Towards WCDA: Introduce geos_marine, keep geos_ocean

[Dooruk]

This introduces geos_marine as a 3rd model interface and allows assimilating ocean and sea-ice data (first proper Weakly-Coupled DA with SWELL!). geos_marine will eventually replace geos_ocean but we can keep it around for now while we figure out ocean-only DA vs. marine DA changes. So that creates a lot of code duplication in terms of observation YAMLs under geos_ocean and geos_marine but that is temporary.

Some highlights:

• GEOSgcm version is bumped to v11.6.0, which required some changes in terms of Boundary Conditions linking.
• We switched from using CICE4 to CICE6. As far as SWELL is concerned, this is a one line change in run_geos_executable as we point to the external GEOS builds and experiment directories.
• Introduced the new task SOCA2CICE variable change, which should only be active in flow.cylc while the sea-ice DA is active.
• SOCA is designed to handle different model types (ocean, sea-ice, BGC, waves), hence some of the configuration creation will depend on if these models are turned on or off. This is a longer term consideration but an important one as combinations of different model types are concerned.
• Some minor EVA related changes, mainly concerning projections for sea-ice related plots.
• Introduces --no-requeue #SBATCH option to prevent Cylc tasks from resubmitting jobs and "spreading" to other compute nodes.

Closes the following:

• Modifying JEDI files to handle CICE files #265
• GEOS version update and iced.nc handling #264
• Add --no-requeue flag to SLURM submissions #388

[ashiklom]

I had a few minor comments on the Python implementation, but I'm not equipped (yet) to comment on the YAML configs that are the bulk of this PR or the science behind them. I'll have to defer to somebody with a better understanding of JEDI for that.

[Dooruk]

I had a few minor comments on the Python implementation, but I'm not equipped (yet) to comment on the YAML configs that are the bulk of this PR or the science behind them. I'll have to defer to somebody with a better understanding of JEDI for that.

Thanks for reviewing this PR. I'm the only one exercising the cyling GEOS+SOCA capability currently so I added geos_marine as a first step. Hopefully in the second step (another PR) when I take geos_ocean out David can test it to make sure ocean-only DA still works.

[asewnath]

Just some comments related to a bug found when building swell experiments with the CLI

[Dooruk]

I am going to merge this to move forward with the FMS updates.