Add parametrized tidal mixing effects to ice shelf basal melt #6306

irenavankova · 2024-03-26T20:09:41Z

An implementation of prescribed tidal effect on ice shelf basal melting following (Jourdain et al., 2019):

https://doi.org/10.1016/j.ocemod.2018.11.001

The tidal flow speeds used to produce a forcing data file are calculated from the CATS 2008 (an update of Padman et al., 2002). When the file is available and compset TMIX on, the Jourdain formula is used. The main change is that what used to be a constant config_land_ice_flux_rms_tidal_velocity is now a spatially variable (horizontal) field.

Related Ocean Discussion is here:

E3SM-Ocean-Discussion#76

[NML]
[BFB]

…ll ISMF and add TMIX when PISMF

…ompass

xylar · 2024-03-28T13:03:28Z

@jonbob, I've marked this as "Stealth" but I'm not sure if that is correct. Let me know what you think. I'm going to run an Icos G-case with this configuration for 10 years and see what we get. I'll compare with the same configuration without tidal mixing. Hopefully, that's sufficient as a test of this capability as a stealth feature. I'm also going to figure out where to add an appropriate test.

xylar · 2024-03-28T15:12:54Z

@jonbob, it's certainly conceivable that we could make this the default behavior for Polar runs. That's the sense in which I was thinking of it as a stealth feature.

jonbob · 2024-03-28T15:23:44Z

I think in our current definition, this is a separate configuration and we can remove the stealth label and add a new test to extra coverage. But it's still important to document what impact it has as part of the PR

xylar · 2024-03-28T17:03:09Z

@irenavankova, I hope you don't mind that I pushed a test.

xylar · 2024-03-28T17:04:12Z

cime_config/tests.py

@@ -239,6 +239,7 @@
        "tests" : (
            "ERS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF.mpaso-jra_1958",
            "PEM_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF.mpaso-jra_1958",
+            "SMS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF-TMIX.mpaso-jra_1958",


@jonbob, does this look right?

I was able to run this on Chrysalis by excluding mpaso-jra_1958 but didn't know how to test that part.

I can do that when I work on this PR -- there is a bit of a trick to specifying tests that use testdefs

Okay, thanks!

The easiest way is to add machine_compiler to the test name.
./creat_test SMS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF-TMIX.chrysalis_intel.mpaso-jra_1958

Thanks @rljacob!

xylar

I ran 2 Icos30 G-cases for 10 years, one with and one without tidal mixing:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.xylar/analysis/tmix_vs_no_tmix_20240328.GMPAS-JRA1p5-DIB-PISMF.TL319_IcoswISC30E3r5.chrysalis/yrs-01-10/ocean/index.html

The results are as expected: only the melt rates below ice shelves and the adjacent ocean properties are affected, the rest of the climate is nearly identical.

Interestingly, there is substantially less melting overall and particularly in the Amundsen and Bellingshausen regions:

This likely suggests that other coefficients related to melting will need to be retuned in coordinating with adopting this parameterization in Polar simulations, but that is work for a follow-up PR.

xylar · 2024-04-04T14:05:09Z

cime_config/tests.py

@@ -239,6 +239,7 @@
        "tests" : (
            "ERS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF.mpaso-jra_1958",
            "PEM_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF.mpaso-jra_1958",
+            "SMS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF-TMIX.mpaso-jra_1958",


Thanks @rljacob!

jonbob · 2024-05-15T19:53:48Z

@irenavankova -- I ran the scripts that make the bld files consistent with Registry and they came up with some differences in namelist_definition_mpaso.xml:

--- a/components/mpas-ocean/bld/namelist_files/namelist_definition_mpaso.xml
+++ b/components/mpas-ocean/bld/namelist_files/namelist_definition_mpaso.xml
@@ -1608,25 +1608,25 @@ Default: Defined in namelist_defaults.xml
 
 <entry id="config_land_ice_flux_tidal_Jourdain_alpha" type="real"
        category="land_ice_fluxes" group="land_ice_fluxes">
-Constant in parameterization of tidal velocity used in computing the sub-ice-shelf friction velocity
+alpha from parameterization of tidal velocity by Jourdain et al., 2019 - Equation 9, alpha = 0.777 when TMIX compset on
 
-Valid values: Any positive real number
+Valid values: Any non-negative real number
 Default: Defined in namelist_defaults.xml
 </entry>
 
 <entry id="config_land_ice_flux_tidal_Jourdain_A0" type="real"
        category="land_ice_fluxes" group="land_ice_fluxes">
-Constant in parameterization of tidal velocity used in computing the sub-ice-shelf friction velocity
+A0 from parameterization of tidal velocity by Jourdain et al., 2019 - Equation 9, A0 = 0.656 when TMIX compset on
 
-Valid values: Any positive real number
+Valid values: Any non-negative real number
 Default: Defined in namelist_defaults.xml
 </entry>
 
 <entry id="config_land_ice_flux_tidal_Jourdain_U0" type="real"
        category="land_ice_fluxes" group="land_ice_fluxes">
-Constant in parameterization of tidal velocity used in computing the sub-ice-shelf friction velocity
+U0 from parameterization of tidal velocity by Jourdain et al., 2019 - Equation 9, U0 = 0.003 m/s when TMIX compset on
 
-Valid values: Any positive real number
+Valid values: Any non-negative real number
 Default: Defined in namelist_defaults.xml
 </entry>

Can you please help me resolve the differences?

