Adjust MALI SMB in G-cases and update hist fields #7463

matthewhoffman · 2025-06-25T23:48:08Z

This branch updates a number of details about MALI coupling in preparation of production runs:

When ELM is not active (e.g. G-cases), MALI will now use a data SMB from the MALI initial file
The data volume of MALI history files is vastly reduced

[NML] only for configurations with MALI
[non-BFB] only for configurations with MALI

This commit adds the ability for MALI to use a data surface mass balance field from the MALI input file in compsets where ELM is not active. This is done by adding a new namelist option in MALI called config_smb_source and making its value controlled by the CIME variable GLC_TWO_WAY_COUPLING which is only true when both ELM and MALI are active.

Diagnostic/debugging fields that had been added previously have been removed to reduce file size. Output frequency has been reduced to 5 days.

trhille

@matthewhoffman, these changes look correct to me. Is there anything you'd like me to test before I approve?

matthewhoffman · 2025-07-01T14:50:49Z

@trhille , no, I just wanted you to look over the implementation and naming choices for config_smb_source, and to have you do a sanity check on the changes to the MALI history files.

trhille · 2025-07-01T15:13:08Z

components/mpas-albany-landice/cime_config/buildnml

-        lines.append('    <var name="ismip6shelfMelt_zOcean"/>')
-        lines.append('    <var name="ismip6shelfMelt_zBndsOcean"/>')
-        lines.append('    <var name="faceMeltSpeed"/>')
        lines.append('    <var name="faceMeltingThickness"/>')


It's a little odd to have faceMeltingThickness, but not calvingThickness. Do we intend to eventually create a face-melting flux variable analogous to avgCalvingFlux?

Thanks for pointing that out - we do actually have the time-averaged facemelting flux already available. I've pushed a new commit that replaces faceMeltingThickness with that.

matthewhoffman · 2025-07-03T19:08:29Z

@trhille , I've pushed a commit further adjusting the MALI history file contents to include your suggestion, remove a few fields I don't think are necessary, and add the 2d TF field that facemelting uses.

In principle, we don't need either of the TF fields included right now, as we have the fields for the resulting fluxes calculated from them. At the same time, it might be useful to have the full 3d TF field that comes from MPAS-Ocean for visualization and possible debugging. But that field alone is bigger than the rest of the contents of this file (by a few times) because of the (standard) 30 vertical levels. So I am taking the approach of leaving it out by default, and one can modify their streams file if they want to have it available. That is probably something we will want to do frequently while evaluating ocn/glc TF coupling, but I want to avoid the default MALI hist file being too big.

If you (or others) have a different preference, let me know.

stephenprice

No suggested changes but a few comments to look at (responses on slack are fine too).

stephenprice · 2025-07-08T21:29:37Z

driver-mct/cime_config/config_component_e3sm.xml

      <value compset="_DATM.*_CICE.*_DOCN">24</value>
      <value compset="_DATM.*_DOCN%US20">24</value>
      <value compset="_MPASO">48</value>
      <value compset="_DLND.*_MALI">1</value>


If the base coupling interval for MALI is set to 1 year, does the above value for "_DLND.*_MALI" compsets also need to be updated? Even if only for consistency? Or, does this compset really only couple w/ the atmos. once per year?

stephenprice · 2025-07-08T21:33:29Z

driver-mct/cime_config/config_component_e3sm.xml

      <value compset="_DATM.*_SLND.*MPAS" grid="oi%oQU240">12</value>
      <value compset="_DATM.*_SLND.*MPAS" grid="oi%oQU240wLI">12</value>
+      <value compset="_DATM.*_SLND.*MPAS.*_MALI" grid="oi%oQU240wLI">4380</value>
      <value compset="_DATM.*_SLND.*MPAS" grid="oi%oQU120">24</value>


This one ("_DATM.*_SLND.MPAS._MALI" grid="oi%oQU240wLI") and line 523 below are diff. (default coupling once every 2 hrs versus once every 30 mins) on purpose because of the low-res ocean mesh?

