Update develop from gsd/develop 2020/06/30 #134

climbfuji · 2020-06-30T14:25:06Z

Changes in this PR:

output of auxiliary arrays (2d, 3d) from CCPP to assist developers debugging their physics
correctly output 3-dim tendencies for t, u, v, qv, ozone from CCPP physics
mandatory changes for the CCPP framework updates to handle blocked data structures correctly: fix inconsistent use of horizontal_dimension (ix) versus horizontal_loop_extent (im) and other metadata corrections
remove old pset logic from ccpp_prebuild_config.py (was required for CCPP dynamic build)
updates to all suite definition files: change the order of interstitial schemes get_phi_fv3+GFS_suite_interstitial_3 and GFS_DCNV_generic_pre to match the logical flow (does not change the answer)
move CCPP physics initialization calls in atmos_model.F90 so that the physics can see the data from the restart files at initialization time
add capability to read large-scale and small-scale orographic data when using GSL drag suite
add additional fields to restart files as required for GSL physics updates
new suite definition files for RAP and HRRR configurations
capability to call radiation for the first nhfrad physics time steps for cold starts, afterwards on the cycle defined by the radiation time step

This PR needs to be squashed into one commit when merged, using the description of the PR as commit message.

Associated PRs:
NCAR/ccpp-physics#465
NCAR/ccpp-framework#307
NOAA-EMC/GFDL_atmos_cubed_sphere#30
#134
ufs-community/ufs-weather-model#156

For regression testing information, see ufs-community/ufs-weather-model#156.

…ute GFS_GWD_generic_pre,post for drag_suite_pre,post

…adiation_driver.F90

…odule pointers

2. in GFS_typedefs.F90, correct nvdiff definition for FA.

GWD interstitial reorganization

…S_fer_hires

…e from existing suites

…4b identical results on macOS with 32-bit dynamics

…schemes and satmedmfvdifq scheme

…for CCPP for sas and satmedmfvdifq

…gs about non-existent include directories with GNU

…eepcnv_samf instead of Model%imfdeepcnv_sas

…s directory to avoid warnings about non-existent include directories with GNU

Add SAS deep/shallow convection and satmedmfvdifq (updated version of satmedmfvdif) to CCPP

…S_fer_hires

…routines in ugwp_driver_v0.f (note that the code that generated the warnings is in a non-active section)

add Ferrier-Aligo MP scheme changes on host model side

dtc/develop: update submodule pointers 2019/11/22

… and ccpp-physics for code review and testing

…ore_bugfixes_and_sgscloud_update_gsd_develop_20200629

…tmos_cubed_sphere and ccpp-physics

…cloud_update_gsd_develop_20200629 Two FV3 dycore bug fixes and SGS cloud update gsd/develop 2020/06/29

…, ccpp-framework and ccpp-physics for code review and testing

…into tendencies

…tmos_cubed_sphere, ccpp-framework, ccpp-physics

junwang-noaa · 2020-07-02T01:29:12Z

gfsphysics/GFS_layer/GFS_diagnostics.F90

+      ExtDiag(idx)%desc = 'v momentum tendency due to non-physics processes'
+      ExtDiag(idx)%unit = 'm s-2'
+      ExtDiag(idx)%mod_name = 'gfs_phys'


for tendency from non-physics processes, maybe it's good not to put it in "gfs_phys" model in the future

Good point. The only reason why it ended up there was "convenience" so that all tendencies were in the same file. We can easily change this.

junwang-noaa · 2020-07-02T01:31:46Z

atmos_model.F90

@@ -1456,7 +1491,7 @@ subroutine update_atmos_chemistry(state, rc)
          ib = i + Atm_block%isc - 1
          nb = Atm_block%blkno(ib,jb)
          ix = Atm_block%ixp(ib,jb)
-          hpbl(i,j)   = IPD_Data(nb)%IntDiag%hpbl(ix)
+          hpbl(i,j)   = IPD_Data(nb)%Tbd%hpbl(ix)


May I ask why hpbl is changed from diag_type to Tbd type?

This was required, because it is not a diagnostic quantity but a prognostic variable for MYNN. In other words, to prevent a reset when the IntDiag DDT is reset and make it persistent.

junwang-noaa · 2020-07-02T01:41:19Z

gfsphysics/GFS_layer/GFS_typedefs.F90

@@ -5590,6 +5783,7 @@ end subroutine diag_create
  subroutine diag_rad_zero(Diag, Model)
    class(GFS_diag_type)               :: Diag
    type(GFS_control_type), intent(in) :: Model
+    integer :: i


not seeing i is used. may be remove in the future?

junwang-noaa · 2020-07-02T01:41:48Z

gfsphysics/GFS_layer/GFS_typedefs.F90

@@ -5612,6 +5806,7 @@ subroutine diag_phys_zero (Diag, Model, linit, iauwindow_center)
    logical,optional, intent(in)       :: linit, iauwindow_center

    logical set_totprcp
+    integer :: i


junwang-noaa · 2020-07-02T01:48:38Z

gfsphysics/GFS_layer/GFS_typedefs.meta

@@ -3991,6 +4074,12 @@
  units = flag
  dimensions = ()
  type = logical
+[cycling]


I am curious about this variable, model can cold start in cycling mode (e.g. GEFS). Is this used to identify cold start or warm start?

This was added by @joeolson42 when he was working on the MYNN PBL/SFC improvements. It has to do with cycling runs being conceptually different from cold starts or warm starts. Maybe we don't understand how the cycling capability works in FV3 and it could have been implemented differently. I am starting an issue to collect improvements to this PR from code review.

junwang-noaa

I reviewed the code for outputting diagnostic fields, it looks good to me. There are some very minor suggestions that have no impact on results, which can be fixed later if verified. I assume the new physics updates have been tested and evaluated separately.

climbfuji · 2020-07-02T02:01:12Z

Thank you, @junwang-noaa . I created an issue to address the questions and make the recommended changes as soon as possible.

climbfuji · 2020-07-02T02:02:18Z

I reviewed the code for outputting diagnostic fields, it looks good to me. There are some very minor suggestions that have no impact on results, which can be fixed later if verified. I assume the new physics updates have been tested and evaluated separately.

I am hopeful that we can soon share results for larger retrospective runs with the updated GSL physics. Several people at EMC have asked for this already.

- output of auxiliary arrays (2d, 3d) from CCPP to assist developers debugging their physics - correctly output 3-dim tendencies for t, u, v, qv, ozone from CCPP physics - mandatory changes for the CCPP framework updates to handle blocked data structures correctly: fix inconsistent use of `horizontal_dimension` (ix) versus `horizontal_loop_extent` (im) and other metadata corrections - remove old pset logic from `ccpp_prebuild_config.py` (was required for CCPP dynamic build) - updates to all suite definition files: change the order of interstitial schemes get_phi_fv3+GFS_suite_interstitial_3 and GFS_DCNV_generic_pre to match the logical flow (does not change the answer) - move CCPP physics initialization calls in `atmos_model.F90` so that the physics can see the data from the restart files at initialization time - add capability to read large-scale and small-scale orographic data when using GSL drag suite - add additional fields to restart files as required for GSL physics updates - new suite definition files for RAP and HRRR configurations - capability to call radiation for the first `nhfrad` physics time steps for cold starts, afterwards on the cycle defined by the radiation time step

