202411 Release of SHiELD_physics #58

laurenchilutti · 2024-12-02T14:56:27Z

Description
This PR publishes SHiELD_physics 202411 release. This release coincides with the release of GFDL_atmos_cubed_sphere 202411.

The changes included in this PR are from the GFDL FV3 Team. Full description of changes can be seen in the RELEASE.md

PRs that are all related and should be merged with this one:
GFDL_atmos_cubed_sphere PR #364
SHiELD_build PR #50

Fixes # (issue)

How Has This Been Tested?

Tested on Gaea

Checklist:

Please check all whether they apply or not

My code follows the style guidelines of this project
I have performed a self-review of my own code
I have commented my code, particularly in hard-to-understand areas
I have made corresponding changes to the documentation
My changes generate no new warnings
Any dependent changes have been merged and published in downstream modules

bensonr · 2024-12-02T17:02:35Z

GFS_layer/GFS_driver.F90

+       !ONLY call with the first block, which will be equal to the blocksize.
+       ! this will (should?) allocate the right amount of data. Then calls to
+       ! COSP will have the right data set up.
+       call cosp2_init (size(Grid(1)%xlon,1), Model%levs)


Please make sure this works when non-uniform blocks are created (e.g. [mod (nx, blocksize) != 0]

This does not work with nonuniform blocksizes, to the best of my knowledge. This was a quick 'hack' I put in to get COSP to work in X-SHiELD without running out of memory. This is a limitation of the method at this time.

My original code requires the block size to be the same as the layout size, making it much less efficient. Lucas's simple fix resolves this issue, although I haven't tested it in non-uniform blocks.

bensonr · 2024-12-02T17:16:46Z

gsmphys/mfpbltq.f

+          if (use_tke_ent_det) then
+             xlamax(i) = ce0t(i) / delz(i)
+          else
+             xlamax(i) = ce0 / delz(i)
+          endif 


Since the only difference is whether to use ce0t(i) or ce0 based on use_tke_ent_det==T/F you could have set ce0t(i)==ce0 above for the case where it is False and not had to use the if-test here and in similar loops below.

bensonr · 2024-12-02T17:17:44Z

gsmphys/mfscuq.f

+          ! kgao 12/08/2023
+          if (use_tke_ent_det) then
+             xlamax(i) = ce0t(i) / delz(i)
+          else
+             xlamax(i) = ce0 / delz(i)
+          endif


Since the only difference is whether to use ce0t(i) or ce0 based on use_tke_ent_det==T/F you could have set ce0t(i)==ce0 above for the case where it is False and not had to use the if-test here and in similar loops below.

bensonr · 2024-12-02T17:21:54Z

gsmphys/rad_initialize.f

@@ -136,12 +137,13 @@ subroutine rad_initialize                                         &

      ictmflg= ictm                     ! data ic time/date control flag
      ico2flg= ico2                     ! co2 data source control flag
+      co2_scaling= fco2_scaling         ! co2 level scaling


May want to put a comment here that you are re-setting the value of co2_scaling in physparam.

gsmphys/physparam.f

gsmphys/sfc_drv.f

bensonr · 2024-12-02T17:39:05Z

gsmphys/sfc_noahmp_drv.f

+            chh(i) = -maxevap(i)/evap(i)*chh(i)
+            hflx(i) = -maxevap(i)/evap(i)*hflx(i)


Checking order of operations is supposed to be (maxevap/evap) * _var_ and you want to negate it.

There were two questions here. One was to make sure it wasn't supposed to be -maxevap/(evap * var) and secondly questioning the use of the negative sign.

gsmphys/sfc_nst.f

gsmphys/sfc_sice.f

gsmphys/sfc_ocean.f

lharris4

Looks OK to me. @kaiyuan-cheng @kaiyuan-cheng @spencerkclark could you check your changes to make sure they are OK?

lharris4 · 2024-12-06T14:52:52Z

GFS_layer/GFS_driver.F90

+       !ONLY call with the first block, which will be equal to the blocksize.
+       ! this will (should?) allocate the right amount of data. Then calls to
+       ! COSP will have the right data set up.
+       call cosp2_init (size(Grid(1)%xlon,1), Model%levs)


This does not work with nonuniform blocksizes, to the best of my knowledge. This was a quick 'hack' I put in to get COSP to work in X-SHiELD without running out of memory. This is a limitation of the method at this time.

kaiyuan-cheng · 2024-12-06T21:24:14Z

Looks OK to me. @kaiyuan-cheng @kaiyuan-cheng @spencerkclark could you check your changes to make sure they are OK?

Just submitted a PR addressing the CO2 scaling comments.

Regarding
chh(i) = -maxevap(i)/evap(i)*chh(i)
hflx(i) = -maxevap(i)/evap(i)*hflx(i),
the original formulation looks good to me.

