Fix mistake in linearization of alpha and beta

alpha and beta of one column originally were only linearized. However,
because alpha of the searched from interface also changes with depth,
it should also be linearized.

Merge branch 'dev/master' of github.com:NOAA-GFDL/MOM6 into dev/master

Ulm release (21.0.0)

Ulm prerelease (21.0.0-beta)

Draft of SCM idealized hurricane wind forcing

- New initialization and forcing for a single column model
  with idealized hurricane wind forcing.

Draft of SCM idealized hurricane wind forcing

- New initialization and forcing for a single column model
  with idealized hurricane wind forcing.

Updated OMP directives for issue NOAA-GFDL#78

- Added varaibles to shared(), private() OMP directives
  so that model compiles with OPENMP=1 on command-line.
- I have not run with openmp, only compiled.
- No answer changes (with non-openmp executable).
- Closes NOAA-GFDL#78.

Last version before splitting repository

- About to move tools/ and examples/ to CommerceGov/NOAA-GFDL-MOM6-examples.

Answer changing:

- Turned on MLE parameterization in OM4_025.
- Re-write of MEKE.
- Fix underflow in conservation test in remapping_core().

Significant non-answer changing:
- Switch most analysis to ocean_annual_z output.
- Verified end-to-end preprocessing for OM4_025.
- CMORization of some diagnostics.

Missed a necessary fix for openmp (cf PR NOAA-GFDL#68)