Add the subroutines of GF cps for CCPP-FV3 by haiqinli · Pull Request #139 · NCAR/ccpp-physics

haiqinli · 2018-08-08T21:34:08Z

Create the new branch of gfcps to include the subroutines of Grell-Freitas cps for CCPP-FV3.

climbfuji · 2018-08-09T15:59:56Z

       random_clds    = .true.
       trans_trac     = .true.
       cnvcld         = .true.
-       imfshalcnv     = 2


I am not sure if you are supposed to change the default model config. Maybe @grantfirl has a better idea on how to create a different input.nml that uses GF instead of SAS (and a corresponding suite definition file for both FV3 and SCM).

Which brings up the question why a copy of input.nml resides in ccpp-physics - it shouldn't be there. @grantfirl is it actually used? Not by FV3 at least, but maybe by SCM?

To follow up on this, @haiqinli, Dom has (rightly) moved the input.nml used in the SCM from the ccpp-physics repository to the gmtb-scm repository. It is now located in gmtb-scm/scm/etc/case_config in the master branch. Also as Dom suggested, for testing the GF scheme, it is best to create a copy of input.nml for your testing, perhaps input_GF.nml. The original represents the operational namelist values and should stay that way. Then, when you go to run the SCM, you'll also want to modify the case configuration namelists to point to your modified input_GF.nml. For example, if you run the TWP-ICE case, make a copy of gmtb-scm/scm/etc/case_config/twpice.nml, rename it twpice_GF.nml (or similar) and edit the physics_nml variable within to point to your new input_GF.nml file.

As far as I can see, this has not been addressed yet.

climbfuji · 2018-08-09T16:02:30Z

+
+!> \brief brief description of the subroutine
+!!
+!! \section arg_table_sasas_deep_init argument table


This should be "arg_table_cu_gf_deep_init"

Or, even better, since these routines are not CCPP entry points, no argument table is needed - delete lines 46-47

climbfuji · 2018-08-09T16:03:06Z

+
+!> \brief brief description of the subroutine
+!!
+!! \section arg_table_sasas_deep_finalize argument table


This should be "arg_table_cu_gf_deep_finalize"

Or, even better, since these routines are not CCPP entry points, no argument table is needed - delete lines 54-55

climbfuji · 2018-08-09T16:05:31Z

+   use machine   , only: kind_phys
+   use cu_gf_deep, only: cu_gf_deep_run,neg_check,autoconv,aeroevap
+   use cu_gf_sh  , only: cu_gf_sh_run
+


missing:

private public :: cu_gf_driver_init, cu_gf_driver_run, cu_gf_driver_finalize

climbfuji · 2018-08-09T16:05:55Z

@@ -0,0 +1,786 @@
+!
+module cu_gf_driver
+   use physcons  , g => con_g, cp => con_cp, xlv => con_hvap, r_v => con_rv


These physical constants should come in via the subroutine's argument lists.

climbfuji · 2018-08-09T16:06:22Z

+!! | hfx2           | kinematic_surface_upward_sensible_heat_flux               | kinematic surface upward sensible heat flux         | K m s-1 |    1 | real      | kind_phys | out    | F        |
+!! | qfx2           | kinematic_surface_upward_latent_heat_flux                 | kinematic surface upward latent heat flux           | kg kg-1 m s-1 |    1 | real             | kind_phys | out    | F        |
+!! | clw            | convective_transportable_tracers                          | array to contain cloud water and other convective trans. tracers         | kg kg-1       |    3 | real        | kind_phys | none   | F        |
+!! | pbl           | atmosphere_boundary_layer_thickness    | PBL thickness     | m             |    1      | real      | kind_phys | out    | F        |


adjust intents

climbfuji · 2018-08-09T16:07:12Z

+               pbl,ud_mf,dd_mf,dt_mf,cnvw,cnvc,errmsg,errflg)
+!              pbl,ud_mf,dd_mf,dt_mf,gdc,gdc2,cnvw,cnvc,ishal_cnv)
+!-------------------------------------------------------------
+   implicit none


preferably all interface variables (i.e. those coming in via the argument list) are listed first, followed by local parameters and variables

climbfuji · 2018-08-09T16:09:53Z

+      subroutine cu_gf_driver_post_finalize()
+      end subroutine cu_gf_driver_post_finalize
+
+!> \section arg_table_cu_gf_driver_post_run Argument Table


instead of having Model, Stateout, Grid, Tbd in the argument list (of which only Tbd and Stateout are used anyway), it would be better to pass in the components phy_f3d(:,:,5), phy_f3d(:,:,6), gt0(:,:) and gq0(:,:,1) using their respective standard names (see GFS_typedefs.F90) and giving them meaningful local names.

