Master esm1.6 merge #68

anton-seaice · 2025-10-29T04:20:25Z

This pull request merges the access-esm1.6 branch of CICE5 into the master branch.

The access driver is used for access-esm1.6, whilst the auscom driver is used for access-om2

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@357 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

…models/cice_GC3_GA7_hxyo/ git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@358 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@359 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@360 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@366 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@367 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@368 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@372 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@393 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@394 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@395 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@396 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

… years. git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@402 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@403 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

…tion errors. git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@404 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@405 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@406 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

* Add build ci for access-esm1.6 branch

* Needed to support oneapi 2025 compiler.

CICE actually uses a proleptic gregorian calendar, see https://cfconventions.org/Data/cf-conventions/cf-conventions-1.12/cf-conventions.html#calendar

Add license file from CICE-consortium

Ported https://github.com/ACCESS-NRI/cice4/blob/694a9fbd4ac29dc841b38aff002eb36da5b650f1/source/ice_history.F90#L2184-L2198 to cice5. This avoids post-processing interpreting end of the time interval timestamps as the subsequent time interval.

Implement dump last in ESM1.6 driver

This makes the assumed salinity of sea ice configurable, so we can set it the same as the MOM value (configured in MOM [here](https://github.com/ACCESS-NRI/access-esm1.6-configs/blob/88ac5aab5d6d2500209b7d610e68e5c2928222d4/ocean/input.nml#L367)) We use 4ppt in MOM for historical reasons, and did that in [CICE4](https://github.com/ACCESS-NRI/cice4/blob/694a9fbd4ac29dc841b38aff002eb36da5b650f1/source/ice_init.F90#L243) too.

It's hard to set this in the namelist because it is used at compile time to define several parameters Set per MOM5 value: https://github.com/ACCESS-NRI/FMS/blob/bf9b80423ea4f66efa389e03d3c19b26c009e8b6/constants/constants.F90#L90

* Expose ksno (thermal conductivity of snow ) to the namelist Typical default value is 0.3, for CM2, use 0.2

Co-authored-by: Spencer Wong <88933912+blimlim@users.noreply.github.com>

anton-seaice · 2026-01-09T00:44:16Z

drivers/access/cpl_interface.F90

https://github.com/ACCESS-NRI/cice5/compare/access-esm1.6..ACCESS-NRI:cice5:master-esm1.6-merge#diff-fc19414c40fff80d4d66454635b302d42cec694887445d8c275eaf2a219f9285 shows minor formatting change

anton-seaice · 2026-01-09T00:45:10Z

drivers/access/cpl_parameters.F90

https://github.com/ACCESS-NRI/cice5/compare/access-esm1.6..ACCESS-NRI:cice5:master-esm1.6-merge#diff-42d5bf23150be2eb7f74b5edb9ce795669a7a91bcb3eef33c78acc43d0a6afdb shows minor formatting change

anton-seaice · 2026-01-09T00:46:01Z

drivers/access/ice_constants.F90

https://github.com/ACCESS-NRI/cice5/compare/access-esm1.6..ACCESS-NRI:cice5:master-esm1.6-merge#diff-da3f41b82e09ada9d81d1b1e62ee31852b99a7e9bb813f6eaa861f71fd6c63ad only changed to move iceruf

Does iceruf also need to be removed from auscom/ice_constants?

anton-seaice · 2026-01-09T00:46:23Z

drivers/access/ice_coupling.F90

https://github.com/ACCESS-NRI/cice5/compare/access-esm1.6..ACCESS-NRI:cice5:master-esm1.6-merge shows no change to this file

Read grid on each processor per COSIMA/access-om2#221

…6-merge

fix line truncation build failure

anton-seaice · 2026-01-09T04:15:01Z

@dougiesquire and @blimlim

Here, I present a monster PR for your enjoyment.

This merges the access-esm1.6 branch with the master branch, all repro tests show no change in results, see:

1 and 2

There is a bit of stuffing around to make this happen:

refined CMake to work with esm1.6
some constants are configured in the namelist (this is consistent now between esm1.6 and om2, no defaults are changed).
for esm1.6 specific code, this is marked as #ifdef ACCESS. #ifdef AusCOM is used for bother esm1.6 and om2, so sometimes #if defined(AusCOM) && !defined(ACCESS) is used for OM2 only. This is used a lot for calendar code.

I snuck in changing the history output precision of the time coordinate to double, technically it's unrelated to this change.

the drivers/access folder comes from the esm1.6 branch. There are minor changes to the drivers/auscom folder to support some constants becoming configurable and improvements made in esm1.6.

anton-seaice · 2026-01-12T23:39:54Z

I ran 12 consecutive single months of ESM1.6, and 2 consecutive single years of OM2 - and restarts are identical at the end of each case :)

dougiesquire · 2026-01-20T22:59:27Z

@anton-seaice, taking a look at this now. Could you rebase onto main (or merge main, whatever you prefer) when you get a chance.

anton-seaice · 2026-01-20T23:02:58Z

Rebase with merge commits with conflicts in them is a pain, so i merged master in

dougiesquire

Thanks @anton-seaice. I've had a look across all files and didn't notice anything awry, but for a PR this large and broad I think the testing you have done is about the best we have...

@blimlim were you interested in taking a look?

blimlim · 2026-01-21T01:02:35Z

I'll try to take a quick look this afternoon!

blimlim

Thanks @anton-seaice, I've had a read through and it's reassuring that no answers change! There's a lot I don't understand too well, especially around the mpi differences wrt the ESM branch.

I've just added a couple small comments – probably not crucial as they aren't having any effects on results.

blimlim · 2026-01-21T23:25:57Z

source/ice_itd.F90

+      if (present(arg_calc_Tsfc)) then
+         calc_Tsfc = arg_calc_Tsfc
+      else
+         calc_Tsfc = .false.
+      endif


Is this doing anything? I don't think calc_Tsfc isn't being used elsewhere in init_itd

blimlim · 2026-01-21T23:46:00Z

source/ice_itd.F90

+      if (present(arg_heat_capacity)) then
+         heat_capacity = arg_heat_capacity
+      else
+         heat_capacity = .false.
+      endif


I'm getting a bit confused about the setup here especially for OM2. The heat_capacity doesn't seem consistent with the value used everywhere else. E.g. OM2 hasktherm=1and so should haveheat_capacity=.true.` in a lot of other places:

cice5/source/ice_therm_vertical.F90

Lines 646 to 647 in 1086aec

heat_capacity = .true.

if (ktherm == 0) heat_capacity = .false. ! 0-layer thermodynamics

However, init_itd is called without any arguments, meaning that heat_capacity=.false. for this subroutine. If we were to use heat_capacity=.true. though, we'd then end up with wrong values for hs_min and hi_min for OM2:

cice5/source/ice_itd.F90

Lines 249 to 255 in 1086aec

if (heat_capacity) then

! Set higher values to help with stability

hi_min = p2 ! 0.2m

hs_min = p1 ! 0.1m

else

hs_min = 1.e-4_dbl_kind ! min snow thickness for computing zTsn in OM2 (m)

endif

Would it be any clearer to do something like:

#ifdef ACCESS subroutine init_itd (arg_calc_Tsfc, arg_heat_capacity) #else subroutine init_itd #endif

and

#ifdef ACCESS if (heat_capacity) then ! Set higher values to help with stability hi_min = p2 ! 0.2m hs_min = p1 ! 0.1m else hs_min = 1.e-4_dbl_kind ! min snow thickness for computing zTsn in OM2 (m) endif #endif

blimlim · 2026-01-21T23:53:02Z

drivers/auscom/ice_constants.F90

+#ifndef AusCOM
+      real (kind=dbl_kind), parameter, public :: &
+         !!! dragio    = 0.00536_dbl_kind ,&! ice-ocn drag coefficient
+         dragio    = 0.01_dbl_kind ,&!!! 20170922 test new value as per spo 


This one's already set in line 39. Which value should it be?

(not that we use either)

blimlim · 2026-01-21T23:54:41Z

drivers/auscom/ice_constants.F90

+      real (kind=dbl_kind), parameter, public :: &
+         !!! dragio    = 0.00536_dbl_kind ,&! ice-ocn drag coefficient
+         dragio    = 0.01_dbl_kind ,&!!! 20170922 test new value as per spo 
+         Tocnfrz   = -1.8_dbl_kind    ,&! freezing temp of seawater (C),


This is also set in line 53

blimlim · 2026-01-21T23:57:44Z

drivers/access/ice_constants.F90

Does iceruf also need to be removed from auscom/ice_constants?

blimlim · 2026-01-22T00:16:04Z

io_netcdf/ice_restart.F90

        n,     &      ! number of dimensions for variable
-        varid         ! variable id
+        varid, &      ! variable id
+        status        ! status variable from netCDF routine


