PR Checklist

• Ths PR is up-to-date with the top of all sub-component repositories except for those sub-components which are the subject of this PR. Please consult the ufs-weather-model wiki if you are unsure how to do this.

• This PR has been tested using a branch which is up-to-date with the top of all sub-component repositories except for those sub-components which are the subject of this PR

• An Issue describing the work contained in this PR has been created either in the subcomponent(s) or in the ufs-weather-model. The Issue should be created in the repository that is most relevant to the changes in contained in the PR. The Issue and the dependent sub-component PR
  are specified below.

• If new or updated input data is required by this PR, it is clearly stated in the text of the PR.

Description

This PR only updates the submodule pointers for fv3atm and ccpp-physics for the changes described in the dependent PRs below (convert real(kind_phys) vegetation, slope and soil type arrays into integer arrays without affecting input/output files).

Also included:

• Update submodule pointer for CMakeModules (updates to FindNetCDF.cmake and FindESMF.cmake)
• Gaea only: switch to hpc-stack maintained by EMC (@kgerheiser)
• Update to utest (now opnReqTest) CI testing from @MinsukJi-NOAA

No changes to the input files.

Update 2021/09/30 The regression test fv3_gsd now has different results with Intel on Hera. It starts at 24h forecast time with a minimal difference in one of the tiles, and then propagates to all tiles. The regression test fv3_gsd_debug (running out to 6h only) is b4b identical. With GNU, both the 48h forecast fv3_gsd and the 6h forecast fv3_gsd_debug are b4b identical. I then ran the DEBUG version out to 48h, and the results between the original code andd this PR were identical. Thus, the b4b difference for fv3_gsd with Intel in PROD mode is due to an optimization round-off difference in RUC LSM.

Update 2021/10/01 Switching to hpc-stack maintained by EMC (@kgerheiser) on gaea changes the "results" for the following regression tests - on gaea only, of course, and only in the Grib2 files produced by the regional inline post:

Dom.Heinzeller@gaea12:/lustre/f2/scratch/Dom.Heinzeller/ufs-weather-model/ufs-weather-model-remove-noah-wrfv4-bugfix-mp-thompson/tests [intel|esrl_bmcs]> cat log_gaea.intel/rt_038_regional_quilt_2threads.log

baseline dir = /lustre/f2/pdata/ncep_shared/emc.nemspara/RT/NEMSfv3gfs/develop-20210928/INTEL/fv3_regional_quilt
working dir  = /lustre/f2/scratch/Dom.Heinzeller/FV3_RT/rt_38214/regional_quilt_2threads
Checking test 038 regional_quilt_2threads results ....
 Comparing dynf000.nc .........OK
 Comparing dynf024.nc .........OK
 Comparing phyf000.nc .........OK
 Comparing phyf024.nc .........OK
 Comparing PRSLEV.GrbF00 .........NOT OK
 Comparing PRSLEV.GrbF24 .........NOT OK
 Comparing NATLEV.GrbF00 .........NOT OK
 Comparing NATLEV.GrbF24 .........NOT OK

 0: The total amount of wall time                        = 267.801442

Test 038 regional_quilt_2threads FAIL

Dom.Heinzeller@gaea12:/lustre/f2/scratch/Dom.Heinzeller/ufs-weather-model/ufs-weather-model-remove-noah-wrfv4-bugfix-mp-thompson/tests [intel|esrl_bmcs]> cat log_gaea.intel/rt_041_regional_quilt_RRTMGP.log

baseline dir = /lustre/f2/pdata/ncep_shared/emc.nemspara/RT/NEMSfv3gfs/develop-20210928/INTEL/fv3_regional_quilt_RRTMGP
working dir  = /lustre/f2/scratch/Dom.Heinzeller/FV3_RT/rt_38214/regional_quilt_RRTMGP
Checking test 041 regional_quilt_RRTMGP results ....
 Comparing dynf000.nc .........OK
 Comparing dynf024.nc .........OK
 Comparing phyf000.nc .........OK
 Comparing phyf024.nc .........OK
 Comparing PRSLEV.GrbF00 .........NOT OK
 Comparing PRSLEV.GrbF24 .........NOT OK
 Comparing NATLEV.GrbF00 .........NOT OK
 Comparing NATLEV.GrbF24 .........NOT OK

 0: The total amount of wall time                        = 469.675394

Test 041 regional_quilt_RRTMGP FAIL

Dom.Heinzeller@gaea12:/lustre/f2/scratch/Dom.Heinzeller/ufs-weather-model/ufs-weather-model-remove-noah-wrfv4-bugfix-mp-thompson/tests [intel|esrl_bmcs]> cat log_gaea.intel/rt_037_regional_quilt.log

baseline dir = /lustre/f2/pdata/ncep_shared/emc.nemspara/RT/NEMSfv3gfs/develop-20210928/INTEL/fv3_regional_quilt
working dir  = /lustre/f2/scratch/Dom.Heinzeller/FV3_RT/rt_38214/regional_quilt
Checking test 037 regional_quilt results ....
 Comparing dynf000.nc .........OK
 Comparing dynf024.nc .........OK
 Comparing phyf000.nc .........OK
 Comparing phyf024.nc .........OK
 Comparing PRSLEV.GrbF00 .........NOT OK
 Comparing PRSLEV.GrbF24 .........NOT OK
 Comparing NATLEV.GrbF00 .........NOT OK
 Comparing NATLEV.GrbF24 .........NOT OK

 0: The total amount of wall time                        = 858.850208

Test 037 regional_quilt FAIL

@WenMeng-NOAA helped me to look at the differences, and they are only in the grib2 header section. The data itself is 100% identical. Please look at the discussion in NOAA-EMC/hpc-stack#337 (towards the bottom) for more details.

Issue(s) addressed

• fixes Remove real land/sea/ice mask, use integer variable instead NCAR/ccpp-physics#704
• fixes Transfer Gaea responsibilities from Dom to EMC NOAA-EMC/hpc-stack#273

Testing

Initial regression testing

Full regression tests passed on hera.intel against existing baseline.

Final regression testing

Regression tests will be run on all tier-1 platforms:

• hera.intel
• hera.gnu
• orion.intel
• cheyenne.intel
• cheyenne.gnu
• gaea.intel
• jet.intel
• wcoss_cray
• wcoss_dell_p3
• CI: https://github.com/ufs-community/ufs-weather-model/actions/runs/1303798496

Dependencies

• waiting on Convert real(kind_phys) vegetation, slope and soil type arrays into integer arrays NCAR/ccpp-physics#730
• waiting on Convert real(kind_phys) vegetation, slope and soil type arrays into integer arrays without affecting input/output files NOAA-EMC/fv3atm#388