Co-authored-by: Kun Gao <[email protected]>

laurenchilutti · 2024-12-09T13:41:13Z

Just merged in @gaokun227 and @kaiyuan-cheng updates based on @bensonr review.

bensonr

thanks for making the updates

spencerkclark

Thanks @laurenchilutti! Just a couple minor comments related to code that was touched in both the public and private repos.

spencerkclark · 2024-12-11T16:08:21Z

GFS_layer/GFS_radiation_driver.F90

+      if (Model%ldiag3d) then
+        call compute_column_integrated_moles_of_dry_air_and_co2(Statein, gasvmr, IM, LMK, NF_VGAS, Diag)
+      endif


Following #50, I think we no longer need to guard against calling this when Model%ldiag3d is .false.:

Suggested change

if (Model%ldiag3d) then

call compute_column_integrated_moles_of_dry_air_and_co2(Statein, gasvmr, IM, LMK, NF_VGAS, Diag)

endif

call compute_column_integrated_moles_of_dry_air_and_co2(Statein, gasvmr, IM, LMK, NF_VGAS, Diag)

spencerkclark · 2024-12-11T23:05:03Z

GFS_layer/GFS_typedefs.F90

+      allocate (Diag%column_moles_co2_per_square_meter(IM))
+      allocate (Diag%column_moles_dry_air_per_square_meter(IM))


Following #50, these are always allocated above, so these lines are no longer needed:

Suggested change

allocate (Diag%column_moles_co2_per_square_meter(IM))

allocate (Diag%column_moles_dry_air_per_square_meter(IM))

laurenchilutti and others added 10 commits November 8, 2024 09:47

release of FV3Team development of phsyics for SHiELD

e2efc0c

revert changes on atmos_shared

16a337b

update RELEASE and README.

373f6c6

rearrange vpert to avoid multiple scaling divisions

191539e

enforce DP parameters

91d6ecd

remove unnecessary temp arrays

174f695

ensure DP cutoff values

8c5cdda

Merge remote-tracking branch 'joseph/repro_fix' into 2024release

b2804a6

enforcing DP in the pbl coefficients

f4b02f6

Merge remote-tracking branch 'origin/main' into 2024release

8fe3638

laurenchilutti requested review from bensonr and lharris4 December 2, 2024 14:56

laurenchilutti mentioned this pull request Dec 2, 2024

202411 Release of FV3 Dynamical Core NOAA-GFDL/GFDL_atmos_cubed_sphere#364

Open

6 tasks

bensonr reviewed Dec 2, 2024

View reviewed changes

lharris4 approved these changes Dec 6, 2024

View reviewed changes

gaokun227 and others added 2 commits December 9, 2024 08:39

optimize PBL entrainment/detrainment rate calculations (#2)

614748b

Co-authored-by: Kun Gao <[email protected]>

Add notes and stdout related to co2 scaling (#3)

6816b01

bensonr approved these changes Dec 9, 2024

View reviewed changes

laurenchilutti requested a review from spencerkclark December 11, 2024 13:53

spencerkclark approved these changes Dec 11, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

202411 Release of SHiELD_physics #58

202411 Release of SHiELD_physics #58

laurenchilutti commented Dec 2, 2024 •

edited

Loading

bensonr Dec 2, 2024

lharris4 Dec 6, 2024

linjiongzhou Dec 12, 2024

bensonr Dec 2, 2024

bensonr Dec 2, 2024

bensonr Dec 2, 2024

bensonr Dec 2, 2024

bensonr Dec 4, 2024

lharris4 left a comment

lharris4 Dec 6, 2024

kaiyuan-cheng commented Dec 6, 2024 •

edited

Loading

laurenchilutti commented Dec 9, 2024

bensonr left a comment

spencerkclark left a comment

spencerkclark Dec 11, 2024

spencerkclark Dec 11, 2024

		chh(i) = -maxevap(i)/evap(i)*chh(i)
		hflx(i) = -maxevap(i)/evap(i)*hflx(i)

		allocate (Diag%column_moles_co2_per_square_meter(IM))
		allocate (Diag%column_moles_dry_air_per_square_meter(IM))

202411 Release of SHiELD_physics #58

Are you sure you want to change the base?

202411 Release of SHiELD_physics #58

Conversation

laurenchilutti commented Dec 2, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lharris4 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kaiyuan-cheng commented Dec 6, 2024 • edited Loading

laurenchilutti commented Dec 9, 2024

bensonr left a comment

Choose a reason for hiding this comment

spencerkclark left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

laurenchilutti commented Dec 2, 2024 •

edited

Loading

kaiyuan-cheng commented Dec 6, 2024 •

edited

Loading