Ufs/dev PR#246

[hertneky]

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

Associated ufs/dev PR:

• ufs-community/ccpp-physics#246

Associated fv3atm PR:

• NOAA-EMC/fv3atm#914

Associated NCAR PR:

• NCAR/ccpp-physics#1121

---

REGRESSION TEST CHANGES:
Expected changes from UWM tests:
cpld_control_p8_faster intel
control_p8_faster intel
hafs_regional_storm_following_1nest_atm_ocn_wav intel
hafs_regional_storm_following_1nest_atm_ocn_wav_inline intel

[scrasmussen]

Checked the changes and they look the same, good job!

Just out of curiosity, because these changes changed the UWM RTs, would we expect them to also change our RTs?

[grantfirl]

@scrasmussen This PR wasn't supposed to change any UFS RTs and only ended up doing so due to compiler optimization changes (this was the hypothesis anyway). The changed RTs from this PR are suspicious to me since there are so many. I'm running the RT CI workflow manually on the main branch to double-check that the baselines are OK.

[grantfirl]

@hertneky @scrasmussen Also, please hold off on merging this until I add a test tracer file. Lisa Bengtsson emailed me a snippet from a field_table that we can use. We can add an RT to test this if we want, even though one was not added for UFS. I'm on the fence about adding an additional RT, but I can at least test with using the new tracer locally before merging.

[grantfirl]

@mkavulich @scrasmussen @hertneky It looks like the baselines weren't updated since the last time they changed. I ran the RTs on the top of main and there look to be the same differences as in this PR: https://github.com/NCAR/ccpp-scm/actions/runs/14132388691/job/39595745350. Could someone at NCAR take care of uploading this to mohawk please?

[grantfirl]

@hertneky @scrasmussen Also, please hold off on merging this until I add a test tracer file. Lisa Bengtsson emailed me a snippet from a field_table that we can use. We can add an RT to test this if we want, even though one was not added for UFS. I'm on the fence about adding an additional RT, but I can at least test with using the new tracer locally before merging.

@hertneky OK, I've added omegab to the tracers for the GFS_v17_p8_ugwpv1 suite, added the progomega variable to the namelist for that suite (is default false) and added the new variable to the output. I've also run the TWPICE case with progomega turned on and I see nonzero values of omegab coming out, although I didn't look at the results anymore than that, so it seems to work.

[grantfirl]

This should be ready to merge as soon as someone updates the baselines on mohawk and the RTs are kicked off again to show the true RT differences.

[hertneky]

This should be ready to merge as soon as someone updates the baselines on mohawk and the RTs are kicked off again to show the true RT differences.

I updated about 5 mins ago and they are already running.

[grantfirl]

CI RTs only show differences to the GFS_v17_p8_ugwpv1 suite where progsigma is active, and the differences (miniscule) are consistent with the optimization change theory from the UFS RTs, so this should be good to go. CI is running again due to a tiny metadata fix in the physics: NCAR/ccpp-physics@0d4b4b9. We can wait until that finishes to merge.