climbfuji · 2018-08-09T16:10:40Z

+!> \section arg_table_cu_gf_driver_pre_run Argument Table
+!! | local_name     | standard_name                                          | long_name                                                                | units         | rank | type              |    kind   | intent | optional |
+!! |----------------|--------------------------------------------------------|--------------------------------------------------------------------------|---------------|------|-------------------|-----------|--------|----------|
+!! | Model          | FV3-GFS_Control_type                                   | Fortran DDT containing FV3-GFS model control parameters                  | DDT           |    0 | GFS_control_type  |           | in     | F        |


see above comment for cu_gf_driver_post_run (replace GFS derived data types with individual components)

climbfuji · 2018-08-09T16:13:33Z

@@ -0,0 +1,4754 @@
+module cu_gf_deep


this is a pretty minor request, but it would be great to declare all variables/subroutines as private and only those variables and subroutines used as public. For instance, adding

private public :: LIST OF VARIABLES AND SUBROUTINES THAT GET USED BY THE CU_GF_DRIVER MODULE

haiqinli · 2018-08-09T16:26:19Z

Hi Dom, Thanks a lot for your comments and suggestions. I will work on it. Best Wishes Haiqin

…

On Thu, Aug 9, 2018 at 10:13 AM, Dom Heinzeller ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In physics/cu_gf_deep.F90 <#139 (comment)>: > @@ -0,0 +1,4754 @@ +module cu_gf_deep this is a pretty minor request, but it would be great to declare all variables/subroutines as private and only those variables and subroutines used as public. For instance, adding private public :: LIST OF VARIABLES AND SUBROUTINES THAT GET USED BY THE CU_GF_DRIVER MODULE — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#139 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ak4AOOkRdOb1thx1FTPK0rPaSc3F7pnyks5uPF-ugaJpZM4V0sw5> .

climbfuji · 2018-08-09T16:41:57Z

Sure, no problems. I am still working my way through, but all in all your work looks great.

climbfuji · 2018-08-09T16:44:08Z

+
+!> \brief brief description of the subroutine
+!!
+!! \section arg_table_sasas_deep_init argument table


same as for cu_gf_deep, no variable table needed here

climbfuji · 2018-08-09T16:44:25Z

+
+!> \brief brief description of the subroutine
+!!
+!! \section arg_table_sasas_deep_finalize argument table


no variable table needed here

climbfuji · 2018-08-09T16:44:43Z

@@ -0,0 +1,866 @@
+! module cup_gf_sh will call shallow convection as described in grell and
+! freitas (2016). input variables are:


As for cu_gf_deep: this is a pretty minor request, but it would be great to declare all variables/subroutines as private and only those variables and subroutines used as public. For instance, adding

private public :: LIST OF VARIABLES AND SUBROUTINES THAT GET USED BY THE CU_GF_DRIVER MODULE

climbfuji

Great job overall, I noted a few issues and a few nice to haves.

climbfuji · 2018-10-05T15:02:30Z

Associated PRs:

NCAR/ccpp-framework#119
#139
https://github.com/NCAR/FV3/pull/53
NCAR/ccpp-scm#45

climbfuji · 2018-10-05T23:31:39Z

Superseded by #166.

The change gets around snow initialization with the use of RUC LSM

clm lake fix: pressure is not density + NCAR#142 and NCAR#148

haiqinli added 2 commits August 8, 2018 17:09

"add for gfcps"

3427618

"for gfcps"

e340df5

haiqinli requested review from climbfuji, grantfirl and llpcarson as code owners August 8, 2018 21:34

climbfuji reviewed Aug 9, 2018

View reviewed changes

climbfuji requested changes Aug 9, 2018

View reviewed changes

climbfuji mentioned this pull request Aug 9, 2018

GF CU scheme in CCPP #98

Closed

climbfuji added work in progress do not merge labels Sep 3, 2018

haiqinli added 2 commits September 21, 2018 17:27

"restore to the default setting"

db4f729

"update GF to be independent of microphysics"

f1fa403

climbfuji mentioned this pull request Oct 5, 2018

GF CU scheme in CCPP NCAR/ccpp-scm#45

Closed

This was referenced Oct 5, 2018

Add GF cps for CCPP-FV3 NCAR/ccpp-framework#119

Merged

CCPP-compliant Grell-Freitas convection scheme #166

Merged

climbfuji added the wontfix label Oct 5, 2018

climbfuji merged commit f1fa403 into NCAR:master Oct 10, 2018

SamuelTrahanNOAA added a commit to SamuelTrahanNOAA/ccpp-physics that referenced this pull request May 17, 2022

Merge pull request NCAR#139 from tanyasmirnova/rrfs_snow_ini

5a8a8f9

The change gets around snow initialization with the use of RUC LSM

scrasmussen pushed a commit to scrasmussen/ccpp-physics that referenced this pull request Dec 25, 2023

Merge pull request NCAR#139 from SamuelTrahanNOAA/clm-unit-fix

0cdfc9d

clm lake fix: pressure is not density + NCAR#142 and NCAR#148

		@@ -0,0 +1,866 @@
		! module cup_gf_sh will call shallow convection as described in grell and
		! freitas (2016). input variables are:

Conversation

haiqinli commented Aug 8, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

haiqinli commented Aug 9, 2018 via email

Uh oh!

climbfuji commented Aug 9, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

climbfuji left a comment

Choose a reason for hiding this comment

Uh oh!

climbfuji commented Oct 5, 2018

Uh oh!

climbfuji commented Oct 5, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants