Updates for CMIP7 by dabail10 · Pull Request #1066 · CICE-Consortium/CICE

dabail10 · 2025-11-13T23:51:11Z

For detailed information about submitting Pull Requests (PRs) to the CICE-Consortium,
please refer to: https://github.com/CICE-Consortium/About-Us/wiki/Resource-Index#information-for-developers

PR checklist

This PR should address issue #1048 and #1038.

I have updated all of the variable names, long names, and units to correspond to the CMIP7 data request.
I have added new variables requested in the CMIP7 data request.
I have added documentation about the CMIP6 to CMIP7 update.
I have simplified the accumulation of some fields where possible and added prognostic sea ice density.
I have added accumulation of variables relative to aice_init or aice.
Bug fix for flwout (sifllwutop) where aice_init = 0, but aice > 0.
Bug fix for shortwave abosrbed and albedo computation (more coming later)
Bug fix: Some variables that were scaled by aice, should be multiplied by aice (not aice_init) to get the _ai quantities. This includes fswabs, fsens, flat, ...
Remove f_CMIP flag and add set_nml.cmip option instead.
Added comment field for SIMIP variables that uses part of the description field in the CMIP data request table.
Add long_name field for time_bounds, lon?_bounds and lat?_bounds. Issue time_bounds, lat?_bounds, lon?_bounds attributes #1057
Partly address aice versus aice_init aice vs. aice/aice_init factor in ice_history #1033
Partial fix for albedo variables [albedo]_ai history variables over 100% #1051
Also addresses issue Some CMIP variables are computed using a mix of U and T quantities #904
Updated test results

Closes #904
Closes #1033
Closes #1038
Closes #1048
Closes #1057

variables for CMIP7. - Add new variables that were not there before - Change names for some of the variables - Add an f_CICE flag to be able to turn off CICE duplicate variables - Added more variables to the daily stream

anton-seaice

Thanks @dabail10 - I will give this a proper review next week

I dropped a couple of comments - i am not sure about mask_ice_free_points=.true. for extensive variables

In general the cell_methods in the data request are a bit of a mess now, there are several methods more then there used to be and some errors. If we want to be setting them directly in CICE it will take some work (e.g. its more than just one for intensive and one for extensive!)

cicecore/cicedyn/analysis/ice_history.F90

…iate

- add a few more SIMIP variables (sisndmasswind, ...) - remove sialb as this is no longer in the CMIP7 table - start some work on the aice versus aice_init thing

apcraig · 2025-11-17T16:03:40Z

Documentation will need to be updated. I'm also confused about the f_CICE and f_CMIP implementation. Why not just set all the CMIP f_variables=f_CMIP? And are all or some of the CMIP variables also exposed independently with f_ namelist and which takes priority and/or should there be an ability to "merge" the independent setting and the f_CMIP value instead of overwriting? If the implementation introduces limitations on the settings, we should make that clear in the documentation.

dabail10 · 2025-11-17T16:10:16Z

Documentation will need to be updated. I'm also confused about the f_CICE and f_CMIP implementation. Why not just set all the CMIP f_variables=f_CMIP? And are all or some of the CMIP variables also exposed independently with f_ namelist and which takes priority and/or should there be an ability to "merge" the independent setting and the f_CMIP value instead of overwriting? If the implementation introduces limitations on the settings, we should make that clear in the documentation.

This is a good question. This was the way I thought I would deal with some variables only on the monthly stream and some variables on the daily stream with the one flag, f_CMIP. I'm open to other ideas.

- change avg_ice_present to be a character array - add capability to average when aice_init > 0 - add long_name (title) for SIMIP variables from the CMIP7 table - update some ice density calculations to use the mushy-layer brine+ice density

dabail10 · 2025-11-19T21:31:21Z

Just added a whole bunch of changes. I think this is getting pretty close. I have a one-month run with the output here. Outstanding issues:

I have added the CF long_name from the CMIP7 table. Do we want more in the comment field here?
I'm still not happy with fllwutop. I am getting some excessively large values here at points with very tiny ice concentration (1.0e-8). Maybe this is just what it it is. @eclare108213 and I have been iterating on the math.
Still working on the siitd fields.

anton-seaice · 2025-11-20T23:42:18Z

should there be an ability to "merge" the independent setting and the f_CMIP value instead of overwriting?