* Add NSSL 2-Moment Microphysics * Update data structures to handle the additional tracer species in NSSL microphysics * Add the code files for NSSL microphysics to the CCPP prebuild config * Add suite for NSSL microphysics * Add a CCPP suite to test NSSL microphysics * Initialize ice fluxes and add "tiice" array (#119) (1) Introducing a separate array "tiice" to store internal ice temperature. "tiice" will be added to phyf*nc output only when frac_grid=T. (2) When cplflx=T, initializing ice fluxes by PBL calculated values when fluxes from CICE are unavailable. (3) Adding frac_grid to namelist and no longer modifying it inside FV3GFS_io.F90. frac_grid=F by default. * Remove cheyenne.pgi from CCPP config (#117) remove cheyenne.pgi from CCPP config (ccpp/set_compilers.sh, ccpp/build_ccpp.sh) * develop: merge HWRF version of saSAS with GFS version (#93) * Merge branch 'man_hafs_sas_without_updates_of_dtc_develop_49b73a8f4149ca88f071dbaae81f1769b120f1e0' into HEAD * gfsphysics/physics/GFS_debug.F90: add capability to debug 1-d logical arrays * Clean up error messages in gfsphysics/GFS_layer/GFS_typedefs.F90 Co-authored-by: Dusan Jovic <dusan.jovic@noaa.gov> * Merge GFDL FV3 dynamic core new (201912) version (#65) * Change 'mytile' to 'mygrid' corresponding to the change in atmos_cubed_sphere * GFS.v16.0.4 physics update (#128) * physics update from GFS.v16.0.4 * GFS.v16.0.4 physics udpate * remove whitespaces in sfc_diff.f * keep z0 unchanged in coupled mode in sfc_diff.f * fix IPD run * point to NCAR ccpp-physics branch Co-authored-by: Jun Wang <junwang-noaa@users.noreply.github.com> * Update for GFDL_atmos_cubed_sphere (dycore GNU bugfix) (#131) * Bugfix for regional runs when dycore is compiled in double precision (#133) * Update .gitmodules and submodule pointer for fv3atm for code review and testing * Revert change to .gitmodules and update submodule pointer for GFDL_atmos_cubed_sphere * Update develop from gsd/develop 2020/06/30 (#134) - output of auxiliary arrays (2d, 3d) from CCPP to assist developers debugging their physics - correctly output 3-dim tendencies for t, u, v, qv, ozone from CCPP physics - mandatory changes for the CCPP framework updates to handle blocked data structures correctly: fix inconsistent use of `horizontal_dimension` (ix) versus `horizontal_loop_extent` (im) and other metadata corrections - remove old pset logic from `ccpp_prebuild_config.py` (was required for CCPP dynamic build) - updates to all suite definition files: change the order of interstitial schemes get_phi_fv3+GFS_suite_interstitial_3 and GFS_DCNV_generic_pre to match the logical flow (does not change the answer) - move CCPP physics initialization calls in `atmos_model.F90` so that the physics can see the data from the restart files at initialization time - add capability to read large-scale and small-scale orographic data when using GSL drag suite - add additional fields to restart files as required for GSL physics updates - new suite definition files for RAP and HRRR configurations - capability to call radiation for the first `nhfrad` physics time steps for cold starts, afterwards on the cycle defined by the radiation time step * Add regression tests with MULTI_GASES (#108) * Update suite_FV3_GFS_2017_fv3wam.xml * Remove unnecessary file from the CCPP prebuild configuration * Add TACC frontera configuration * Update to RAS in CCPP, update submodule pointer for ccpp-framework (CI tests) (NOAA-EMC#252) * Reactivate mass flux arrays * fix the argument list in inline_post_getattr and remove prints and post_gfs_stub.F90 * ccpp/data/GFS_typedefs.F90: correct initialization of Model%fscav * Add FV3_GFS_v16_ras * Cld perts (NOAA-EMC#233) Addition of stochastic physics restarts in netCDF format that will be written at the restart interval, and at end of run. Also, addition of stochastic cloud perturbation and microphyiscs perturbations for SPPT. * upgrade OPAC aerosol with more advanced MERRA2 aerosol (NOAA-EMC#254) * changed .gitmodules to point to merra2 ccpp/physics * revert .gitmodule * update to Larissa version * add atmos_cube_sphere and ccpp/physics as submodules * add submodules of emc-2021-cloud * update ccpp/physics: remove CAPS version mp_nssl_2mom * ccpp/physics update: fix nssl mp reproducibility issue * Updating pointer to ccpp/physics. * Pointing to Christin's branch * Adding new suite for NSSL physics. * Updating suites to match HWT. * Updating gitmodules * Updating pointer to atmos_cubed_sphere. Co-authored-by: Tim Supinie <tsupinie@ou.edu> Co-authored-by: shansun6 <48043606+shansun6@users.noreply.github.com> Co-authored-by: Dom Heinzeller <dom.heinzeller@icloud.com> Co-authored-by: Dusan Jovic <dusan.jovic@noaa.gov> Co-authored-by: XiaqiongZhou-NOAA <48254930+XiaqiongZhou-NOAA@users.noreply.github.com> Co-authored-by: Jun Wang <37633869+junwang-noaa@users.noreply.github.com> Co-authored-by: Jun Wang <junwang-noaa@users.noreply.github.com> Co-authored-by: Phil Pegion <38869668+pjpegion@users.noreply.github.com> Co-authored-by: AnningCheng-NOAA <48297505+AnningCheng-NOAA@users.noreply.github.com> Co-authored-by: Jili Dong <jili.dong@noaa.gov>

…mposites.F90 (NOAA-EMC#134) * Updated submodules for Bug fixes to drag_suite.F90 and GFS_surface_composites.F90 * Point to RRFS_dev for ccpp/physics

@climbfuji

…gi, check for GNU compiler version (NOAA-EMC#139) * update stampede config, remove cheyenne.pgi config, bugfixes for Python3 on stampede (from @climbfuji) * check version of GNU compiler NOAA-EMC#134 (from @edwardhartnett) * submit compile jobs with ecflow to batch system instead of running them on frontend nodes (from @DusanJovic-NOAA) Co-authored-by: Dusan Jovic <dusan.jovic@noaa.gov> Co-authored-by: Edward Hartnett <Edward.Hartnett@noaa.gov>

grantfirl and others added 30 commits November 4, 2019 10:30

add GFS_GWD_generic_post scheme to SDFs using default GWD and substit…

e45b4f4

…ute GFS_GWD_generic_pre,post for drag_suite_pre,post

Add CCPP annotations to GFS_driver.F90, GFS_physics_driver.F90, GFS_r…

759997d

…adiation_driver.F90

Update .gitmodules for NCAR fork, branch dtc/develop, and update subm…

8bd0c2e

…odule pointers

add FA scheme changes on host model side

e6e2c81

1. using explicit simulation configuration in SDF file

685722d

2. in GFS_typedefs.F90, correct nvdiff definition for FA.

Merge pull request #1 from grantfirl/gwd_interstitial_reorg

eaed2d3

GWD interstitial reorganization

Merge branch 'dtc/develop' of https://github.com/NCAR/fv3atm into HAF…

9f91d6d

…S_fer_hires

add gwd_post in FA SDF file following other SDFs

c985af4

Update submodule pointers for ccpp-framework and ccpp-physics

b3f213d

Add new SDFs for sas and satmedmfvdifq, remove samfshalcnv_post schem…

9ee222b

…e from existing suites

ccpp/CMakeLists.txt: cleanup bitforbit compiler options, adjust for b…

c69850f

…4b identical results on macOS with 32-bit dynamics

ccpp/config/ccpp_prebuild_config.py: add sas shallow/deep convection …

8cba04d

…schemes and satmedmfvdifq scheme

Update CCPP annotations in gfsphysics/GFS_layer/GFS_physics_driver.F90

2e27a29

gfsphysics/GFS_layer/GFS_typedefs.*: add new control variables/flags …

daba616

…for CCPP for sas and satmedmfvdifq

ccpp/build_ccpp.sh: correct ESMF_DEP_INCPATH settings to avoid warnin…

b9a3887

…gs about non-existent include directories with GNU

gfsphysics/GFS_layer/GFS_typedefs.F90: bugfix, use correct Model%imfd…

a4a0418

…eepcnv_samf instead of Model%imfdeepcnv_sas

makefile: remove include directory and library entries for old FV3/fm…

ae916b3

…s directory to avoid warnings about non-existent include directories with GNU

Update submodule pointers for ccpp-framework and ccpp-physics

a333ffc

Update submodule pointer for ccpp/physics

8c7f231

Update submodule pointers for ccpp/framework and ccpp/physics

1d1b1f8

Merge pull request #3 from climbfuji/ccpp_add_sas_and_satmedmfvdifq

332fb61

Add SAS deep/shallow convection and satmedmfvdifq (updated version of satmedmfvdif) to CCPP

remove update_moist module

c8c7d06

change TRAIN name to accumulated change...

