Check the results of the energy fixer after running YOG

[jatkinson1000]

It would be good to see how much (if any) lifting the energy fixer is having to do after the YOG parameterisation is run.

Diagnostics should be output to the CAM log files which can be used to check this.

[jatkinson1000]

This goes to the logfiles and is output.
Judith can provide guidance but look at the check_energy subroutine.

[MarionBWeinzierl]

@paogorman said that this is lower priority for us, as it involves more science (might be picked up by research associate in the new year?).

[paogorman]

Yes, this could be a lower priority than other issues. One caveat: checking the energy fixer for YOG could be a way to find a bug in the code if it flagged that energy was clearly not conserved.

[gretamiller]

Yes, Jack pointed out the line I think Judith was referring to:

CAM-ML/src/physics/cam/physpkg.F90

Line 1975 in 4a4a0a6

call check_energy_chng(state, tend, "chkengyfix", nstep, ztodt, zero, zero, zero, flx_heat)

...but I'm not sure about log files for this

[MarionBWeinzierl]

Information about the dycore energy fixer can be found here: https://www.cesm.ucar.edu/models/simple/change-dycore . @judithberner mentioned, though, that we are only interested in the physics energy fixer.

[MarionBWeinzierl]

Doing a grep of EFIX, as was mentioned at the above link, shows that this field is output in the logs/atm.log* files.

[MarionBWeinzierl]

I output the EFIX field for the YOG simulation:

Is this the kind of output we would expect?

[paogorman]

Thanks Marion. That's a larger energy correction than I would expect for a GCM, but it would be a good idea to try it with the ZM configuration to see what you get for comparison. ps: Just check: Is this the output of the energy fixer for all processes? (not just YOG)

…

On Wed, Nov 20, 2024 at 11:28 AM Marion ***@***.***> wrote: I output the EFIX field for the YOG simulation: image.png (view on web) <https://github.com/user-attachments/assets/8aa208e3-0935-411b-89d9-e64c8acd299c> Is this the kind of output we would expect? — Reply to this email directly, view it on GitHub <#34 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARPLW6ZHNIAEIXVTAB34JUD2BS2CLAVCNFSM6AAAAABMQSZMMGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOBZGA2DKOBWGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[MarionBWeinzierl]

This is without the bugfix that Jack has done, btw. I believe it is for all processes -- they seem to all write to the same EFIX field.

This is the file with axis etc:

[paogorman]

Hi Marion, I'm not completely sure, but I think the EFIX you are outputting is related to the finite volume dynamical core rather than the parameterizations. Apparently the FV core has about a 2 W/m^2 energy loss due to discretization of the momentum equations. To get the code to write out information about the energy conservation of YOG, I think we need to set the parameter "print_energy_errors" in the check_energy_nl namelist to .true. You can see how this information is written out by looking at the check_energy_chng subroutine in check_energy.f90. Also, in the code below in physpkg.f90, I would change "chkengyfix" to "yog_nn" so that what is written out to the logfiles has a name to identify yog. if (yog_scheme=='on') then call t_startf('yog_nn') call yog_tend(ztodt, state, ptend, pbuf) call physics_update(state, ptend, ztodt, tend) call t_stopf('yog_nn') flx_cnd(:ncol) = prec_dp(:ncol) call check_energy_chng(state, tend, "chkengyfix", nstep, ztodt, zero, flx_cnd, zero, flx_heat) end if

…

On Wed, Nov 20, 2024 at 12:03 PM Marion ***@***.***> wrote: This is without the bugfix that Jack has done, btw. I believe it is for all processes -- they seem to all write to the same EFIX field. This is the file with axis etc: image.png (view on web) <https://github.com/user-attachments/assets/b73744ed-dab6-4c81-97e9-faa507594994> — Reply to this email directly, view it on GitHub <#34 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARPLW62BYNBZ3HJ32GPDQ2D2BS6FVAVCNFSM6AAAAABMQSZMMGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOBZGEZDKNRUG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[jatkinson1000]

From a look into the source of check_energy_chng I believe one needs to set a namelist variable print_energy_errors = .true. to enable printing of any columns violating the energy fixer tolerance to the logfile (iulog). This will include the name of the check where it occurs which is patched in #51