…to next (PR #6306) Add parametrized tidal mixing effects to ice shelf basal melt An implementation of prescribed tidal effect on ice shelf basal melting following (Jourdain et al., 2019). The tidal flow speeds used to produce a forcing data file are calculated from the CATS 2008 (an update of Padman et al., 2002). When the file is available and compset TMIX on, the Jourdain formula is used. The main change is that what used to be a constant config_land_ice_flux_rms_tidal_velocity is now a spatially variable (horizontal) field. This implmentation is accessed via a compset option to mpaso [NML] [BFB]

jonbob · 2024-05-28T21:57:34Z

passes:

ERS_Ld5.T62_oQU120.CMPASO-NYF.chrysalis_intel
SMS_D_Ld1.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF.chrysalis_intel.mpaso-jra_1958

with expected NML DIFFs. Also passes:

SMS_P480_Ld5.TL319_IcoswISC30E3r5.GMPAS-JRA1p5-DIB-PISMF-TMIX.chrysalis_intel.mpaso-jra_1958

which is the new test introduced by this PR

merged to next

jonbob · 2024-05-28T22:03:56Z

verified that all new files are available on the inputdata repo and have correct permissions

jonbob · 2024-05-29T19:32:30Z

@irenavankova and @xylar -- it looks like this PR was non-BFB on pm-cpu using the gnu compiler. From looking at the changes, it can only be order-of-operations level though of course with the model any little change becomes larger. How do you want to handle this? It is only the one machine and one compiler. I tested it on chrysalis with gnu and did not see any DIFFs and overnight testing with intel passed

xylar · 2024-05-29T19:42:27Z

Oh, that's odd. I guess since you merged this, you think it's okay. I guess we'll need to update baselines for pm-cpu? I think that's likely okay since it was BFB on other machines.

xylar · 2024-05-29T19:46:55Z

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

-                                   + config_land_ice_flux_rms_tidal_velocity**2))
+                                   + config_land_ice_flux_tidal_Jourdain_alpha**2 * &
+                                   (config_land_ice_flux_tidal_Jourdain_A0 * velocityTidalRMS(iCell) + &
+                                   config_land_ice_flux_tidal_Jourdain_U0)**2))


This is the only term I could imaging causing the non-BFB results. It's hard for me to understand how a reasonable compiler doesn't produce BFB results when config_land_ice_flux_tidal_Jourdain_alpha = 1.0, config_land_ice_flux_tidal_Jourdain_A0 = 0.0 and config_land_ice_flux_tidal_Jourdain_U0 is identical to the old config_land_ice_flux_rms_tidal_velocity but I guess gnu on pm-cpu might just not be reasonable...

That was my thought as well -- but I did want to point it out just in case

xylar · 2024-05-29T19:47:29Z

Thanks for the help with this, @jonbob and thanks for working so hard on it, @irenavankova!

jonbob · 2024-05-29T19:49:58Z

I did go ahead and merge it because it passed on our production platforms and compilers -- it seems like a one-off. I can just bless those DIFFs, or if you want we can rework the code section that actually does the work and try to get it BFB on pm-cpu_gnu? I'm guessing it would be possible using the new configs that were added? Or decrease the level of optimization for the affected file?

xylar · 2024-05-29T19:52:07Z

I think we should just bless the difference. The only simulations where this is going to have any impact at all (apart from ensemble variability) are those with prognostic ice-shelf melting and we don't have any production simulations going like that right now.

jonbob · 2024-05-29T21:59:54Z

OK, I blessed the DIFFs for the PISMF tests on pm-cpu. There is another new DIFF that doesn't make sense to me:

ERS_Ld5.TL319_oQU240wLI_ais20.MPAS_LISIO_JRA1p5.pm-cpu_gnu.mpaso-ocn_glcshelf

xylar · 2024-05-30T08:32:21Z

That's also PISMF, but through the coupler so it makes sense to me. The friction velocity, modified by this PR, will be used there as well.

jonbob · 2024-05-30T15:12:38Z

Thanks @xylar -- I'll bless that test as well

This has been changed in E3SM-Project/E3SM#6306

This merge updates the E3SM-Project submodule from [31e0924](https://github.com/E3SM-Project/E3SM/tree/31e0924) to [8939709](https://github.com/E3SM-Project/E3SM/tree/8939709). This update includes the following MPAS-Ocean and MPAS-Frameworks PRs (check mark indicates bit-for-bit with previous PR in the list): - [ ] (ocn) E3SM-Project/E3SM#6263 - [ ] (ocn) E3SM-Project/E3SM#6310 - [ ] (fwk) E3SM-Project/E3SM#6427 - [ ] (ocn) E3SM-Project/E3SM#6397 - [ ] (ocn) E3SM-Project/E3SM#6306 - [ ] (ocn) E3SM-Project/E3SM#6454

irenavankova and others added 24 commits March 26, 2024 16:08

replace with new name for landIceTopDragCoeff

5159c84

Resolved - keep ocn_ismf and tidal cahnges in buildnml

48aa754

Resolved - keep both date change and tidal file name change

350108e

typo fix

71aab2f

Resolved: again replace with new name for landIceTopDragCoeff

65c3a82

resolve

0aa76c3

Resolved: kept both

cb79214

add jourdain constants to namelist_definition_mpaso.xml

66f1bcc

more forgotten namelist stuff

976037a

ismf to pismf

0c8044f

more TMIX compsets - NYF and IAF

7627d8f

Resolved - keep both date change and tidal file name change

d9eef97

tell mpaso to read tidal file at init

e27358d

>

ae11914

remove > on jb backup

6235dae

added compsets for C cases

6a74f0e

corrected compsets for C cases

3ecba56

add Jourdain defaults for TMIX, remove redundant variables, correct a…

09cac5b

…ll ISMF and add TMIX when PISMF

resolve

c1c717f

add velocityTidalRMS to monthly output

a177ec2

Add CRYO1850-TMIX and CRYO1950-TMIX compsets

1fe8bbd

Updated tidal mixing files and filenames that were regenerated with c…

b50d180

…ompass

remove short names compsets

70b4d72

add tidal files for FRIS meshes

d4fb283

irenavankova force-pushed the ocn/add-tidal-mixing-to-ice-shelf-melt branch from 4d22b59 to d4fb283 Compare March 26, 2024 22:40

xylar self-requested a review March 27, 2024 14:50

xylar assigned jonbob Mar 27, 2024

xylar added mpas-ocean Stealth PR has feature which, if turned on, could change climate. fka FCC labels Mar 27, 2024

xylar added BFB PR leaves answers BFB and removed Stealth PR has feature which, if turned on, could change climate. fka FCC labels Mar 28, 2024

Add a test with the TMIX compset

203a45e

xylar reviewed Mar 28, 2024

View reviewed changes

xylar approved these changes Apr 4, 2024

View reviewed changes

jonbob added the NML label May 2, 2024

Irena Vankova and others added 2 commits May 28, 2024 12:16

remove new registry variables from namelist_definition_mpaso.xml

1ddd297

Update bld files to match Registry

bd446b2

jonbob merged commit 24bdc2d into E3SM-Project:master May 29, 2024
2 checks passed

xylar reviewed May 29, 2024

View reviewed changes

xylar added a commit to xylar/compass that referenced this pull request Jun 4, 2024

Update the namelist option for tidal velocity

ce05afe

This has been changed in E3SM-Project/E3SM#6306

xylar mentioned this pull request Jun 4, 2024

Update the namelist option for tidal velocity MPAS-Dev/compass#826

Merged

1 task

xylar mentioned this pull request Jun 4, 2024

Update E3SM-Project submodule MPAS-Dev/compass#827

Merged

6 tasks

xylar mentioned this pull request Jul 25, 2024

Add subglacial discharge at grounding line #6508

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add parametrized tidal mixing effects to ice shelf basal melt #6306

Add parametrized tidal mixing effects to ice shelf basal melt #6306

irenavankova commented Mar 26, 2024 •

edited by jonbob

Loading

xylar commented Mar 28, 2024 •

edited

Loading

xylar commented Mar 28, 2024

jonbob commented Mar 28, 2024

xylar commented Mar 28, 2024

xylar Mar 28, 2024

jonbob Mar 28, 2024

xylar Mar 28, 2024

rljacob Mar 28, 2024

xylar Apr 4, 2024

xylar left a comment •

edited

Loading

xylar Apr 4, 2024

jonbob commented May 15, 2024

jonbob commented May 28, 2024

jonbob commented May 28, 2024

jonbob commented May 29, 2024

xylar commented May 29, 2024

xylar May 29, 2024

jonbob May 29, 2024

xylar commented May 29, 2024

jonbob commented May 29, 2024

xylar commented May 29, 2024

jonbob commented May 29, 2024

xylar commented May 30, 2024 •

edited

Loading

jonbob commented May 30, 2024

Add parametrized tidal mixing effects to ice shelf basal melt #6306

Add parametrized tidal mixing effects to ice shelf basal melt #6306

Conversation

irenavankova commented Mar 26, 2024 • edited by jonbob Loading

xylar commented Mar 28, 2024 • edited Loading

xylar commented Mar 28, 2024

jonbob commented Mar 28, 2024

xylar commented Mar 28, 2024

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

xylar left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jonbob commented May 15, 2024

jonbob commented May 28, 2024

jonbob commented May 28, 2024

jonbob commented May 29, 2024

xylar commented May 29, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

xylar commented May 29, 2024

jonbob commented May 29, 2024

xylar commented May 29, 2024

jonbob commented May 29, 2024

xylar commented May 30, 2024 • edited Loading

jonbob commented May 30, 2024

irenavankova commented Mar 26, 2024 •

edited by jonbob

Loading

xylar commented Mar 28, 2024 •

edited

Loading

xylar left a comment •

edited

Loading

xylar commented May 30, 2024 •

edited

Loading