Updates for FMS 2022.01 and a GFDL_amos_cubed_sphere branch

[shlyaeva]

Description

All work done by @danholdaway.

Updates that are needed for using FMS 2022.01 from stack and a branch of GFDL_atmos_cubed_sphere (see description of changes in NOAA-GFDL/GFDL_atmos_cubed_sphere#231).

Dependencies

Goes with:

• https://github.com/JCSDA-internal/fv3-bundle/pull/103
• https://github.com/JCSDA-internal/fv3-jedi-linearmodel/pull/17
• https://github.com/JCSDA-internal/fv3-jedi/pull/812

[danholdaway]

Amazing work whoever did this ;)

[shlyaeva]

I am going to merge this PR now: @svahl991 ran tests showing that using develop on all branches and fv3-jedi-data on this branch doesn't break the tests, and we also ran CI confirming that (https://github.com/JCSDA-internal/fv3-jedi/pull/815).
Merging this now would allow us to put this in before the data freeze date (today), and potentially have JEDI w/ FMS2022.04 functionality in v4 release.
The other corresponding PRs (fv3-bundle, fv3-jedi, fv3-jedi-linearmodel and TBD jedi-bundle and possibly mom6/soca) would have to be merged in a coordinated merge when everything is ready and approved. Code freeze is later than data freeze, so we still have time for that.

[fmahebert]

An early merge of this PR sounds good to me with tests that have been performed