+Corrected diagnostics and diagnostic units

[Hallberg-NOAA]

Corrected the units reported for 9 diagnostics, and altered the code so that
the diagnostics N2_u and N2_v are only offered if the can be calculated and the
proper diagnostics are written if these diagnostics are requested (previously
arrays of zeros were always output). All solutions are bitwise identical, but
some diagostics in output files change and the available_diags files have
altered entries.

[Hallberg-NOAA]

The failed Travis check is due to changes in the values of two diagnostics, which had previously been reporting all 0 values, and offering diagnostics in some cases when they are not in fact available.

Pipeline testing for this PR has been triggered (although I can not see the pipelines from outside of the GFDL building, so someone else will have to let me know the results).

[marshallward]

Gaea regression test: https://gitlab.gfdl.noaa.gov/ogrp/MOM6/pipelines/10010

The tests have passed up to a parameter diagnostic unit change, as expected.

I have some major Lustre issue on Gaea (it currently takes me several hours to build+run the test), so someone else may need to do the parameter update.