Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) #4835

mark-petersen · 2022-03-16T21:25:29Z

Currently the internal gravity wave speed (first baroclinic mode) in the ocean GM parameterization is a constant, previously named config_GM_constant_gravWaveSpeed but changed in this PR to config_GM_constant_bclModeSpeed. This remains the default, but this PR adds a new flag, config_GM_minBclModeSpeed_method: If 'constant' (the default) then use config_GM_constant_bclModeSpeed. If 'computed' then compute the gravity wave speed at every edge at every time step using the Brunt-Vaisala frequency. This functionality was tested by the AMOC focus team in 2021 on the branch amoc-experimental-branch.

[NML]
[BFB]

vanroekel

@mark-petersen this is a great idea to implement, thanks for the PR. I had some confusion though when going through it. It's possible I'm not understanding something

components/mpas-ocean/src/shared/mpas_ocn_gm.F

vanroekel · 2022-03-17T04:24:08Z

@mark-petersen I realize I got ,confused by your PR description and it led to incorrect comments on my review. Your title suggests you are implementing a spatially variable computation of c and this made me thing of cGMphaseSpeed, but this PR is really just to do a spatially variable c_min. So a few suggestions

change config_GM_gravwavespeed to something like config_GM_gravwavespeed_min
same for the gm_constant_gravWaveSpeed variable
Can the PR description be changed as well without too much trouble?
Although more likely the subject of a separate PR gravWaveSpeed is not a correct description. It is actually the first baroclinic mode, it would be nice to drop the grav or change that to something like bcl, but again, I don't think that needs to be here.

I think with those first 3 clarifications this PR will be good to go from my view.

mark-petersen · 2022-03-20T14:34:41Z

components/mpas-ocean/src/shared/mpas_ocn_gm.F

+               c_min = gm_minBclModeSpeed_constant + gm_minBclModeSpeed_compute_on*max(0.01_RKIND,sumN2/ltSum*(0.5*(lt1+lt2)))
+
+               ! Compute the speed of the first baroclinic mode from the Brunt-Vaisala frequency.
+               cGMphaseSpeed(iEdge) = max(c_min,                          &
                                          sumN2/(config_GM_spatially_variable_baroclinic_mode*3.141592_RKIND))


@vanroekel I changed the flag names to

config_GM_constant_bclModeSpeed = 0.3 config_GM_minBclModeSpeed_method = 'constant'

The only substantial change is the line above, but I use fixed coefficient values set on init to avoid if statements inside the loop.

mark-petersen · 2022-03-20T14:37:42Z

I rebased this on today's master. It now passes nightly suite bfb with both gnu and intel against master. Also passes:

SMS_D.T62_oQU120_ais20.MPAS_LISIO_TEST.anvil_intel
SMS_D.T62_oQU120_ais20.MPAS_LISIO_TEST.anvil_gnu

@stephenprice, this is ready for the next cryo test in the sequence. You only needed to change this flag to

config_GM_minBclModeSpeed_method = 'computed'

to invoke the changes in this PR.

milenaveneziani · 2022-03-23T19:36:56Z

My 5-year test with ECwISC30to60E2r1, this branch, and the setting config_GM_minBclModeSpeed_method = 'computed' did not produce any visible difference with a previous 5-year test that used a baseline configuration.

MPAS-Analysis for this branch:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.milena/20220322.CRYO1850.ne30pg2_ECwISC30to60E2r1.chrysalis.CryoBranchGMphase/Years1-5/
MPAS-Analysis for the baseline:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.milena/20220301.CRYO1850.ne30pg2_ECwISC30to60E2r1.chrysalis.CryoBranchBaseline/Years1-5/

vanroekel

@mark-petersen this looks great. This clears up the misunderstanding I had on this PR in my initial comments. I'm approving by visual inspection.

vanroekel · 2022-03-29T18:59:45Z

@mark-petersen but to verify this PR is BFB right? It looked it to me

mark-petersen · 2022-03-29T19:31:38Z

Yes, if config_GM_gravwavespeed='const', the default, then it is bfb with previous, and that was shown in the testing. Thanks @vanroekel

mark-petersen · 2022-03-30T14:17:23Z

@jonbob this is ready to merge when the repo is open. Thanks.

jonbob · 2022-03-30T21:14:03Z

@mark-petersen - the PR description is no longer current with the code changes? I don't see those variable names in the code

jonbob · 2022-04-04T16:14:41Z

@mark-petersen - I updated the bld files with output from the auto-generating scripts. There were some changes to the namelist definition file, if you want to make sure you agree.

mark-petersen · 2022-04-04T16:28:33Z

Thanks @jonbob, I updated the description. Your updated scripts are correct.

jonbob · 2022-04-04T21:20:30Z

@mark-petersen - this PR is failing SMS_D_Ld3.T62_oQU120.CMPASO-IAF.chrysalis_intel. I'll look at it more this evening

mark-petersen · 2022-04-05T18:37:14Z

@jonbob this now passes SMS_D_Ld3.T62_oQU120.CMPASO-IAF.chrysalis_intel. I initialized thread-private scalars outside the threaded loop.