This would be great but i can't think of how to implement it within the current framework. (e.g. something like, if the variable is in the ice_in file, then override with the f_CMIP or f_CICE value)

My view is that all diagnostics should just be off by default, and then let the user set the ones they want. That would break a lot of configurations though, so takes some care if others agree!

anton-seaice

Thanks @dabail10

I didn't go through everything, but the main themes are:

in some places this used fixed values of rhoi, rhos and in others they come from icepack. Is this correct ?
Melt processes should always be less than 0 in cmip diagnostics
There's lots of loops that aren't needed once if aice>puny condition is removed, they can just be call accum_hist_field(n_..., iblk, ...(:,:,iblk), a2D)
Can all the _ai variables be saved before scale_fluxes is called rather than recalculating them

cicecore/cicedyn/analysis/ice_history.F90

dabail10 · 2025-11-21T20:02:04Z

Just added a whole bunch of changes. I think this is getting pretty close. I have a one-month run with the output here. Outstanding issues:

I have added the CF long_name from the CMIP7 table. Do we want more in the comment field here?

I'm still not happy with fllwutop. I am getting some excessively large values here at points with very tiny ice concentration (1.0e-8). Maybe this is just what it it is. @eclare108213 and I have been iterating on the math.

Still working on the siitd fields.

I have fixed the fllwutop field and the siitd fields. Thoughts on what to put in the comments field? The description from the table? the CF standard_name?

dabail10 · 2025-11-21T20:25:07Z

Thanks @dabail10

I didn't go through everything, but the main themes are:

in some places this used fixed values of rhoi, rhos and in others they come from icepack. Is this correct ?

Melt processes should always be less than 0 in cmip diagnostics

There's lots of loops that aren't needed once if aice>puny condition is removed, they can just be call accum_hist_field(n_..., iblk, ...(:,:,iblk), a2D)

Can all the _ai variables be saved before scale_fluxes is called rather than recalculating them

I have fixed all occurrences of rhoi as a constant to use the brine + ice density.
changed the melt processes to negative
removed the unnecessary accumulation in worka where appropriate.

dabail10 · 2025-11-21T20:28:12Z

Thanks @dabail10

I didn't go through everything, but the main themes are:

in some places this used fixed values of rhoi, rhos and in others they come from icepack. Is this correct ?

Melt processes should always be less than 0 in cmip diagnostics

There's lots of loops that aren't needed once if aice>puny condition is removed, they can just be call accum_hist_field(n_..., iblk, ...(:,:,iblk), a2D)

Can all the _ai variables be saved before scale_fluxes is called rather than recalculating them

On point 4, this is a larger issue. Ultimately we want to get rid of the _ai quantities. However, that is for a future PR.

anton-seaice · 2025-11-24T02:58:11Z

Thanks for your work @dabail10 !

Can this calculation be a subroutine or done once in the file and stored, rather than repeated through the code ? :

       if (ktherm == 2) then
                 rho_ocn = icepack_mushy_density_brine(sss(i,j,iblk))
                 rho_ice = c0
                 salt_ice = c0
                 do k = 1, nzilyr
                    Tice = icepack_mushy_temperature_mush(trcr(i,j,nt_qice+k-1,iblk),trcr(i,j,nt_sice+k-1,iblk))
                    Sbr = trcr(i,j,nt_sice+k-1,iblk)
                    salt_ice = salt_ice + Sbr / real(nzilyr,kind=dbl_kind)
                    phi = icepack_mushy_liquid_fraction(Tice,Sbr)
                    rhob = icepack_mushy_density_brine(Sbr)
                    rho_ice = rho_ice + min(phi*rhob+(c1-phi)*rhoi,rho_ocn)
                 enddo
                 rho_ice = rho_ice / real(nzilyr,kind=dbl_kind)

anton-seaice · 2025-11-24T03:04:18Z

Thoughts on what to put in the comments field? The description from the table? the CF standard_name?

Description field can be quite long ? It's a bit of an annoying thing to hard-code .

I don't have a strong view

dabail10 · 2025-11-24T16:55:58Z

Thanks for your work @dabail10 !

