Updates to module_SGSCloud_RadPre.F90 and .meta (from Tanya)

[joeolson42]

Changes:

1. revert addition of GF clouds to previous form (basically minimize use of them)
2. update calculation of Xu-Randal to the exact form used in GFS
3. (2) requires using more species (qr, qs, etc), which requires a change to the meta file. Note that adding these species to the cloud fraction field will make it unusable for ceiling calculations - must output and use CLDFRA_BL instead.

[climbfuji]

Associated PRs:
#42
NOAA-GSL/GFDL_atmos_cubed_sphere#5
NOAA-GSL/fv3atm#39
NOAA-GSL/ufs-weather-model#32

For regression testing information, see NOAA-GSL/ufs-weather-model#32.

[tanyasmirnova]

1. revert addition of GF clouds to previous form (basically minimize use of them)
   _This is not correct: the convective clouds are added as they were before (hash: 6671b15). But this commit removes use of Xu-Randall cloud fraction in the areas of convection (again as it was in hash: 6671b15). The Xu-Randall cloud fraction will be used with the GF scheme only when MYNN PBL or GFDL MP are not used in the physics suite (again as it was in hash: 6671b15).

2. update calculation of Xu-Randal to the exact form used in GFS
   Yes, as it was in progcld6 that was removed in hash: 2b14e69 as a clean-up.

3. requires using more species (qr, qs, etc), which requires a change to the meta file. Note that adding these species to the cloud fraction field will make it unusable for ceiling calculations - must output and use CLDFRA_BL instead.
   Again, this is done only because we had it in progcld6 that was removed.

Overall, the modifications to module_SGSCloud_RadPre.F90 are done to revert the changes in the cloud-radiation interaction introduced with the 2b14e68 commit that were causing significant warming above 700 mb in the convection areas. This new commit will be just a cleaned-up version of the 6671b15 commit in terms of cloud-radiation interactions (used in Hannah's May 2020 retro).