Fix atmospheric model's field import

[rmontuoro]

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 fixes the coupling issue described in #401, which occurs when running the atmospheric model's with separate run phases.

Issue(s) addressed

• fixes Replacing single NUOPC run phase with two sequential run phases leads to surface layer differences in P7 configurations NOAA-EMC/fv3atm#401

Testing

These changes were tested on the following platforms:

• hera.intel
• hera.gnu
• orion.intel
• cheyenne.intel
• cheyenne.gnu
• gaea.intel
• jet.intel
• wcoss_cray
• wcoss_dell_p3: not run due to Venus issues
• CI: https://github.com/ufs-community/ufs-weather-model/actions/runs/1304893686

Dependencies

• waiting on Properly import coupling fields when running with separate run phases NOAA-EMC/fv3atm#406

[climbfuji]

@BrianCurtis-NOAA @junwang-noaa I removed the Cheyenne labels so that they won't overwrite my logs once Brian fixed the problem.

[BrianCurtis-NOAA]

Note: Cheyenne missing a file in AutoRT, issue fixed.

[BrianCurtis-NOAA]

Automated RT Failure Notification
Machine: orion
Compiler: intel
Job: RT
Repo location: /work/noaa/nems/emc.nemspara/autort/pr/749037300/20211004140010/ufs-weather-model
Please manually delete: /work/noaa/stmp/bcurtis/stmp/bcurtis/FV3_RT/rt_367581
Test cpld_control_c192_p7 006 failed failed
Test cpld_control_c192_p7 006 failed in run_test failed
Please make changes and add the following label back:
orion-intel-RT

[MinsukJi-NOAA]

Automated RT Failure Notification Machine: orion Compiler: intel Job: RT Repo location: /work/noaa/nems/emc.nemspara/autort/pr/749037300/20211004140010/ufs-weather-model Please manually delete: /work/noaa/stmp/bcurtis/stmp/bcurtis/FV3_RT/rt_367581 Test cpld_control_c192_p7 006 failed failed Test cpld_control_c192_p7 006 failed in run_test failed Please make changes and add the following label back: orion-intel-RT

cpld_control_c192_p7 timed out. Will manually run cpld_control_c192_p7 and cpld_restart_c192_p7

[MinsukJi-NOAA]

@rmontuoro , can you please run regression tests on Cray and Dell?

[junwang-noaa]

Venus has a bad network switch. the RT does not get identical results from mars. We will skip dell RT for this PR.

[climbfuji]

fv3atm was merged, new hash is 95ff1ca

[climbfuji]

Will approve after .gitmodules and submodule pointer update.