Can this calculation be a subroutine or done once in the file and stored, rather than repeated through the code ? :

       if (ktherm == 2) then
                 rho_ocn = icepack_mushy_density_brine(sss(i,j,iblk))
                 rho_ice = c0
                 salt_ice = c0
                 do k = 1, nzilyr
                    Tice = icepack_mushy_temperature_mush(trcr(i,j,nt_qice+k-1,iblk),trcr(i,j,nt_sice+k-1,iblk))
                    Sbr = trcr(i,j,nt_sice+k-1,iblk)
                    salt_ice = salt_ice + Sbr / real(nzilyr,kind=dbl_kind)
                    phi = icepack_mushy_liquid_fraction(Tice,Sbr)
                    rhob = icepack_mushy_density_brine(Sbr)
                    rho_ice = rho_ice + min(phi*rhob+(c1-phi)*rhoi,rho_ocn)
                 enddo
                 rho_ice = rho_ice / real(nzilyr,kind=dbl_kind)

You know, I bet there is a mushy function for this already ...

cicecore/cicedyn/analysis/ice_history_mechred.F90

cicecore/cicedyn/analysis/ice_history_pond.F90

cicecore/cicedyn/general/ice_flux.F90

doc/source/user_guide/ug_implementation.rst

cicecore/cicedyn/analysis/ice_history.F90

anton-seaice · 2026-01-11T22:13:57Z

I still think there is something inconsistent with snowfrac / sisnconc:

Snowfrac is per grid cell area

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 686 to 689 in d0455f6

    
           call define_hist_field(n_snowfrac,"snowfrac","1",tstr2D, tcstr, & 
        
               "grid cell mean snow fraction",                             & 
        
               "snow fraction per unit grid cell area", c1, c0,            & 
        
               ns1, f_snowfrac)

Sisnconc is per sea ice area (and extensive)

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 1780 to 1784 in d0455f6

    
           call define_hist_field(n_sisnconc,"sisnconc","%",tstr2D, tcstr,  & 
        
               "snow area percentage",                                      & 
        
               "percentage of the sea-ice surface that is covered by snow", & 
        
               c100, c0,                                                    & 
        
               ns1, f_sisnconc, avg_ice_present='none', mask_ice_free_points=.false.)

But they are calculated the same:

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 2401 to 2402 in d0455f6

    
           if (f_snowfrac(1:1) /= 'x') & 
        
               call accum_hist_field(n_snowfrac, iblk, snowfrac(:,:,iblk), a2D)

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 2818 to 2820 in d0455f6

    
           if (f_sisnconc(1:1) /= 'x') then 
        
             call accum_hist_field(n_sisnconc, iblk, snowfrac(:,:,iblk), a2D) 
        
           endif

dabail10 · 2026-01-12T16:17:17Z

I still think there is something inconsistent with snowfrac / sisnconc:

Snowfrac is per grid cell area

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 686 to 689 in d0455f6

call define_hist_field(n_snowfrac,"snowfrac","1",tstr2D, tcstr, &

"grid cell mean snow fraction", &

"snow fraction per unit grid cell area", c1, c0, &

ns1, f_snowfrac)

Sisnconc is per sea ice area (and extensive)

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 1780 to 1784 in d0455f6

call define_hist_field(n_sisnconc,"sisnconc","%",tstr2D, tcstr, &

"snow area percentage", &

"percentage of the sea-ice surface that is covered by snow", &

c100, c0, &

ns1, f_sisnconc, avg_ice_present='none', mask_ice_free_points=.false.)

But they are calculated the same:

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 2401 to 2402 in d0455f6

if (f_snowfrac(1:1) /= 'x') &

call accum_hist_field(n_snowfrac, iblk, snowfrac(:,:,iblk), a2D)

CICE/cicecore/cicedyn/analysis/ice_history.F90

Lines 2818 to 2820 in d0455f6

if (f_sisnconc(1:1) /= 'x') then

call accum_hist_field(n_sisnconc, iblk, snowfrac(:,:,iblk), a2D)

endif

Ah, I think I understand. The variable snowfrac is only the fraction on the sea ice. This comment is incorrect.

dabail10 · 2026-01-12T16:21:29Z

In icepack we have the following:

            snowfracn(n) = fsn ! for history

Where fsn is the snow fraction on the sea ice per category and this is aggregated like apeff_ai.

                  apeff_ai(i,j,iblk) = apeff_ai(i,j,iblk) &
                       + apeffn(i,j,n,iblk)*aicen(i,j,n,iblk)
                  snowfrac(i,j,iblk) = snowfrac(i,j,iblk) &
                       + snowfracn(i,j,n,iblk)*aicen(i,j,n,iblk)

I will modify the comment.

eclare108213

After an ENORMOUS amount of work by @dabail10 following up on tons of critiques, I think this is ready to go! Just one more very minor request from me. Thank you @dabail10!

cicecore/cicedyn/analysis/ice_history.F90

dabail10 · 2026-01-15T20:54:13Z

Getting a lot of compare fails.

https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#b653260a0f28509a80598e07635d01d0c56ce434

apcraig · 2026-01-15T22:06:58Z

Getting a lot of compare fails.

https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#b653260a0f28509a80598e07635d01d0c56ce434

It looks like your branch was created off a version of CICE on Nov 12. There was a commit to main on Nov 26 that changed answers. You either need to compare to the version of CICE you branched from or update your branch to something more recent.

