CCPPize dme_adjust

[cacraigucar]

Originator(s): cacraig

Description (include issue title and the keyword ['closes', 'fixes', 'resolves'] and issue number):
closes #103 - CCPP-ized version of physics_dme_adjust

List all namelist files that were added or changed: N/A

List all files eliminated and why: (created conservation_adjust subdirectory and moved check_energy subdirectory into it)

R100 schemes/check_energy/check_energy_chng.F90 schemes/conservation_adjust/check_energy/check_energy_chng.F90
R100 schemes/check_energy/check_energy_chng.meta schemes/conservation_adjust/check_energy/check_energy_chng.meta
R100 schemes/check_energy/check_energy_chng_namelist.xml schemes/conservation_adjust/check_energy/check_energy_chng_namelist.xml
R100 schemes/check_energy/check_energy_fix.F90 schemes/conservation_adjust/check_energy/check_energy_fix.F90
R100 schemes/check_energy/check_energy_fix.meta schemes/conservation_adjust/check_energy/check_energy_fix.meta
R100 schemes/check_energy/check_energy_gmean/check_energy_gmean.F90 schemes/conservation_adjust/check_energy/check_energy_gmean/check_energy_gmean.F90
R100 schemes/check_energy/check_energy_gmean/check_energy_gmean.meta schemes/conservation_adjust/check_energy/check_energy_gmean/check_energy_gmean.meta
R100 schemes/check_energy/check_energy_save_teout.F90 schemes/conservation_adjust/check_energy/check_energy_save_teout.F90
R100 schemes/check_energy/check_energy_save_teout.meta schemes/conservation_adjust/check_energy/check_energy_save_teout.meta
R100 schemes/check_energy/check_energy_scaling.F90 schemes/conservation_adjust/check_energy/check_energy_scaling.F90
R100 schemes/check_energy/check_energy_scaling.meta schemes/conservation_adjust/check_energy/check_energy_scaling.meta
R100 schemes/check_energy/check_energy_zero_fluxes.F90 schemes/conservation_adjust/check_energy/check_energy_zero_fluxes.F90
R100 schemes/check_energy/check_energy_zero_fluxes.meta schemes/conservation_adjust/check_energy/check_energy_zero_fluxes.meta
R100 schemes/check_energy/dycore_energy_consistency_adjust.F90 schemes/conservation_adjust/check_energy/dycore_energy_consistency_adjust.F90
R100 schemes/check_energy/dycore_energy_consistency_adjust.meta schemes/conservation_adjust/check_energy/dycore_energy_consistency_adjust.meta

List all files added and what they do:
A test/test_suites/suite_dme_adjust.xml
A schemes/conservation_adjust/dme_adjust/dme_adjust.F90
A schemes/conservation_adjust/dme_adjust/dme_adjust.meta

List all existing files that have been modified, and describe the changes:
(Helpful git command: git diff --name-status development...<your_branch_name>)
M suites/suite_cam7.xml
M suites/suite_kessler.xml
M suites/suite_tj2016.xml

List all automated tests that failed, as well as an explanation for why they weren't fixed:

Is this an answer-changing PR? If so, is it a new physics package, algorithm change, tuning change, etc?

If yes to the above question, describe how this code was validated with the new/modified features:

[jimmielin]

Thanks @cacraigucar! I had two clarification questions but maybe they are outside the scope of the PR, I am not sure, depending what the plans for dme_adjust in SIMA are.

[jimmielin]

I suppose these are currently read from the snapshot but when SIMA runs standalone there might be the need for an init routine (or change in CAM-SIMA if q/liq/ice needs to be recorded at somewhere outside of physics) to initialize those?

[cacraigucar]

@nusbaume - what are your thoughts on this? These variables are actually initialized in tphysbc (not the tphysac layer where the run routine is called). I'd thought about writing an initialization routine, but this would require having a separate module for it. When I mentioned there are going to be a number of little routines that may need to be added to the main SDF's when we have all the physics translated, this is a concrete example of what I meant. I figure translating the limited code which resides in tphysac and tphysbc will be one of the final steps.

[jimmielin]

I am happy to discuss this in more detail (including whether to do this conversion now or later) but I realized I bumped into a similar issue in rk_stratiform. The "trick" I used (to permit this field to be both readable from a snapshot and initializable in a full CAM-SIMA run) is to set <initial_value>-1.0_kind_phys</initial_value> and detect this negative value in the (timestep?) initialization routine of this scheme, or if you need it to be outside of physics since it's for "before coupling", it could be in SIMA itself. In fact, rk_stratiform.F90 in ESCOMP/CAM already does this (but uses HUGE(1.0_r8) for safety)

   ! check that qcwat and tcwat were initialized; if not then do it now.
   if (qcwat(1,1) == huge(1._r8)) then
      qcwat(:ncol,:) = state%q(:ncol,:,1)
   end if
   if (tcwat(1,1) == huge(1._r8)) then
      tcwat(:ncol,:) = state%t(:ncol,:)
   end if

Since ESCOMP/CAM-SIMA#365 was developed, the model is capable of reading the actual, physical value from the snapshot when available (and thus no negatives), and if not (in an actual run), it'll have a negative value you can capture at timestep_init (or elsewhere) and properly initialize.

[jimmielin]

I presume that this is not enabled because CAM-SIMA cannot run dme_adjust standalone due to qini/liqini/iceini not being set, could you confirm? It would be good to make a note of the reason why this will remain disabled, after this scheme is available, to signal that this is intentional.

[jimmielin]

Hi @cacraigucar thanks for adding the comment, I think you may have missed a ! in the xml comment syntax (should be <!-- starting instead of <--)

[jimmielin]

Thanks @cacraigucar! just a couple of minor changes while we wait for discussion on the qini/liqini/iceini

[jimmielin]

Hi @cacraigucar thanks for adding the comment, I think you may have missed a ! in the xml comment syntax (should be <!-- starting instead of <--)

[jimmielin]

Same issue regarding the comment syntax here (should be <!-- instead of <--)