cam6_3_028: Science Updates and bug fixes for SE dycore. #401

jtruesdal · 2021-07-13T15:24:47Z

These mods are necessary to fix science bugs, improve algorithms and numerical stability of SE dycore. Namelist default changes to the gravity wave parameterization and updates to the Test Tracers will also affect the other dycores. This PR will close issue #392.

Improvements:

alleviate noise over steep orography
sponge layer changes for stability
Changes to TT_CCOSB TT_COSB and TT_lCCOSB for idealized testing of dynamical cores (change to Lauritzen and Thuburn,2012, QJRMS, formulation).
New parameterization defaults for improved stability and to provide better baseline defaults.

Bug fixes:

species dependent thermodynamics in WACCM-X
mods to physconst.F90 for WACCM-x and SE (mid pressure calculation consistency)
initialization bug in angular momentum diagnostics

Refactor:

Remove Rayleigh Friction Parameterization from SE
remove vertical diffusion code in SE dycore

… stability

src/dynamics/se/dp_coupling.F90

src/dynamics/se/dyn_comp.F90

PeterHjortLauritzen · 2021-07-20T18:03:45Z

Cecile brought to our attention that del4 sponge layer settings are hard coded. We have code to move those settings to namelist.

jtruesdal · 2021-07-20T18:18:20Z

@PeterHjortLauritzen I have finished my testing of the refactored code that generalizes the sponge layer setup and removes the hard coded layer settings. I will commit the fix for the rest of the reviewers to look at.

nusbaume

Looks good to me, but I did have some questions and requests.

src/dynamics/se/dp_coupling.F90

src/dynamics/se/dycore/dimensions_mod.F90

src/dynamics/se/dycore/global_norms_mod.F90

src/dynamics/se/dycore/control_mod.F90

src/dynamics/se/dycore/viscosity_mod.F90

src/dynamics/se/dycore/prim_advance_mod.F90

…- cam6_3_027

nusbaume · 2021-07-22T16:57:14Z

bld/namelist_files/namelist_definition.xml

+<entry id="se_sponge_del4_nu_div_fac" type="integer" category="se"
+       group="dyn_se_inparm" valid_values="" >
+Divergence damping hyperviscosity coefficient se_nu_div [m^4/s] for u,v is increased to
+se_nu_p*se_sponge_del4_nu_fac following a hyperbolic tangent function


I believe it should be se_nu_p*se_sponge_del4_nu_div_fac here.

nusbaume · 2021-07-22T16:59:32Z

doc/ChangeLog

+	   If less than 0, se_sponge_del4_nu_fac is automatically set based on model top location.
+	se_sponge_del4_nu_div_fac
+	.  Divergence damping hyperviscosity coefficient se_nu_div [m^4/s] for u,v is increased to
+           se_nu_p*se_sponge_del4_nu_fac following a hyperbolic tangent function


I believe it should be se_nu_p*se_sponge_del4_nu_div_fac here as well.

src/dynamics/se/dycore/global_norms_mod.F90

cacraigucar · 2021-07-22T14:55:32Z

src/dynamics/se/dycore/control_mod.F90

  logical, public :: variable_nsplit=.false.

-  integer, public :: phys_dyn_cp = 0 !=0; no thermal energy scaling of T increment
+  integer, public :: phys_dyn_cp = 1 !=0; no thermal energy scaling of T increment


Since you modified the default value of this namelist, I wanted to point out that we are moving towards not having namelist values initialized to anything other than the following (from the Fortran Coding Standards wiki). Something for the next PR.

All namelist variables except for logical quantities are initialized to invalid values (integer: -HUGE(1), real: Nan, character: 'UNSET').

Got it. I'll look over the Fortran Coding Standards wiki before the next PR.

cacraigucar · 2021-07-22T15:57:28Z

src/utils/physconst.F90

       kmcnd(:ncol,:,lchnk), pcnst, tracer=mmr(:ncol,:,:), fact=to_moist_fact(:ncol,:),                &
       active_species_idx_dycore=thermodynamic_active_species_idx)
-
+    cappav(:ncol,:,lchnk) = rairv(:ncol,:,lchnk)/cpairv(:ncol,:,lchnk)


Sorry we missed this being removed in cam6_2_034. Did it have a big impact on results? This is obviously one of the bug fixes which needs to be applied to CESM2.2.

Yes this is a bug ("only" impacts WACCM-x applications) that needs to be fixed in the release branch and, as far as I understand, will be tackled in a separate pull request. We have not assessed impacts on simulation but the WACCM-x group did a sanity check with the buggy code and did not report issues.

bld/namelist_files/namelist_defaults_cam.xml

nusbaume · 2021-07-22T17:37:46Z

src/utils/physconst.F90

+                 residual     = residual - mm
+               end do
+               icnst=dry_air_species_num
+               ispecies = idx_local(icnst)


I don't believe you need ispecies here, as it is not being used.