jonbob · 2022-04-05T18:45:48Z

That fix makes sense, @mark-petersen. I'll re-test and aim to merge today

Add choice of computed or constant ocean GM gravity wave speed Currently the internal gravity wave speed (first baroclinic mode) in the ocean GM parameterization is a constant, previously named config_GM_constant_gravWaveSpeed but changed in this PR to config_GM_constant_bclModeSpeed. This remains the default, but this PR adds a new flag, config_GM_minBclModeSpeed_method: If 'constant' (the default) then use config_GM_constant_bclModeSpeed. If 'computed' then compute the gravity wave speed at every edge at every time step using the Brunt-Vaisala frequency. This functionality was tested by the AMOC focus team in 2021 on the branch amoc-experimental-branch. [NML] [BFB]

jonbob · 2022-04-05T19:35:00Z

passed:

ERS.ne11_oQU240.WCYCL1850NS.chrysalis_intel
SMS_D_Ld3.T62_oQU120.CMPASO-IAF.chrysalis_intel

merged to next

jonbob · 2022-04-06T16:49:02Z

merged to master and expected NML DIFFs blessed

mark-petersen added BFB PR leaves answers BFB NML labels Mar 16, 2022

mark-petersen requested a review from vanroekel March 16, 2022 21:25

mark-petersen assigned jonbob Mar 16, 2022

mark-petersen force-pushed the mark-petersen/ocn/gm-min-phase-speed branch from 96a2036 to f2253c1 Compare March 16, 2022 21:35

vanroekel requested changes Mar 16, 2022

View reviewed changes

components/mpas-ocean/src/shared/mpas_ocn_gm.F Show resolved Hide resolved

components/mpas-ocean/src/shared/mpas_ocn_gm.F Outdated Show resolved Hide resolved

components/mpas-ocean/src/shared/mpas_ocn_gm.F Outdated Show resolved Hide resolved

mark-petersen added 4 commits March 19, 2022 22:14

add choice of computed or constant GM gravity wave speed

d46c766

add initial definitions of lt1 and lt2

486b9c9

Update name to bclModeSpeed

b2365dd

Add explanation in comments

91b5bac

mark-petersen commented Mar 20, 2022

View reviewed changes

mark-petersen force-pushed the mark-petersen/ocn/gm-min-phase-speed branch from 6ab06ba to 91b5bac Compare March 20, 2022 14:45

mark-petersen added the mpas-ocean label Mar 21, 2022

vanroekel approved these changes Mar 29, 2022

View reviewed changes

jonbob mentioned this pull request Mar 31, 2022

Add horizontal tapering function to mesoscale eddy parameterizations (AMOC PR 2/6) #4864

Merged

Update bld files with output from scripts to match changes to Registry

0195995

mark-petersen mentioned this pull request Apr 4, 2022

Add ocean option for Redi kappa to be set equal to GM (AMOC PR 3/6) #4867

Closed

mark-petersen mentioned this pull request Apr 4, 2022

limit Visbeck eddy length to dcEdge (AMOC PR 4/6) #4868

Merged

mark-petersen changed the title ~~add choice of computed or constant ocean GM gravity wave speed~~ Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) Apr 5, 2022

move lt1,lt2 initialization within loop for threading

c99879a

jonbob merged commit 565da94 into master Apr 6, 2022

jonbob deleted the mark-petersen/ocn/gm-min-phase-speed branch April 6, 2022 16:49

xylar mentioned this pull request May 4, 2022

Update E3SM-Project to E3SM/master MPAS-Dev/compass#383

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) #4835

Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) #4835

mark-petersen commented Mar 16, 2022 •

edited

Loading

vanroekel left a comment

vanroekel commented Mar 17, 2022

mark-petersen Mar 20, 2022

mark-petersen commented Mar 20, 2022 •

edited

Loading

milenaveneziani commented Mar 23, 2022

vanroekel left a comment

vanroekel commented Mar 29, 2022

mark-petersen commented Mar 29, 2022

mark-petersen commented Mar 30, 2022

jonbob commented Mar 30, 2022 •

edited

Loading

jonbob commented Apr 4, 2022

mark-petersen commented Apr 4, 2022

jonbob commented Apr 4, 2022

mark-petersen commented Apr 5, 2022 •

edited

Loading

jonbob commented Apr 5, 2022

jonbob commented Apr 5, 2022

jonbob commented Apr 6, 2022

Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) #4835

Add choice of computed or constant ocean GM gravity wave speed (AMOC PR 1/6) #4835

Conversation

mark-petersen commented Mar 16, 2022 • edited Loading

vanroekel left a comment

Choose a reason for hiding this comment

vanroekel commented Mar 17, 2022

mark-petersen Mar 20, 2022

Choose a reason for hiding this comment

mark-petersen commented Mar 20, 2022 • edited Loading

milenaveneziani commented Mar 23, 2022

vanroekel left a comment

Choose a reason for hiding this comment

vanroekel commented Mar 29, 2022

mark-petersen commented Mar 29, 2022

mark-petersen commented Mar 30, 2022

jonbob commented Mar 30, 2022 • edited Loading

jonbob commented Apr 4, 2022

mark-petersen commented Apr 4, 2022

jonbob commented Apr 4, 2022

mark-petersen commented Apr 5, 2022 • edited Loading

jonbob commented Apr 5, 2022

jonbob commented Apr 5, 2022

jonbob commented Apr 6, 2022

mark-petersen commented Mar 16, 2022 •

edited

Loading

mark-petersen commented Mar 20, 2022 •

edited

Loading

jonbob commented Mar 30, 2022 •

edited

Loading

mark-petersen commented Apr 5, 2022 •

edited

Loading