[production/RRFS.v1] Reorganization of NWGES directory structure and removal of SLASH_ENSMEM_SUBDIR

[BenjaminBlake-NOAA]

DESCRIPTION OF CHANGES:

• When I started working on reorganizing the run directory structure for the remaining DA tasks in the RRFS workflow, I realized that two things also need to be modified: the NWGES directory structure, and the removal of SLASH_ENSMEM_SUBDIR. This PR does that for the non-DA and firewx workflows to keep things simple.
• The NWGES directories are re-structured to be more similar to COMOUT, so the cycle-specific directories are replaced by ${RUN}.${PDY}/${cyc} directories. The filenames in nwges will also contain ${PDY}${cyc} in them. To see an example of this on Cactus, you can look in /lfs/h2/emc/ptmp/Benjamin.Blake/test_nonDA_netcdf/nwges. In addition, $GESROOT is now used instead of ${NWGES} for the path to the nwges directory.
• The SLASH_ENSMEM_SUBDIR variable is removed. If the task/directory is for an RRFS ensemble member, the member number will be incorporated into the directory name itself using the ${mem_num} variable.
• Also included in this PR are the changes from PR Turn off FAA grib2 output for ensemble forecasts #353 which have not been added to production/RRFS.v1 yet (exrrfs_run_prdgen.sh changes).

TESTS CONDUCTED:

These changes were tested with the non-DA engineering test and the RRFS fire weather workflow.

Machines/Platforms:

• WCOSS2
  • Cactus/Dogwood
  • Acorn
• RDHPCS
  • Hera
  • Jet
  • Orion
  • Hercules

Test cases:

• Engineering tests
  • Non-DA engineering test
  • DA engineering test
    • Retro
    • Ensemble
    • Parallel
• RRFS fire weather
• RRFS_A:
• RRFS_B:
• RTMA:
• Others:

ISSUE:

• Continues work on issue Revamp RRFS run directory structure  #237

[BenjaminBlake-NOAA]

@MatthewPyle-NOAA Having a NWGES directory structure similar to COMOUT also makes sense to me if we want to go that route. So for example we could have directories within nwges like this instead of the current ${PDY}${cyc} directories:
rrfs.${PDY}/${cyc}
enkfrrfs.${PDY}/${cyc}
refs.${PDY}/${cyc}

[MatthewPyle-NOAA]

@BenjaminBlake-NOAA I like that overall structure

[BenjaminBlake-NOAA]

@MatthewPyle-NOAA Sounds good to me as well. I've updated my PR to reflect that structure and tested the changes.

[MatthewPyle-NOAA]

FWIW, I do use a GESROOT definition in the HREF system. I know I was surprised that NCO doesn't really talk about it in the production standards - need to dig out some old e-mail conversations with Steven Earle about it. But in HREF I do stuff like:

export GESROOT=${GESROOT:-$(compath.py $envir/$NET/${href_ver})/nwges}

[BenjaminBlake-NOAA]

I think using GESROOT makes sense and I'll change NWGES to GESROOT. The compath.py stuff can be added later but that should be the eventual goal.

[MatthewPyle-NOAA]

Generally looked good to me - but did have a comment about spinup cycles.

[MatthewPyle-NOAA]

One comment that applies to RUN_POST as well - I don't believe there are "spinup" cycles for the ensemble forecasts.

[BenjaminBlake-NOAA]

You're right, thanks for catching that. I'll update those J-jobs.

[MatthewPyle-NOAA]

@BenjaminBlake-NOAA Did you retest with your latest changes in this PR?

[BenjaminBlake-NOAA]

@MatthewPyle-NOAA I did test the GESROOT changes and they worked well, but I didn't run any tests for the more recent spinup changes. Would you like me to run any more tests for those?

[MatthewPyle-NOAA]

No need to run more tests at this point - will approve and get it merged.

[MatthewPyle-NOAA]

Looks good, but a complete test will require testing the full system. So will merge now so as not to hold things up.