I'll get rid of ispecies here.

src/dynamics/se/dycore/viscosity_mod.F90

nusbaume

Looks great to me now! Just a few final (very minor) requests I missed the first go-around.

gold2718

I think it is important to log namelist values, especially if they are set in the code (see my comment). Did I miss the print statements?

src/dynamics/se/dycore/global_norms_mod.F90

doc/ChangeLog

cacraigucar · 2021-07-23T15:34:30Z

doc/ChangeLog

+. bld/build-namelist
+	- Remove SE Rayleigh Friction Parameterization


Totally at your discretion (i.e. no need to change). I typically combine several files that have the exact same description together and then have the description follow. In this case, build-namelist, user_nl_cam, dyn_comp.F90 could all be grouped together with the "Remove SE Rayleigh Friction Parameterization" description.

gold2718

Looking good now, thanks!

gold2718 · 2021-07-23T22:45:06Z

src/dynamics/se/dycore/global_norms_mod.F90

+    if (hybrid%masterthread.and.div_set) &
+         write(iulog, '(a,e9.2)')   '  sponge_del4_nu_div_fac = ',sponge_del4_nu_div_fac
+    if (hybrid%masterthread.and.lev_set) &
+         write(iulog, '(a,i3)')   '  sponge_del4_lev        = ',sponge_del4_lev


FYI, i0 will use the exact needed number of characters to write out an integer value.

Thanks! It is new information to me and I will change to i0.

jtruesdal added 2 commits July 8, 2021 15:13

Full-physics physics-dynamics coupling with MPAS dynamical core

f06ba04

Changelog testing update, se_hypervis_subcycle_sponge mod for WACCM-X…

9d9e8f6

… stability

cacraigucar added this to the CESM2.3 milestone Jul 13, 2021

updates to ChangeLog (Had wrong description)

bba0979

gold2718 requested review from cacraigucar, fvitt, gold2718, nusbaume and peverwhee July 19, 2021 21:13

gold2718 added enhancement New feature or request bug-fix This PR was created to fix a specific bug. labels Jul 19, 2021

fvitt reviewed Jul 19, 2021

View reviewed changes

src/dynamics/se/dp_coupling.F90 Outdated Show resolved Hide resolved

src/dynamics/se/dyn_comp.F90 Show resolved Hide resolved

move del4 sponge layer parameters to namelist

e639045

peverwhee approved these changes Jul 20, 2021

View reviewed changes

bug in previous commit

ecdb4ca

Generalize sponge layer setup and fix hard coded layer settings.

6f759fc

nusbaume requested changes Jul 20, 2021

View reviewed changes

jtruesdal added 4 commits July 21, 2021 09:39

updates requested as part of PR#401

ee0edee

PR request to rename nu_s/nu_s_lev to nu_t/nu_t_lev for clarity

ef544d0

bring science updates for CESM2.2 from cam6_3_025 to cam_development …

b5a6e0d

…- cam6_3_027

ChangeLog merge with cam6_3_027

3db5d65

nusbaume reviewed Jul 22, 2021

View reviewed changes

cacraigucar approved these changes Jul 22, 2021

View reviewed changes

nusbaume reviewed Jul 22, 2021

View reviewed changes

nusbaume approved these changes Jul 22, 2021

View reviewed changes

cacraigucar changed the title ~~Science Updates for CESM2.2 release and cam_development~~ cam6_3_028?: Science Updates for CESM2.2 release and cam_development Jul 22, 2021

gold2718 requested changes Jul 23, 2021

View reviewed changes

src/dynamics/se/dycore/global_norms_mod.F90 Outdated Show resolved Hide resolved

src/dynamics/se/dycore/global_norms_mod.F90 Show resolved Hide resolved

doc/ChangeLog Outdated Show resolved Hide resolved

jtruesdal changed the title ~~cam6_3_028?: Science Updates for CESM2.2 release and cam_development~~ cam6_3_028?: Science Updates and bug fixes for SE dycore. Jul 23, 2021

PR requests

5c88161

jtruesdal requested a review from gold2718 July 23, 2021 15:30

cacraigucar approved these changes Jul 23, 2021

View reviewed changes

gold2718 approved these changes Jul 23, 2021

View reviewed changes

fvitt approved these changes Jul 25, 2021

View reviewed changes

cacraigucar changed the title ~~cam6_3_028?: Science Updates and bug fixes for SE dycore.~~ cam6_3_028: Science Updates and bug fixes for SE dycore. Jul 26, 2021

cacraigucar changed the title ~~cam6_3_028: Science Updates and bug fixes for SE dycore.~~ Science Updates and bug fixes for SE dycore. Jul 27, 2021

cacraigucar changed the title ~~Science Updates and bug fixes for SE dycore.~~ cam6_3_028: Science Updates and bug fixes for SE dycore. Aug 18, 2021

jtruesdal added 3 commits August 18, 2021 18:09

i0 format suggestion by Steve

97f989b

update Changelog with climate validation runs

a8dcb53

update Changelog for tag

de2e10f

jtruesdal merged commit 857ca55 into ESCOMP:cam_development Aug 19, 2021

gold2718 mentioned this pull request Sep 21, 2021

Spectral Element science updates/bug fixes/refactor (aka CESM2.2-updates) #392

Closed

nusbaume mentioned this pull request Nov 22, 2021

Need to bring CAM SE-dycore updates into CAMDEN ESCOMP/CAM-SIMA#168

Open

		. bld/build-namelist
		- Remove SE Rayleigh Friction Parameterization

cam6_3_028: Science Updates and bug fixes for SE dycore. #401

cam6_3_028: Science Updates and bug fixes for SE dycore. #401

Uh oh!

Conversation

jtruesdal commented Jul 13, 2021

Uh oh!

Uh oh!

Uh oh!

PeterHjortLauritzen commented Jul 20, 2021

Uh oh!

jtruesdal commented Jul 20, 2021

Uh oh!

nusbaume left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

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!

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!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

nusbaume left a comment

Choose a reason for hiding this comment

Uh oh!

gold2718 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gold2718 left a comment

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants