Option to read horizontally varying KHTH from file

[NoraLoose]

• Adds the option to set the diffusivity KHTH as horizontally varying
• Can be enabled via READ_KHTH = True, filename is provided by user via KHTH_FILE
• Will return error if user sets both READ_KHTH = True and KHTH > 0

I tested the functionality of this new option in the CPT NeverWorld2 setup.

Left: Prescribed spatially (here: latitudinally) varying KHTH, read from netcdf file.
Middle: Diagnosed KHTH via diagnostic KHTH_t
Right: Difference.

Small differences are expected since

1. the diffusivities at velocity points KH_u, KH_v are computed as arithmetic means from the read-in KHTH, where the latter is assumed to live on tracer points;
2. the diagnostic KHTH_t is computed from KH_u, KH_v weighted by the thickness fluxes (rather than being computed as a simple arithmetic mean).

[codecov]

Codecov Report

Merging #236 (a6039ba) into dev/gfdl (aa44c32) will decrease coverage by 0.00%.
The diff coverage is 24.00%.

@@             Coverage Diff              @@
##           dev/gfdl     #236      +/-   ##
============================================
- Coverage     37.09%   37.09%   -0.01%     
============================================
  Files           263      263              
  Lines         73368    73388      +20     
  Branches      13672    13677       +5     
============================================
+ Hits          27219    27220       +1     
- Misses        41121    41136      +15     
- Partials       5028     5032       +4     

Impacted Files	Coverage Δ
...arameterizations/lateral/MOM_thickness_diffuse.F90	46.38% <24.00%> (-0.96%)	⬇️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[Hallberg-NOAA]

Thank you for this useful contribution. I think that it will be a valuable new capability for users.

However before this is accepted, there are several modifications (described separately below) that I would strongly encourage to follow the patterns used elsewhere in the MOM6 code when reading in variables from files. These should not be very onerous to make, but would help users by providing some of the constructs that they might expect.

In addition, this is really an isopycnal height diffusivity or an interface height diffusivity that is being read in, and not a thickness diffusivity (although these are the same in the limit of flat topography and vertically uniform diffusivities). I have regretting using the terminology "thickness diffusion", and perhaps we could consider modifying some of the new comments to read "isopycnal height diffusivity" (your choice) rather than "thickness diffusivity".

[NoraLoose]

Thanks for your comments, @Hallberg-NOAA! I have addressed the 3 points you made via 3 more commits.

[marshallward]

Gaea regression: https://gitlab.gfdl.noaa.gov/ogrp/MOM6/-/jobs/89375 ✔️ 🟡

New parameter:

• READ_KHTH