UFS-dev PR#183

[grantfirl]

Changes similar to NOAA-EMC/fv3atm#798

Additional changes:

• Make the interstitial DDT in the SCM physics DDT an array of (n_threads) size so that GFS_Debug can work (this is how UFS does it)
  -- downstream changes for the above change

RT notes: This PR doesn't change RT results except for the addition of a new test that exercise the GFS_debug.F90 schemes. Note that the GFS_abort scheme in GFS_debug.F90 is untested because it causes a run failure (on purpose) and CI flags it as a failure, which we don't necessarily want.

[grantfirl]

@dustinswales @mkavulich @scrasmussen RTs run as expected now. There were a couple of failures related to the original PR -- one was that I failed to compile the new suite(s) in the RT CI workflow and the second was that, for testing purposes, I include the GFS_abort scheme from GFS_debug.F90 in the new suites, but this sets a non-zero error flag upon run termination and the CI flags it as a failure. All of that is expected, but we don't want to have CI fail from henceforth, so I just removed the GFS_abort scheme. All is good at this point, so I'll go ahead and merge after the ccpp-physics one.