Update gx1 initial condition, Icepack, tests, and version. Fix hmix default value

[apcraig]

PR checklist

• Short (1 sentence) summary of your PR:
  Update gx1 initial condition, Icepack, tests, and version number. Fix hmix default value CESM ocean forcing - hmix division by zero #585
• Developer(s):
  apcraig, dabail10
• Suggest PR reviewers from list in the column to the right.
• Please copy the PR test results link or provide a summary of testing completed below.
  tests are bit-for-bit except gx1 with new initial condition. debug tests with gx1 fail due to Nans in ic file. https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#c0364f4f6924009f9d949d6c0484e832c1c5a9c6
• How much do the PR code changes differ from the unmodified code?
  • bit for bit
  • different at roundoff level
  • more substantial
• Does this PR create or have dependencies on Icepack or any other models?
  • Yes
  • No
• Does this PR add any new test cases?
  • Yes
  • No
• Is the documentation being updated? ("Documentation" includes information on the wiki or in the .rst files from doc/source/, which are used to create the online technical docs at https://readthedocs.org/projects/cice-consortium-cice/. A test build of the technical docs will be performed as part of the PR testing.)
  • Yes
  • No, does the documentation need to be updated at a later time?
    • Yes
    • No
• Please provide any additional information or relevant details below:

Also

• Remove old and unused jra55_data subroutine, should have been done before.
• Fix default hmix value , closes CESM ocean forcing - hmix division by zero #585
• Update landice tests, closes Add new landfast tests #573
• Update gx1prod settings to match up with recent gx1 jra55 multi-year run
• Update gx1 and gx3 initial filenames
• Update SOM forcing dataset name for gx1 runs
• Update CI wget implementation, closes Robustify GitHub Actions workflow #588
• Create variable in ice_forcing.F90 called mixed_layer_depth_default and use it instead of c20
• Fix bug in hourly output, created in time manager update , closes Hourly output broken by recent time manager update #589
• Set start year for all runs to 2005 and turn leap years on by default
• Update documentation in history output section to add information about hist_avg namelist, noted in Update Time Manager #566
• Update color links on test results wiki page, move from https to local images
• JRA forcing initialization/spinup #481, Replace forcing for gx1 #249

[eclare108213]

Just a note, mainly as a reminder: We are making JRA the default, but the older forcing options are still available. Those will need to be deprecated via our process (not in this release).

[eclare108213]

shouldn't year_init match the new forcing data? maybe it doesn't matter, as long as it's properly set in the -s options. I see below that fyear_init=2005, but I thought that one wasn't used now. It's a little confusing, would be better to be consistent.

[apcraig]

I saw this and wondered the same things. I was reluctant to make this change as I didn't understand the importance or history. The other thing is that it will change all the gx3 results as well in our regression test. But I agree 100% that it would be nice if all of this were consistent. Are there other defaults that need to change. Should use_leap_years be always true, for instance? I can make this change as part of this PR, but it will change answers.

[eclare108213]

Is this just temporary because of NaNs in the file? If not, then it would be better to set this as a parameter at the top of the module, to make it more explicit. 'default mixed layer depth'

[apcraig]

I think this avoids a divide by zero. I assume there are also some science reasons for setting it to 20. I can create a parameter in the code as suggested and will do that later today. I think that's a good idea.

[TillRasmussen]

This is a way of avoiding divison by zero. @dabail10 suggested max(c20,work1(i,j,iblk)). I dont have a preference. C20 is set to match the only other fixed setting in subroutine init_coupler_flux.

[apcraig]

This is ready to merge. I ran a full test suite on cheyenne this morning, https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#4dc3bf15908deffbd68dd400fc0513a6bae7b81f and everything looks good.

It would be nice to merge this afternoon to leverage automated weekend testing and to carry out release testing this weekend as well. Someone should review and approve. I'm happy to make additional changes as needed too.

[eclare108213]

Sorry I didn't notice this was waiting for a review earlier. A few questions/comments to consider, but it looks good.

[eclare108213]

We need to fix the ocn_data_oned subroutine when we clean up the forcing module. sss probably shouldn't be hardwired either. We can discuss whether to get rid of this option completely (and just use Icepack) or keep it as an additional test for comparing with Icepack's test, or keep it to have netcdf available in a column configuration.

[eclare108213]

does this test work, starting Sep 2?

[apcraig]

Yes it does. It works just fine. In fact, sec_init = 7200 as well in this setup in case you didn't notice and it all works fine. It runs 5 days to 2005-09-07-07200. The output is also as expected in the sense that things like daily history averages are fine except on the first and last day. If one started on 2005-09-07-07200, you'd probably want to run 22 hours first (which you can do with npt_unit='h' and npt=22) to get on a daily boundary, and then start running days or months after that.

[eclare108213]

cool, thanks

[dabail10]

This looks good to me. I do wonder about setting a mixed-layer depth default. Particularly as I did a fill, so there should be valid values everywhere.