UFS-dev PR#195

[grantfirl]

SOURCE: @grantfirl

DESCRIPTION OF CHANGES:

• All changes enable the GFDL MP v3 and TEMPO schemes
• New input data is required for TEMPO; this is staged in v7.0.0 release assets on GitHub with a new script to download/extract provided in /contrib
• New SDFs based on GFS_v16 for GFDL MP v3 and GFS_v17_p8_ugwpv1 for TEMPO (as in FV3); also new physics namelists to be used with these SDFs
• The flag_for_fast_microphysics_energy_conservation variable is stored in the CCPP_typedefs/GFDL_interstitial DDT in FV3; this doesn't exist in SCM, so this variable was moved to GFS_typedefs/GFS_control (this now differs from FV3)
• New entries in suite_info.py for the namelists/timesteps
• New RTs added

ISSUE: N/A

ASSOCIATED PRs:
NCAR/ccpp-physics#1153

TESTS CONDUCTED: UFS RTs were run in ufs-community/ccpp-physics#195; The new SDFs were exercised on my Mac for the arm_sgp_summer_1997_A and TWPICE cases.

This PR catches the NCAR:main branch up with changes from the ufs-community:ufs/dev branch.

Associated ufs/dev PR:

• GFDLMP-v3 and TEMPO MP microphysics ufs-community/ccpp-physics#195

Associated fv3atm PR:

• GFDLMP-v3 and TEMPO MP microphysics NOAA-EMC/ufsatm#823

Associated NCAR PR:

• UFS-dev PR#195 ccpp-physics#1153

---

REGRESSION TEST CHANGES: None expected from UFS RTs except new tests added. It does look like all tests using GFDL MP v1 are not bit-for-bit, but functionally equivalent. I'm not sure why SCM RTs show differences with the GFDL MP v1 scheme but UFS RTs do not.

[hertneky]

Looks good to me.
Built ursa-gnu/ursa-intel and ran rt tests successfully (test/rt_test_cases.py).

[grantfirl]

@hertneky @scrasmussen Could one of you please update the baselines with the artifacts found here: https://github.com/NCAR/ccpp-scm/actions/runs/17247393582