If this is ready, I can run some tests again. What I do is checkout CICE main then merge the PR branch into my sandbox. This duplicates what will happen when we merge the PR and also ensures testing is done relative the latest main version. You could create a sandbox like this was for testing, you do not necessarily have to update your branch to the latest CICE main.

Just to check again, do we think modifications are complete at this point. If so, I will test too.

dabail10 · 2026-01-15T23:29:33Z

Oh shoot. I forgot about that. I will update the base branch.

dabail10 · 2026-01-15T23:30:50Z

There it was a clean update with no conflict. I will rerun my tests.

dabail10 · 2026-01-16T00:08:28Z

I am now getting some compile errors. I guess the merge was not as clean as I hoped.

/glade/work/dbailey/CICE_ESCOMP/cicecore/cicedyn/general/ice_step_mod.F90(685): error #6627: This is an actual argument keyword name, and not a dummy argument name. [WAVE_HEIGHT_TYPE_OUT]
wave_height_type_out=wave_height_type)
------------------------------------^
/glade/work/dbailey/CICE_ESCOMP/cicecore/cicedyn/general/ice_step_mod.F90(948): error #6627: This is an actual argument keyword name, and not a dummy argument name. [WAVE_HEIGHT]
wave_height = wave_sig_ht (i,j, iblk))
-------------------------------------------^

I didn't update icepack appropriately.

dabail10 · 2026-01-16T00:59:08Z

Much better. 3 PEND and 2 FAIL. I think this is expected for the hdf5/cdf tests?

https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#3a220fa2de4ffe3ba93b69f02d3ff2bc146f9624

apcraig · 2026-01-16T01:22:13Z

Much better. 3 PEND and 2 FAIL. I think this is expected for the hdf5/cdf tests?

https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_hash_forks#3a220fa2de4ffe3ba93b69f02d3ff2bc146f9624

Yes, that looks like what we expect.

eclare108213 · 2026-01-16T16:52:18Z

Are you both done with your testing? @dabail10 have you gone through all of the history fields to make sure they look reasonable?

Before it's merged, please add PR header statements like "closes issue #[whatever]" so that we don't have to close them manually. It would be nice to merge this today for weekend testing!

dabail10 · 2026-01-16T17:04:36Z

Are you both done with your testing? @dabail10 have you gone through all of the history fields to make sure they look reasonable?

Before it's merged, please add PR header statements like "closes issue #[whatever]" so that we don't have to close them manually. It would be nice to merge this today for weekend testing!

I have looked through all the fields numerous times and I am happy. I have copied one year of data to here:

https://ftp.cgd.ucar.edu/pub/dbailey/cicecmip/

It would be great if others had a look in case I missed something. My io_suite test did as expected. I will investigate how to do the automatic closing of issues.

apcraig · 2026-01-16T17:26:30Z

I have run a partial test suite on Derecho with 4 compilers and feel like this is ready to be merged. I will merge this afternoon unless someone does it before me or anyone notes any last minute concerns.

apcraig · 2026-01-16T17:34:25Z

So, are there five issues that should close?

"To automatically close issues when merging a pull request (PR), use specific keywords like
closes, fixes, or resolves followed by the issue number (e.g., closes #123) in the PR's description or commit message, which triggers the closure upon merging into the default branch on platforms like GitHub and GitLab."

It looks like you added the issues that are related, please update the PR description to explicitly add phrasing like

closes #123

for any issue that should automatically close.

dabail10 · 2026-01-16T17:43:37Z

So, are there five issues that should close?

"To automatically close issues when merging a pull request (PR), use specific keywords like closes, fixes, or resolves followed by the issue number (e.g., closes #123) in the PR's description or commit message, which triggers the closure upon merging into the default branch on platforms like GitHub and GitLab."

It looks like you added the issues that are related, please update the PR description to explicitly add phrasing like

closes #123

for any issue that should automatically close.

Added. I thought if they were linked it would do the same.

apcraig · 2026-01-16T19:39:46Z

Added. I thought if they were linked it would do the same.