stephenprice · 2025-07-08T21:40:19Z

components/mpas-albany-landice/bld/namelist_files/namelist_defaults_mali.xml

 <config_rk_order>2</config_rk_order>
 <config_rk3_stages>3</config_rk3_stages>
-<config_adaptive_timestep>.false.</config_adaptive_timestep>
+<config_adaptive_timestep>.true.</config_adaptive_timestep>


I'm wary of turning this on by default without pretty extensive testing first, but I guess that can be part of this PR.

trhille · 2025-07-09T02:54:50Z

components/mpas-albany-landice/cime_config/buildnml

        lines.append('            type="output"')
        lines.append('            filename_template="{}.mali.hist.$Y-$M-$D_$S.nc"'.format(casename))
-        lines.append('            filename_interval="0001-00-00_00:00:00"')
+        lines.append('            filename_interval="0000-01-00_00:00:00"')


Do we really want to write a new output file every month?

That's how the ocean and sea ice history files are. But they don't include the mesh in every file. So maybe we should remove that.

After discussion, I pushed an update to go back to annual history files that include mesh fields and have daily output. A shake-up to that can wait for a future PR after we have more simulations tested.

I think I mis-read this in my review to be 1 day. Given the relatively shorter timescales we are working at right now, and the (default) monthly history files from other components, I think 1 month is a decent compromise.

Include all time-averaged coupling fields, remove temperature field because it is 3d and will be in restart files, remove vertical coordinate fields (they are tiny but rarely used and included in input and restart files), add ismip6_2dThermalForcing for reference

matthewhoffman · 2025-07-09T03:13:18Z

I separated out the change to the coupling interval into a new PR: #7496

trhille

Thanks for taking care of this, @matthewhoffman!

Adjust MALI SMB in G-cases and update hist fields This branch updates a number of details about MALI coupling in preparation of production runs: * When ELM is not active (e.g. G-cases), MALI will now use a data SMB from the MALI initial file * The data volume of MALI history files is vastly reduced [NML] only for configurations with MALI [non-BFB] only for configurations with MALI

jonbob · 2025-07-10T20:40:04Z

Testing shows expected NML DIFFs for e3sm_landice_developer using gnu on chrysalis plus:

ERS.ne30pg2_r05_IcoswISC30E3r5_gis20.IGELM_MLI.chrysalis_gnu.mali-gis20km fieldlist DIFF
ERS_Ld5.TL319_oQU240wLI_ais8to30.MPAS_FOLISIO_JRA1p5.chrysalis_gnu.mpaso-jra_1958 baseline DIFF
ERS_Ld5.TL319_oQU240wLI_gis4to40.MPAS_FOLISIO_JRA1p5.chrysalis_gnu.mpaso-jra_1958 baseline DIFF
SMS.ne30_oECv3_gis.IGELM_MLI.chrysalis_gnu.elm-extrasnowlayers BFB
SMS.ne30pg2_r05_IcoswISC30E3r5_gis20.BGWCYCL1850.chrysalis_gnu.allactive-gis20km fieldlist DIFF
SMS.ne30pg2_r05_IcoswISC30E3r5_gis20.IGELM_MLI.chrysalis_gnu.mali-gis20km fieldlist DIFF
SMS.ne30pg2_r05_IcoswISC30E3r5_gis4to40.BGWCYCL1850.chrysalis_gnu.allactive-gis20km fieldlist DIFF

Passes:

SMS_D_Ld1.ne30pg2_r05_IcoswISC30E3r5.WCYCL1850.chrysalis_intel.allactive-wcprod

merged to next

jonbob

Approved based on visual inspection and testing

jonbob · 2025-07-11T18:54:04Z

merged to master and expected NML and baseline DIFFs blessed except:

pm-cpu integration -- not reported yet
pm-cpu developer -- not reported yet
pm-cpu landice-developer -- not reported yet
mappy developer -- some tests failed due to download issues with new file

matthewhoffman · 2025-07-11T19:10:53Z

@jonbob , FYI, I fixed the download issue after a head's up from @amametjanov . I had placed the new input files in the local mirror on Chrysalis instead of the primary public inputdata location. That has now been corrected.

