adding mynn surface layer as opt_sfc=4 in NoahMP

[HelinWei-NOAA]

This PR is to add MYNN surface scheme as opt_sfc=4 in NoahMP and recode the thermal roughness scheme as a subroutine. A bug with opt_sfc=1 was fixed. 2m T/Q from NoahMP instead of sfc_diag will be used as outputs.

The update is not aimed for UFS coupled prototype 8, but for RRFS.

issue #920

[grantfirl]

@HelinWei-NOAA What kind of testing has been done? I see on https://github.com/ufs-community/ufs-weather-model/pull/1224 that it was tested on wcoss_dell_p3. Do any of the existing regression tests need to change their baseline as a result of the code changes here? Is the MYNN surface layer option in NoahMP exercised in any of the existing regression tests?

[HelinWei-NOAA]

@grantfirl We switch opt_sfc to 4 for some regional tests. Also there are other changes in this PR that will change the results for all tests using Noah MP.

[grantfirl]

@HelinWei-NOAA There was a merge into ccpp-physics this morning. This PR branch needs to be merge in the latest main branch again. Could you do this please? Also, once this is done, please update the upstream (fv3atm, ufs-weather-model) branches to point to the latest code too.

[HelinWei-NOAA]

@HelinWei-NOAA There was a merge into ccpp-physics this morning. This PR branch needs to be merge in the latest main branch again. Could you do this please? Also, once this is done, please update the upstream (fv3atm, ufs-weather-model) branches to point to the latest code too.

Done.

[dustinswales]

@HelinWei-NOAA It appears that the rte-rrtmgp submodule is pointing to the wrong hash, d9594c46c877a2ab8001f5cd37961efdcf08ad8e. Please switch to the submodule back to the HEAD (f9377e81d33e4f73f4433501186465b84dd1111c) of the dtc/ccpp branch, https://github.com/earth-system-radiation/rte-rrtmgp/commits/dtc/ccpp

[dustinswales]

@HelinWei-NOAA Thanks for reverting the rte-rrtmgp submodule. Everything looks good from a RRTMGP perspective.

[HelinWei-NOAA]

@SamuelTrahanNOAA Can you please merge my pull request?