Maybe they do, it's hard to know for sure. Thanks for adding that text.

eclare108213 · 2026-01-16T21:51:00Z

Merged. Yay! Take the weekend off, will you? :)

dabail10 added 3 commits November 13, 2025 09:58

Commit ice_history changes before merge main

ee3602a

Merge branch 'main' of https://github.com/ESCOMP/CICE

80ca1d3

This is a big commit. I have gone through and updated the SIMIP

b6b59c1

variables for CMIP7. - Add new variables that were not there before - Change names for some of the variables - Add an f_CICE flag to be able to turn off CICE duplicate variables - Added more variables to the daily stream

dabail10 requested review from anton-seaice, apcraig and eclare108213 November 13, 2025 23:51

Forgot a couple broadcasts

30ec758

anton-seaice reviewed Nov 14, 2025

View reviewed changes

cicecore/cicedyn/analysis/ice_history.F90 Outdated Show resolved Hide resolved

cicecore/cicedyn/analysis/ice_history.F90 Outdated Show resolved Hide resolved

cicecore/cicedyn/analysis/ice_history.F90 Outdated Show resolved Hide resolved

dabail10 added 3 commits November 14, 2025 10:01

Add sidragbot and change concentrations to percent

bd9cf20

Line up some ampersands and change mask_ice_free_points where appropr…

9f429b8

…iate

Another sizable commit.

1dcff41

- add a few more SIMIP variables (sisndmasswind, ...) - remove sialb as this is no longer in the CMIP7 table - start some work on the aice versus aice_init thing

Another big commit on this.

5dad94a

- change avg_ice_present to be a character array - add capability to average when aice_init > 0 - add long_name (title) for SIMIP variables from the CMIP7 table - update some ice density calculations to use the mushy-layer brine+ice density

dabail10 added 2 commits November 19, 2025 14:42

Fix siitd variables

f2a463e

Add documentation on CMIP7 updates and f_CICE flag

1d15438

eclare108213 mentioned this pull request Nov 20, 2025

documentation for averaging #1056

Merged

18 tasks

anton-seaice reviewed Nov 21, 2025

View reviewed changes

dabail10 added 2 commits November 21, 2025 10:48

Fix lwout bug for coupler and revert change for history

d9ef7d8

Remove some of the worka accumulation where it is not necessary.

a724928

Change sidmasslat to sidmassmeltlat. Fix the signs on the melt terms.

9219a55

eclare108213 requested changes Nov 24, 2025

View reviewed changes

Further refine abort condition

d0455f6

Fix comment for snowfrac

f222ad1

anton-seaice mentioned this pull request Jan 13, 2026

Missing variables in spinup outputs for seaice: sisnconc (monthly) and sisnthick (daily, monthly) ACCESS-NRI/ACCESS-MOPPy#152

Open

Fix aborts for accumulation variables

5abb957

eclare108213 approved these changes Jan 14, 2026

View reviewed changes

cicecore/cicedyn/analysis/ice_history.F90 Outdated Show resolved Hide resolved

One more comment change

b653260

Merge branch 'CICE-Consortium:main' into cmip7

3a220fa

eclare108213 merged commit 27a4982 into CICE-Consortium:main Jan 16, 2026
2 checks passed

Conversation

dabail10 commented Nov 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR checklist

Uh oh!

anton-seaice left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

apcraig commented Nov 17, 2025

Uh oh!

dabail10 commented Nov 17, 2025

Uh oh!

dabail10 commented Nov 19, 2025

Uh oh!

anton-seaice commented Nov 20, 2025

Uh oh!

anton-seaice 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!

Uh oh!

dabail10 commented Nov 21, 2025

Uh oh!

dabail10 commented Nov 21, 2025

Uh oh!

dabail10 commented Nov 21, 2025

Uh oh!

anton-seaice commented Nov 24, 2025

Uh oh!

anton-seaice commented Nov 24, 2025

Uh oh!

dabail10 commented Nov 24, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

anton-seaice commented Jan 11, 2026

Uh oh!

dabail10 commented Jan 12, 2026

Uh oh!

dabail10 commented Jan 12, 2026

Uh oh!

eclare108213 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

dabail10 commented Jan 15, 2026

Uh oh!

apcraig commented Jan 15, 2026

Uh oh!

dabail10 commented Jan 15, 2026

Uh oh!

dabail10 commented Jan 15, 2026

Uh oh!

dabail10 commented Jan 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dabail10 commented Jan 16, 2026

Uh oh!

apcraig commented Jan 16, 2026

Uh oh!

eclare108213 commented Jan 16, 2026

dabail10 commented Nov 13, 2025 •

edited

Loading

dabail10 commented Jan 16, 2026 •

edited

Loading