Test mle diag

[kshedstrom]

@amoebaliz added an lf_bodner diagnostic. This cleans it up a little, allowing it to pass the auto-testing, etc.

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (a03bb95) 37.20% compared to head (118c9d3) 37.20%.

Files	Patch %	Lines
...ameterizations/lateral/MOM_mixed_layer_restrat.F90	87.50%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##           dev/gfdl     #542   +/-   ##
=========================================
  Coverage     37.20%   37.20%           
=========================================
  Files           271      271           
  Lines         80355    80363    +8     
  Branches      14985    14986    +1     
=========================================
+ Hits          29894    29901    +7     
  Misses        44903    44903           
- Partials       5558     5559    +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Hallberg-NOAA]

Thank you for this contribution of a new diagnostic, @amoebaliz, and thank you, @kshedstrom, for fixing the minor problems that were causing the original PR #532. Before we add this PR to dev/gfdl, though, I have a few suggestions that we might want to consider.

1. This commit adds what might be a relatively expensive diagnostic (or at least it would be were it not 2-d) that it calculates regardless of whether the diagnostic is being written out. It might be preferrable to add this diagnostic in a separate block with an if (id_lbod) test outside of it.

2. We have recently added a cuberoot() function (see PR Recurrent failures of the GitHub Actions performance test #583 and PR *Use Halley method iterations in cuberoot function #544), which would allow us to eliminate a lot of the un-scaling and rescaling with this diagnostic, analogously to what is done in a pending PR (PR +(*)Bodner param with cuberoot and non-Boussinesq #545) that refactors mixed_layer_restrat_Bodner() to work properly in non-Boussinesq mode. I think that it would be helpful if we could revise this new diagnostic to work with cuberoot() to avoid changing its answers at roundoff after people have started using it.

3. A number of comments have 'LD:' in them to identify their author, but we have found that the github blame command it much more effective at tracing the history of the various lines of code in MOM6 and giving credit to its contributors, so we are discouraging people from adding their initials to comments like this. (I know that there are some old comments that do still have comments with names or initials, but these are mostly from before our conversion from CVS to git, when the blame option was not yet available.)

[marshallward]

@Hallberg-NOAA I should have mentioned to you that I told @amoebaliz that I'd handle these revisions. I will try to polish it up today.

[marshallward]

Replaced by #568