model update, fhzero=   6.00000000000000      fhour=   12.00000
   12.00000

[bingfu-NOAA]

@lisa-bengtsson please check out.old to see the original log file. The log file you saw are from some tests I made, please ignore it.

[bingfu-NOAA]

@junwang-noaa I will follow your suggestion to take another test. Do you think this can be resolved by setting specific values for the newly-added fhzero* parameters?

[junwang-noaa]

@bingfu-NOAA I made a fix in FV3 PR#857. With this fix, you don't need to use fhzero_array/fhzero_fhour. The previous fhzero=6 in input.nml should still work as before. Please let me know if you still have issues. Thanks

[bingfu-NOAA]

@junwang-noaa Thanks Jun. We will do some tests and let you know if this update can fix the issue.

[bingfu-NOAA]

@junwang-noaa Our tests now show correct bucket after applying this fix.

[junwang-noaa]

@bingfu-NOAA Thanks for testing. The code fix is now combined with fv3atm PR#856 and will be committed in UFS WM PR#2355.

[junwang-noaa]

The PR was committed in PR#2355. Closing the issue.