I don't think this one's being used

blimlim · 2026-01-22T00:21:15Z

io_netcdf/ice_history_write.F90

+                        if (avail_hist_fields(n)%avg_ice_present) then
+                            call check(nf90_put_att(ncid,varid,'cell_methods',&
+                                'area: time: mean where sea_ice (mask=siconc)'), &
+                                'put att cell methods time mean '//avail_hist_fields(n)%vname)
+                        else
+                            if (TRIM(avail_hist_fields(n)%vname(1:2))/='si') then !native diags
+                                call check(nf90_put_att(ncid,varid,'cell_methods','time: mean'), &
+                                'put att cell methods time mean '//avail_hist_fields(n)%vname)
+                            else !cmip diags
+                                call check(nf90_put_att(ncid,varid,'cell_methods', &
+                                'area: mean where sea time: mean'), &
+                                'put att cell methods time mean '//avail_hist_fields(n)%vname)
+                            endif
+                        endif


Should the cell methods changes be copied over to io_pio/ice_history_write?

blimlim · 2026-01-22T01:25:04Z

source/ice_aerosol.F90

+      use ice_itd, only: hs_min
      use ice_state, only: nt_aero 
      use ice_shortwave, only: hi_ssl, hs_ssl



Do the changes compared to the access-esm1.6 branch have any effect? https://github.com/ACCESS-NRI/cice5/compare/access-esm1.6..ACCESS-NRI:cice5:master-esm1.6-merge#diff-f96fef21927cbb81494c664c5f72d85fc5b7cbe2bda3f322723d3b954dc9b29f

Or are they not too relevant as we ESM1.6 doesn't (can't?) run with tr_aero

blimlim · 2026-01-22T05:18:05Z

source/ice_dyn_evp.F90

+      call ice_HaloUpdate(strocnxT, halo_info, &
+                              field_loc_center,  field_type_vector)
+      call ice_HaloUpdate(strocnyT, halo_info, &
+                              field_loc_center,  field_type_vector)


I don't understand enough about the halo updates, but I guess the extra halo updates don't effect answers?

Also compared to the esm1.6 branch,icetmask seems to get replaced by iceumask? I don't understand what effect this would have, but it doesn't seem to affect answers.

- if (maskhalo_dyn) & - call ice_HaloMask(halo_info_mask, halo_info, icetmask) - call ice_timer_stop(timer_bound) + if (maskhalo_dyn) then + call ice_timer_start(timer_bound) + halomask = 0 + where (iceumask) halomask = 1 + call ice_HaloUpdate (halomask, halo_info, & + field_loc_center, field_type_scalar) + call ice_timer_stop(timer_bound) + call ice_HaloMask(halo_info_mask, halo_info, halomask) + endif

blimlim · 2026-01-22T23:21:44Z

source/ice_flux.F90

                               fsens,    flat,     &
                               fswabs,   flwout,   &
                               evap,               &
+                               evap_ice, evap_snow,&


Probably overly pedantic as it's not being called, but should these arguments be added to the (non) scale-fluxes call in drivers/auscom/CICE_RunMod:

cice5/drivers/auscom/CICE_RunMod.F90

Line 617 in 1086aec

evap (:,:,iblk), &

pbd562 and others added 30 commits August 22, 2016 04:02

Import from /short/p66/dhb599/ACCESS-CM2/submodels/cice_GC3_GA7_hxyo/

1071b67

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@357 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Updated cpl_forcing_handler.F90 from /short/p66/dhb599/ACCESS-CM2/sub…

51af00a

…models/cice_GC3_GA7_hxyo/ git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@358 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

I2A coupling bug fix

6f30219

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@359 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

ice errmax tolorance x 10

8d76453

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@360 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

dhb599 mod driver files

4520c26

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@366 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

ukmo:changeset 145,146,163 & 134

2b434c7

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@367 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Bug fix in ice constants

2b9eba7

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@368 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

add dhb599 iceberg scheme

823a987

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_GC3_GA7@372 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

New branch for GSI 8.1

407ade8

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@393 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Remove symliink before adding back as file

a24447b

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@394 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Add code changes from Arnold's cice512_av853

6e363ca

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@395 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Add iceberg_factor (from Dave)

9953401

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@396 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Add extra call to set_calendar to fix naming of history files in leap…

8b553d9

… years. git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@402 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Write messages to stderr and call mpi_abort with correct arguments.