jonbob · 2025-07-11T19:18:50Z

Thanks @matthewhoffman. And that makes sense -- I went and checked the inputdata repo and the new files were there, so I was a little confused. By the way, you might want to cleanup the permissions, which currently look like:

-rwxr-xr-x 1 ac.mhoffman E3SM 154902172 Jul 11 13:32 ais_8to30km.20250625.nc

matthewhoffman · 2025-07-11T19:25:25Z

Thanks, I took care of that too now.

jonbob · 2025-07-15T15:25:11Z

blessed on mappy and pm-cpu

matthewhoffman added 3 commits June 25, 2025 17:20

Update AIS 4km and 8km meshes to include SMB field

c94f643

Update MALI output stream to have minimally required fields

2ee7f50

Diagnostic/debugging fields that had been added previously have been removed to reduce file size. Output frequency has been reduced to 5 days.

matthewhoffman requested review from stephenprice, jonbob and trhille June 25, 2025 23:48

matthewhoffman assigned jonbob Jun 25, 2025

matthewhoffman added mpas-albany-landice NML Stealth PR has feature which, if turned on, could change climate. fka FCC labels Jun 25, 2025

matthewhoffman force-pushed the matthewhoffman/mali/smb-data-option branch 2 times, most recently from 70b890c to 5e7fc9f Compare June 26, 2025 12:32

trhille reviewed Jun 30, 2025

View reviewed changes

trhille reviewed Jul 1, 2025

View reviewed changes

stephenprice approved these changes Jul 8, 2025

View reviewed changes

trhille reviewed Jul 9, 2025

View reviewed changes

matthewhoffman force-pushed the matthewhoffman/mali/smb-data-option branch from 487dd4c to a69161e Compare July 9, 2025 03:10

Switch back to annual MALI history files with daily output

5e8ed20

matthewhoffman changed the title ~~MALI coupling updates for SMB, cpl interval, and hist fields~~ Adjust MALI SMB in G-cases and update hist fields Jul 9, 2025

trhille approved these changes Jul 9, 2025

View reviewed changes

jonbob and others added 2 commits July 10, 2025 10:54

Make bld files consistent with Registry

e701c0f

Fix typos preventing proper parsing; add error check

ab3b0bb

jonbob added non-BFB PR makes roundoff changes to answers. and removed Stealth PR has feature which, if turned on, could change climate. fka FCC labels Jul 10, 2025

jonbob approved these changes Jul 10, 2025

View reviewed changes

jonbob merged commit fa2a3e4 into master Jul 11, 2025
3 checks passed

jonbob deleted the matthewhoffman/mali/smb-data-option branch July 11, 2025 18:34

Adjust MALI SMB in G-cases and update hist fields #7463

Adjust MALI SMB in G-cases and update hist fields #7463

Uh oh!

Conversation

matthewhoffman commented Jun 25, 2025 • edited by jonbob Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

trhille left a comment

Choose a reason for hiding this comment

Uh oh!

matthewhoffman commented Jul 1, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

matthewhoffman commented Jul 3, 2025

Uh oh!

stephenprice 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!

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!

matthewhoffman commented Jul 9, 2025

Uh oh!

trhille left a comment

Choose a reason for hiding this comment

Uh oh!

jonbob commented Jul 10, 2025

Uh oh!

jonbob left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jonbob commented Jul 11, 2025

Uh oh!

matthewhoffman commented Jul 11, 2025

Uh oh!

jonbob commented Jul 11, 2025

Uh oh!

matthewhoffman commented Jul 11, 2025

Uh oh!

jonbob commented Jul 15, 2025

Uh oh!

Uh oh!

matthewhoffman commented Jun 25, 2025 •

edited by jonbob

Loading