Add fix for reading sfc fluxes for DEPHYv1 format #428
Conversation
hertneky
requested review from
dustinswales,
grantfirl and
mkavulich
as code owners
| @@ -490,10 +490,10 @@ def setup_rundir(self): | |||
| if input_type == 1: | |||
| #open the case data file and read the surfaceForcing global attribute | |||
| case_data_dir = case_nml['case_config']['case_data_dir'] | |||
| nc_fid = Dataset(os.path.join(SCM_ROOT, case_data_dir) + '/' + self._case + '_SCM_driver.nc' , 'r') | |||
There was a problem hiding this comment.
I don't think we need these changes to lines 493-494? (Just the change to line 496)
There was a problem hiding this comment.
@dustinswales Thanks, I reverted the L493 change. I just tested and it is working fine (note I made that change awhile back due to some random issue reading the file). Thinking about it, I wonder if surface_forcing_moisture should also be considered here, since they can be set independently. I'm not sure how the SCM would react with them set differently without looking more at the code.
There was a problem hiding this comment.
@dustinswales Revisiting this reverted change. It looks like changes are needed here otherwise the fluxes are not read in correctly. case_nml['case_config']['case_data_dir'] is empty, so either it needs to be assigned the path somewhere above, OR remove L492 and modify L493 case_data_dir > self._case_data_dir. This only shows up when prescribing sfc_fluxes. While the case will appear to run normally without a fix, it will run the default suite and not the *_ps version since nothing is getting read correctly here.
|
For the MPACE case, the name of the attribute was 'surface_forcing_temp'.
So I changed line 494 and 496, but not 493.
if input_type == 1:
#open the case data file and read the surfaceForcing global
attribute
case_data_dir = case_nml['case_config']['case_data_dir']
nc_fid = Dataset(os.path.join(SCM_ROOT, case_data_dir) +
'/' + self._case + '_SCM_driver.nc' , 'r')
surfaceForcing = nc_fid.getncattr('*surface_forcing_temp*')
nc_fid.close()
if (surfaceForcing.lower() == 'flux' or
surfaceForcing.lower() == '*surface_flux*'):
surface_flux_spec = True
…On Thu, Jan 25, 2024 at 12:51 PM Dustin Swales ***@***.***> wrote:
***@***.**** requested changes on this pull request.
------------------------------
In scm/src/run_scm.py
<#428 (comment)>:
> @@ -490,10 +490,10 @@ def setup_rundir(self):
if input_type == 1:
#open the case data file and read the surfaceForcing global attribute
case_data_dir = case_nml['case_config']['case_data_dir']
- nc_fid = Dataset(os.path.join(SCM_ROOT, case_data_dir) + '/' + self._case + '_SCM_driver.nc' , 'r')
I don't think we need these changes to lines 493-494? (Just the change to
line 496)
—
Reply to this email directly, view it on GitHub
<#428 (review)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADMPXGJDL2WPVOZQWUO75QDYQKZULAVCNFSM6AAAAABCJQFI6OVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMYTQNBUGU3DQNJWGA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
@egrell Sorry, I see in the DEPHY standard that its "surface_forcing_temp". |
hertneky
changed the title
…_LEG04A caused by now reading the case_data_dir correctly from previous commit
|
Closing due to changes being combined into #459. |
|
@grantfirl The only remaining question was...do we need to do something different since in DEPHY, the surface flux flags are separate > "surface_forcing_temp" and "surface_forcing_moist". Right now, only "surface_forcing_temp" is checked to determine whether surface fluxes are prescribed or not. |
* upstream/main: Add GP SDFs/NMLs to SCM minor fix/cleanup Add namelist option and sigmab (tracer) for HR3 add SCM_RRFS_v1 SDF, nml, tracer file; update suite_info; update scm_type_defs to recgonize new tracers add script to download and stage climatological aerosol data and fix typo in HR3 namelist merge in changes from NCAR#428 from @hertneky add CITATION.cff file from @scrasmussen from NCAR#458 merge changes from @hertneky in NCAR#446 add HR3 suite and physics namelist from NCAR#445 from @bluefinweiwei changes from @bluefinweiwei for PR#419 Revert submodules Update physics Address reviewers comments update ccpp/physics submodule pointer Moving documentation to ReadTheDocs (NCAR#438) Synced physics update ccpp/physics submodule update ccpp/framework update CCPP_typedefs to work with PR#181 update ccpp/physics and .gitmodules Added step to skip gridpoint in ensemble if error during case generation Added utility script to create input file containing replay points Omission from previous commit Add capability to use input file with longitude and latitude lists update ccpp/physics submodule update .gitmodules for whitespace change removal update ccpp/physics and .gitmodules update ccpp/physics and .gitmodules update ccpp/physics submodule pointer point .gitmodules back to my branch for CI testing update ccpp/physics and .gitmodules update ccpp/physics and .gitmodules update ccpp/physics and .gitmodules Cleanup/Housekeeping. Remove unnecessary interpolation step Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Revert change to CMakeLists Revert change to CMakeLists Revert change to CMakeLists Revert change to CMakeLists Update CI Update CI Add file for Nvidia RTS Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update cmakelists Updated CI Update CI Update CI Update CI Update CI Update CMakeLists for Nvidia support Update CMakeLists for Nvidia support Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Update CI Foundatio for Nvidia CI script Synced physics Small cahnge from physics needed for replay Replay with RUC and NoahMP LSM working Some changes for Man to test If using model initial conditions and forcing from the UFS (e.g. Replay mode) with an active LSM, the SCM performs a cold-start initialization for the LSM. That is, UFS replay mode currently support cases generated from UFS Restart files (e.g spun up (warm) LSM parameters) update ccpp/physics and revert .gitmodules update ccpp/physics submodule and revert .gitmodules changes to work with ufs/dev PR#155 update for ufs/dev PR#94 update ccpp/physics and .gitmodules update ccpp/physics and .gitmodules after testing
These changes address a bug found when running SCM with DEHPYv1 cases where surface fluxes are prescribed. This addresses Issue #427