575c1e7

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@403 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Create netCDF files as netcdf4-classic for better detection of trunca…

d1c40ef

…tion errors. git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@404 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Only write namelist info from master PE

d49e6bc

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@405 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Add coupling timers for ACCESS

61ab72a

git-svn-id: file:///g/data/access/access-svn/cice/branches/access/cice_gsi8.1@406 f6bd92a4-46cf-401b-8a38-b7f7993d28bf

Create README.md

0a84ab1

Add build ci (#21)

8be14fe

* Add build ci for access-esm1.6 branch

bld/makdep.c: add a return type to main() (#22)

2898631

* Needed to support oneapi 2025 compiler.

Set netcdf output calendar attributes correctly (#26)

ec1e2cb

CICE actually uses a proleptic gregorian calendar, see https://cfconventions.org/Data/cf-conventions/cf-conventions-1.12/cf-conventions.html#calendar

Add license file (#29)

77baf8a

Add license file from CICE-consortium

Modify dump_last for use in esm1.6

ed83b97

Keep original implementation for non-coupled model

d87c459

Merge pull request #38 from ACCESS-NRI/esm1.6-dump-last

5471bbf

Implement dump last in ESM1.6 driver

Set cp_ocn to same value as MOM (#40)

15fa7de

It's hard to set this in the namelist because it is used at compile time to define several parameters Set per MOM5 value: https://github.com/ACCESS-NRI/FMS/blob/bf9b80423ea4f66efa389e03d3c19b26c009e8b6/constants/constants.F90#L90

Expose ksno (thermal conductivity of snow ) to the namelist (#41)

267d4be

* Expose ksno (thermal conductivity of snow ) to the namelist Typical default value is 0.3, for CM2, use 0.2

ESM1.6 Readme (#44)

58eeccf

Co-authored-by: Spencer Wong <88933912+blimlim@users.noreply.github.com>

anton-seaice commented Jan 9, 2026

View reviewed changes

anton-seaice force-pushed the master-esm1.6-merge branch from 29a9581 to adef59d Compare January 9, 2026 02:09

anton-seaice added 9 commits January 9, 2026 13:52

Merge 'access-esm1.6' into master (e.g. access-om2) branch

156780b

Read grid on each processor per COSIMA/access-om2#221

Merge remote-tracking branch 'origin/master' into master-esm1.6-merge

1d35801

Merge remote-tracking branch 'origin/master' into master-esm1.6-merge

2bb20e4

Merge remote-tracking branch 'origin/access-esm1.6' into master-esm1.…

b7e4301

…6-merge

Use ACCESS instead of AusCOM for ESM1.6, some merge tidyup

3f8ba8a

Support CMake Build for esm1.6

7a40ca0

fix line truncation build failure

Remove unused files

442d45b

Sync namelist error handling between ACCESS & AUSCOM

5586d4d

Use double precision for time diagnostic

99ce18a

anton-seaice force-pushed the master-esm1.6-merge branch from adef59d to 99ce18a Compare January 9, 2026 03:04

Fix ordering of setting aicenmin

8031135

anton-seaice requested review from blimlim and dougiesquire January 9, 2026 04:15

anton-seaice marked this pull request as ready for review January 9, 2026 04:15

anton-seaice added 3 commits January 12, 2026 13:22

Turn off per process logging - not used

f04d945

Reinstate elevated limits for thermo errors in esm1.6

eb00a6d

Send all nu_diag logging to one file

af61594

Merge branch 'master' into master-esm1.6-merge

1086aec

dougiesquire approved these changes Jan 20, 2026

View reviewed changes

blimlim reviewed Jan 23, 2026

View reviewed changes

	heat_capacity = .true.
	if (ktherm == 0) heat_capacity = .false. ! 0-layer thermodynamics

	if (heat_capacity) then
	! Set higher values to help with stability
	hi_min = p2 ! 0.2m
	hs_min = p1 ! 0.1m
	else
	hs_min = 1.e-4_dbl_kind ! min snow thickness for computing zTsn in OM2 (m)
	endif

Master esm1.6 merge #68

Are you sure you want to change the base?

Master esm1.6 merge #68

Uh oh!

Conversation

anton-seaice commented Oct 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

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!

Choose a reason for hiding this comment

Uh oh!

anton-seaice commented Jan 9, 2026

Uh oh!

anton-seaice commented Jan 12, 2026

Uh oh!

dougiesquire commented Jan 20, 2026

Uh oh!

anton-seaice commented Jan 20, 2026

Uh oh!

dougiesquire left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

blimlim commented Jan 21, 2026

Uh oh!

blimlim 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!

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

anton-seaice commented Oct 29, 2025 •

edited

Loading

dougiesquire left a comment •

edited

Loading