modify subset_data for CRUJRA with a more user-interactive solution

[ekluzek]

subset_data is hardwired to work with datm_gswp3.
The default datm dataset is changing for clm6 to CRUJRA2024.

I added this comment to subset_data.py in #2956:

    # TODO issue #1895: allow datm_crujra, which also affects
    #      tools/site_and_regional/default_data_1850.cfg
    #      tools/site_and_regional/default_data_2000.cfg
    #      python/ctsm/test/testinputs/default_data.cfg

and posted this comment.

#2956 continues to hardwire but for CRUJRA. In the future we want to make this a user-interactive option.

[wwieder]

I'll note that I have an equally hard and brittle application of subset data that works for CRUJRA (currently unavailable on Derecho).

• In the short term, we can bring this option in (having something that's functional with CRUJRA seems critical).
• Alternatively, maybe it's not too difficult to make less brittle python code that can handle both options? (not sure how much we can support multiple datm products in our subsetting tools)?

[samsrabin]

Just met with Katie Rocci, who mentioned this being an issue in her work.

[slevis-lmwg]

I contacted Katie to ask for more info, especially if she is stuck in some way.

[samsrabin]

Oh sorry, no, she worked around it. But it was a pain and she couldn't remember how exactly.

[wwieder]

Sorry, would my hack for doing this with CRUJRA still be helpful here?

[samsrabin]

Depends on how quickly we can get this feature in. If not soon, then we should document your hack.

[wwieder]

Looks like I have it in this WIP PR that's including two separate issues: the CRUJRA workflow to subset data with MIMICS changes. Happy to break these out into two different PRs, but I may need some hand holding to make this more effective with git.

[slevis-lmwg]

@wwieder pretty much the same as your hack is coming in with #2956, so no need to do anything additional at this time. This issue is here to remind us to resolve this with a user-interactive solution rather than a hardwiring.

[wwieder]

OK, Thanks @slevis-lmwg. If we have something that works with #2956, maybe we don't have this be a requirement for point and regional cases in CLM6. It seems like the configurations we want to support should really just include the default datm (and resolutions) we're using in I cases? Should we consider changing the milestone associated with this more interactive option that would be nice to have down the road?

[slevis-lmwg]

[...] Should we consider changing the milestone associated with this more interactive option that would be nice to have down the road?

Sounds reasonable to me.