Use culled ISC as global ocean initial condition for Omega

[cbegeman]

The motivation is to have a global ocean mesh and initial condition available for testing before the global ocean init capability is fully ported to Polaris.

The capability to create MPAS-Ocean and MPAS-SeaIce restart files with culled ISC was introduced here: MPAS-Dev/compass#794. To my knowledge this has only been tested with an Icos30 mesh. The restart files had been spun-up in a G-case, which may not be desirable for Omega.

In reviewing the code, I don't see any changes we need to make. We would be doing the conversions in the state and vertical coordinate after culling. The landIceMask variable is used for culling.

In order to run in coupled configurations, domain mapping files need to be modified along the lines of this script from @jonbob (--method normalize removed based on Slack conversation https://e3sm-project.slack.com/archives/C06P9E3P07P/p1711485832127819 and --correct-areas added based on Slack conversation https://e3sm-project.slack.com/archives/C06P9E3P07P/p1711487743131309):

function makeoverlap {
    # --a should be grid with holes.
    cmd="GenerateOverlapMesh --a $agrid --b $bgrid --allow_no_overlap --out oa_overlap.g"
    echo $cmd
    eval $cmd
}

function makemap {
    echo "src $srcmesh dst $dstmesh map $map"
    cmd="GenerateOfflineMap \
      --in_mesh $srcmesh --out_mesh $dstmesh --ov_mesh oa_overlap.g \
      --in_type fv --in_np 1 --out_type fv --out_np 1 \
      --out_map $map --out_format Classic --correct-areas"
    echo $cmd
    eval $cmd
}

[cbegeman]

Questions:

1. Which RRSwISC6to18E3r5 files are the starting point for culling New Mesh: RRSwISC6to18E3r5 MPAS-Dev/compass#801? Probably the one before SSH adjustment
2. Do we want to derive Omega's vertical coordinate from refLayerThickness which should be (confirm) the z-star version corresponding to SSH=0 everywhere? Probably yes
3. Do we want to use temperature and salinity from an initial condition prior to the 1-month G-case? Yes

[xylar]

Which RRSwISC6to18E3r5 files are the starting point for culling MPAS-Dev/compass#801? Probably the one before SSH adjustment

I agree. This almost certainly requires me pulling this back over from NERSC HPSS, since I don't think we have those files available anywhere else. I did this recently for @vanroekel but then deleted them again.

I'll update here once that's happened.

[xylar]

Do we want to derive Omega's vertical coordinate from refLayerThickness which should be (confirm) the z-star version corresponding to SSH=0 everywhere? Probably yes

That was my thought. We can probably just use layerThickness from the initial condition to compute Omega's reference pseudo-thickness, but it could be that refLayerThickness is safer, once it's corrected for the initial ssh.

[xylar]

Do we want to use temperature and salinity from an initial condition prior to the 1-month G-case? Yes

We don't even want the ssh adjustment or the dynamical adjustment in standalone MPAS-Ocean, I think. Just the raw initial condition before all of that.

[cbegeman]

@xylar Did you have other considerations that I should look into?

[xylar]

@cbegeman, that seems like all I can think of. Thanks very much for getting to this so quickly!

[cbegeman]

@katsmith133 @xylar This is the thread where we discussed the strategy of using an existing wISC mesh. Thought you should know about it in case it seems like an easier strategy than starting with a non-ISC mesh. It may provide an easier route toward using recent ICs.

[xylar]

I don't know. I lean pretty hard on using meshes that just don't have ice-shelf cavities to begin with at this point. The non-z-level coordinate near ice shelves is likely to cause us headaches especially with the 2nd-order pressure gradient in the near term.

[cbegeman]

Good point

[vanroekel]

I agree with Xylar - let's try use wISC meshes.