2e3b41d

Merge branch 'dtc/develop' of https://github.com/NCAR/fv3atm into HAF…

dc11553

…S_fer_hires

fix compilation warnings related to wrong arguments passed into 2 sub…

3dba65f

…routines in ugwp_driver_v0.f (note that the code that generated the warnings is in a non-active section)

Merge pull request #2 from mzhangw/HAFS_fer_hires

72c51fc

add Ferrier-Aligo MP scheme changes on host model side

Update submodule pointers for atmos_cubed_sphere ccpp/physics

5975b8e

Merge pull request #5 from climbfuji/update_submodule_pointers_20191122

3fd71cf

dtc/develop: update submodule pointers 2019/11/22

Update submodule pointers for atmos_cubed_sphere ccpp/physics

d94fedd

Revert "dtc/develop: update submodule pointers 2019/11/22"

ed2116d

climbfuji and others added 6 commits June 29, 2020 08:25

Update .gitmodules and submodule pointers for GFDL_atmos_cubed_sphere…

80732fa

… and ccpp-physics for code review and testing

Merge branch 'develop' of https://github.com/NOAA-EMC/fv3atm into dyc…

eed6517

…ore_bugfixes_and_sgscloud_update_gsd_develop_20200629

Revert change to .gitmodules and update submodule pointers for GFDL_a…

1074e93

…tmos_cubed_sphere and ccpp-physics

Merge pull request NOAA-EMC#39 from climbfuji/dycore_bugfixes_and_sgs…

fb4cc41

…cloud_update_gsd_develop_20200629 Two FV3 dycore bug fixes and SGS cloud update gsd/develop 2020/06/29

Update .gitmodules and submodule pointers for GFDL_atmos_cubed_sphere…

636770a

…, ccpp-framework and ccpp-physics for code review and testing

gfsphysics/GFS_layer/GFS_diagnostics.F90: convert cumulative changes …

856c35a

…into tendencies

Remove README.md

62ef03f

climbfuji marked this pull request as ready for review June 30, 2020 15:26

Update submodule pointer for GFDL_atmos_cubed_sphere

8431be9

climbfuji mentioned this pull request Jul 1, 2020

Change standard name "air_temperature_save_from_cumulus_paramterization" to "air_temperature_save_from_convection/convective_paramterization" NCAR/ccpp-physics#467

Closed

Revert change to .gitmodules and update submodule pointers for GFDL_a…

3e7f743

…tmos_cubed_sphere, ccpp-framework, ccpp-physics

DusanJovic-NOAA requested a review from junwang-noaa July 2, 2020 00:48

junwang-noaa reviewed Jul 2, 2020

View reviewed changes

junwang-noaa approved these changes Jul 2, 2020

View reviewed changes

climbfuji mentioned this pull request Jul 2, 2020

Recommended changes / cleanup work for PR https://github.com/NOAA-EMC/fv3atm/pull/134 #137

Closed

DusanJovic-NOAA merged commit 16e3615 into NOAA-EMC:develop Jul 2, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update develop from gsd/develop 2020/06/30 #134

Update develop from gsd/develop 2020/06/30 #134

climbfuji commented Jun 30, 2020 •

edited

Loading

junwang-noaa Jul 2, 2020

climbfuji Jul 2, 2020

junwang-noaa Jul 2, 2020

climbfuji Jul 2, 2020 •

edited

Loading

junwang-noaa Jul 2, 2020

junwang-noaa Jul 2, 2020

junwang-noaa Jul 2, 2020

climbfuji Jul 2, 2020

junwang-noaa left a comment

climbfuji commented Jul 2, 2020

climbfuji commented Jul 2, 2020

Update develop from gsd/develop 2020/06/30 #134

Update develop from gsd/develop 2020/06/30 #134

Conversation

climbfuji commented Jun 30, 2020 • edited Loading

junwang-noaa Jul 2, 2020

Choose a reason for hiding this comment

climbfuji Jul 2, 2020

Choose a reason for hiding this comment

junwang-noaa Jul 2, 2020

Choose a reason for hiding this comment

climbfuji Jul 2, 2020 • edited Loading

Choose a reason for hiding this comment

junwang-noaa Jul 2, 2020

Choose a reason for hiding this comment

junwang-noaa Jul 2, 2020

Choose a reason for hiding this comment

junwang-noaa Jul 2, 2020

Choose a reason for hiding this comment

climbfuji Jul 2, 2020

Choose a reason for hiding this comment

junwang-noaa left a comment

Choose a reason for hiding this comment

climbfuji commented Jul 2, 2020

climbfuji commented Jul 2, 2020

climbfuji commented Jun 30, 2020 •

edited

Loading

climbfuji Jul 2, 2020 •

edited

Loading