Enable submeso. parameterization in global_ocean

[xylar]

This merge also adds submeso output variables to daily_output test.

Checklist

• Document (in a comment titled Testing in this PR) any testing that was used to verify the changes

closes #532

[xylar]

I tried to run QU240 tests with this branch but they are all failing with state validation errors, perhaps related to the masking PR E3SM-Project/E3SM#5325 that we're trying to debug.

OSX CI failures are unrelated to this.

[cbegeman]

@xylar I had in mind that we would add an additional test with submesoscale turned on. It could be just daily_output_submeso_test or up a level at PHC_submeso

[xylar]

@cbegeman, we could do that, but the default in E3SM is to run with the parameterization on:
https://github.com/E3SM-Project/E3SM/blob/982d11520aec729ef32ee6991bd85162f26b917b/components/mpas-ocean/bld/namelist_files/namelist_defaults_mpaso.xml#L156
So it feels to me like it should be on in global_ocean, too. This is consistent with GM being on by default in global_ocean.

[cbegeman]

@xylar I didn't realize that. In that case, I'm fine with it being on by default.

[cbegeman]

@xylar I've tested the nightly suite plus ocean_global_ocean_QU240_PHC_daily_output_test with the bugfix branch E3SM-Project/E3SM#5458 and found that all tests pass (chrys, intel, openmpi).

[xylar]

Testing

I tested this on Chrysalis with Intel and OpenMPI. I ran all QU240 test cases using the branch for E3SM-Project/E3SM#5458. All tests passed with this branch, whereas 8 forward mode tests failed with the version of MPAS-Ocean in the E3SM-Project submodule.

We will need to update the submodule before this goes in.

[xylar]

The submodule was update to include E3SM-Project/E3SM#5458 in #558 so this can now be merged.

[xylar]

Re-testing

I ran the pr test suite, comparing with the current main as a baseline and (to my surprise!) all tests passed. I verified that submesoscale is turned off for the baseline and on for this run (according to the namelist options), so it seems the tests are short enough not to detect the difference.

I ran all QU240 test cases (this time without a baseline) and they finished successfully.