Handle 4D variables in getsfcensmeanp.x utility.

[aerorahul]

getsfcensmeanp can now handle 4D variables.
Specifically, cldfra was introduced by the UFS-weather-model.
Any/all 4D variables will be averaged in the sfc files.

See commit f1dc784 that adds ability to process 4D variables.

[RussTreadon-NOAA]

Please add @emilyhcliu as a reviewer since this involves cldfra.

[RussTreadon-NOAA]

@aerorahul and @emilyhcliu , I propose a different fix. I added logic to getsfcensmeanp.f90 to handle structures with ndims=4. With this change getsfcensmeanp.f90 can processes a sfc*nc file containing cldfra. However, depending on @yangfanglin 's response to the comment in g-w PR #902, this gsi-utils PR may either be obsolete or not the best resolution of the problem at hand.

[aerorahul]

@RussTreadon-NOAA
Your fix is totally fine.
I would like to know why cldfra is in the sfc file, when it is neither a "surface" field nor a "2D" field.
This has implications not just on the sfc filesize, but also on bandwidth, compressibility, and any "convention" that was followed.
The surface file is not using parallel netcdf when doing lower resolutions. Although, this might not matter, it begs the question on adding a variable such as this in a file that is not meant to carry it.

[RussTreadon-NOAA]

@aerorahul , you ask excellent questions which bounce back to the modeling team (@yangfanglin , @junwang-noaa ). For the time being I will add my change to your branch, bugfix/getsfcensmeanp_cld.

[RussTreadon-NOAA]

@aerorahul , I am unable to push commits to github.

Orion-login-3[26] rtreadon$ git status
On branch bugfix/getsfcensmeanp_cldfra
Your branch is ahead of 'origin/bugfix/getsfcensmeanp_cldfra' by 1 commit.
  (use "git push" to publish your local commits)

nothing to commit, working tree clean
Orion-login-3[27] rtreadon$ git push origin bugfix/getsfcensmeanp_cldfra
remote: Permission to NOAA-EMC/GSI-Utils.git denied to RussTreadon-NOAA.
fatal: unable to access 'https://github.com/NOAA-EMC/GSI-Utils.git/': The requested URL returned error: 403
Orion-login-3[28] rtreadon$

[yangfanglin]

@aerorahul , you ask excellent questions which bounce back to the modeling team (@yangfanglin , @junwang-noaa ). For the time being I will add my change to your branch, bugfix/getsfcensmeanp_cld.

Please see comments I made in NOAA-EMC/global-workflow#902

[RussTreadon-NOAA]

Discuss various paths forward with @aerorahul . For the sake of timeliness, decide to move forward with this PR. A follow up PR can be opened at a later date based on @emilyhcliu's feedback.