Simplify changes to the generic tracer interface for cobalt-p4

[yichengt900]

The new cobalt-p4 requires several variables from MOM6 (e.g. geolat for day-length calculation as well as eqs_of_state for density calculation).

We revisited the original changes and removed unnecessary changes to minimize the generic tracer interface changes at MOM6 side. The results match those obtained from the original codes in the 1-D and NWA12 RT tests.

Dependency: NOAA-CEFI-Regional-Ocean-Modeling/ocean_BGC#8

[charliestock]

Hi YC, thanks. It looks the issue was that my original changes to get geolat and the information required for the equation of state calculation were, for lack of a better descriptor, "overkill". Your code modifications make this change much less disruptive. Thank you.

If I understand correctly, you were able to get these efficiencies by using the MOM_EOS variable rather than passing a broader set of thermodynamic variables, and by only passing them to "generic_tracer_source" subroutine in MOM_generic_tracer.F90. Am I right in thinking that the interface changes will thus be limited to this latter change?

best, Charlie

[charliestock]

Hi YC, this looks good. I added a comment to ensure that I fully understood the interface implications. However, since this is a clearly a cleaner approach and passes all the tests, please forge head with merging.

[yichengt900]

Hi YC, thanks. It looks the issue was that my original changes to get geolat and the information required for the equation of state calculation were, for lack of a better descriptor, "overkill". Your code modifications make this change much less disruptive. Thank you.

If I understand correctly, you were able to get these efficiencies by using the MOM_EOS variable rather than passing a broader set of thermodynamic variables, and by only passing them to "generic_tracer_source" subroutine in MOM_generic_tracer.F90. Am I right in thinking that the interface changes will thus be limited to this latter change?

best, Charlie

Hi @charliestock, You've understood correctly. The focus of these changes is indeed on optimizing efficiency by utilizing the MOM_EOS variable and narrowing the modifications to the "generic_tracer_source" subroutine in MOM_generic_tracer.F90. The intention is to keep the interface changes confined to this specific aspect.

Our primary goal is to streamline the passing of necessary variables through "generic_tracer_source" to COBALT, minimizing broader code changes while achieving the required inputs for daylength and density calculations. I am open to further exploration of improved options for density variables in subsequent PR

[charliestock]

Thanks for the addition clarification YC, nice work!

[yichengt900]

@charliestock , thanks! This PR has a dependency: NOAA-CEFI-Regional-Ocean-Modeling/ocean_BGC#8. It would be great if someone could also review it and provide comments or approval, so I can merge them simultaneously.