Add tracker

[SamuelTrahanNOAA]

Synopsis

This is the second of several PRs to merge improvements from the MPAS HFIP 2025 parallel.

It adds jobs to interpolate to a regular lat-lon grid, a job to run the tracker on that grid, and a job to archive the tracker. (The GFDL Vortex Tracker requires a regular lat-lon grid.)

This is a draft because

• There's no documentation.
• It must be beautified by making it more uwtools-ish.

Type

• Documentation
• Enhancement (adds new functionality)

Impact

• This is a non-breaking change (existing functionality continues to work as expected)

Issues

• Fixes Add GFDL Vortex Tracker #100

Dependencies

• Add archiving and scrubbing #96

Checklist

• I have added myself and any co-authors to the PR's Assignees list. I cannot edit the Assignees list
• I have reviewed the documentation and have made any updates necessitated by this change. Documentation changes are needed, but this branch does not yet have them.

[christinaholtNOAA]

I'm concerned about the tracker pointing to a commit on a feature branch in a repo that we do not control. We cannot ensure the hash will be available indefinitely as opposed to say a hash on a main or develop branch. Perhaps we should keep a fork for GSL and point to the hash there instead.

[christinaholtNOAA]

During the 2024 HFIP experiment, the tracker was added using a driver supported in the MPAS App (here). I don't mind having the grib2 pieces in the bash, but I do think that configuring and running the tracker should use the UW Driver framework that was already established for this task.

[christinaholtNOAA]

We generally do not use environment variables to define the interface for run scripts in the app. Command line arguments are used for the run-time specific information (cycle and lead time, for example), while the experiment invariant values are pulled explicitly from the experiment YAML. This is preferred for testing and documentation purposes.

[SamuelTrahanNOAA]

The environment variables have been removed.

I haven't switched to the uw tools tracker driver yet. It's in the PR's "to do" list.

[SamuelTrahanNOAA]

Status update.

This PR is up to date with main.

It doesn't use the uwtools tracker yet since that script is missing two features. From my reading of the scripts, the uwtools tracker lacks two features:

1. Filter storms from tcvitals by basin.
2. Skip running the tracker and produce an empty track file if there are no storms. (This is because the tracker failed with cryptic messages, which made debugging impossible. It's easier to explicitly log that there's nothing to track and make an empty track.)