dougiesquire
(Dougie Squire, ACCESS-NRI Ocean/Sea-Ice Modelling Team)
1 July 2026 06:51
14
Date: 1 Jul 2026
Attendees: @Aidan , @aekiss , @anton , @NoahDay , @cbull , @sofarrell , @ezhilsabareesh8 , @dougiesquire
OM3 control directory restructure
Aidan has opened an issue suggesting that we add realm subdirectories to the configurations.
opened 01:28AM - 29 Jun 26 UTC
I've (somewhat belatedly) realised you have opted for a flat directory structure… for your [access-om3-configs](https://github.com/ACCESS-NRI/access-om3-configs/tree/dev-MC_25km_jra_ryf).
I understand this is probably due to their not being a necessity for model sub-directories given the single executable of the nupoc coupled approach.
However, I would like to put forward the case for model specific configuration sub-directories, which I think have a number of advantages.
### Proposal
Create sub-directories for each model that contain the configuration files specific to that model, and keep all top level (nuopc) configuration files either in the top level of the control directory, or in a specific coupler/nuopc directory (I'm not fussed).
Current structure (showing only relevant files)
```bash
WW3_PreProc/
├── GenGrid.py
├── OM3_100km.Dpt
├── OM3_100km.Lat
├── OM3_100km.Lon
├── OM3_100km.Mask
├── OM3_100km.Obstr
├── namelists_Global.nml
├── ww3_grid.nml
└── ww3_strt.inp
docs/
├── MOM_parameter_doc.all
├── MOM_parameter_doc.debugging
├── MOM_parameter_doc.layout
├── MOM_parameter_doc.short
└── available_diags.000000
LICENSE
MOM_input
MOM_override
README.md
config.yaml
datm.streams.xml
datm_in
diag_table
diag_table_source.yaml
drof.streams.xml
drof_in
drv_in
fd.yaml
ice_in
input.nml
nuopc.runconfig
nuopc.runseq
wav_in
ww3_points.list
```
Proposed structure:
```
atmosphere/
├── datm.streams.xml
└── datm_in
ocean/
├── docs/
│ ├── MOM_parameter_doc.all
│ ├── MOM_parameter_doc.debugging
│ ├── MOM_parameter_doc.layout
│ ├── MOM_parameter_doc.short
│ └── available_diags.000000
├── MOM_input
├── MOM_override
├── diag_table
├── diag_table_source.yaml
└── input.nml
seaice/
└── ice_in
waves/
├── WW3_PreProc/
│ ├── GenGrid.py
│ ├── OM3_100km.Dpt
│ ├── OM3_100km.Lat
│ ├── OM3_100km.Lon
│ ├── OM3_100km.Mask
│ ├── OM3_100km.Obstr
│ ├── namelists_Global.nml
│ ├── ww3_grid.nml
│ └── ww3_strt.inp
├── wav_in
└── ww3_points.list
LICENSE
README.md
config.yaml
drof.streams.xml
drof_in
drv_in
fd.yaml
nuopc.runconfig
nuopc.runseq
```
(Note I've given an example of having an `atmosphere` directory for `datm`, but maybe that doesn't make sense if `datm` would remain even if there was an atmosphere model)
I believe this new structure has a number of advantages:
1. **Clarity**: it is obvious from inspection which configuration files belong to which model
2. **Consistency**: other coupled models use this structure, so it is familiar to users
3. **Extensibility**: sub-directories mean there is no problem with name clashes when new models are adopted (c.f. the popularity of `input.nml`)
By just doing this rearrangement it becomes obvious the `docs` refer only to the ocean model, and it is obvious where model specific files should live. There is nothing to stop you having a top level `docs` directory, but in that case I would recommend having an `ocean` or `MOM6` sub-dir for those docs. Same for the WW3 pre-processing.
I have used a WW enabled model not to pick on @ezhilsabareesh8, but because it makes the change more compelling, and also highlights issues when more models are added for ESM3.
A similar restructure of the output would also be good, but it more difficult.
General agreement that the new structure is clearer.
Actions:
Dougie will scope out how much work is involved.
Replace panan4km+isf RYF with IAF
Shall we replace the panan4km+isf RYF configuration with an IAF equivalent?
opened 02:48AM - 29 Jun 26 UTC
MC_4km_jra_ryf+regionalpanan+isf
The `dev-MC_panan4km_jra_ryf+isf` configuration is out of sync with what @adele-… morrison, @claireyung and others are now using for research.
Differences at this point in time are: https://github.com/claireyung/access-om3-configs/compare/beb0c82297f79bc7fd640bfa7decf5f689964cb0..ACCESS-NRI:access-om3-configs:c164464e7f118c705d441da67965eee9d452a551
I had a chat with @adele-morrison and @claireyung this morning. These are the key differences/updates:
- [ ] **1.** IAF vs RYF. We should consider replacing `dev-MC_panan4km_jra_ryf+isf` with an `iaf` configuration.
- One complication is that they are currently manually updating the open boundary conditions every 5 years
- [ ] **2.** New input files:
- [ ] **a.** Topo updates:
- `kmt.nc`,
- `topog.nc`,
- `access-om3-4km-ESMFmesh.nc`,
- `access-om3-4km-nomask-ESMFmesh.nc`,
- `ice_shelf_thick.nc`
(note, `topog_seaice.nc` is used to generate `access-om3-4km-ESMFmesh.nc` but not actually used in the config)
- [ ] **b.** Bottom roughness. This has been updated to use the latest version of the script (see my comment [here](https://github.com/ACCESS-NRI/access-om3-configs/issues/1138#issuecomment-4806209700))
- [ ] **c.** Layout/mask table
- [ ] **d.** Rofi spreading
- [ ] **e.** Initial conditions
We decided that **a** and **b** could be updated in `dev-MC_panan4km_jra_ryf+isf`, but the others probably require some more testing. The updated inputs will still need to be checked over - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1138.
- [ ] **3.** Added diagnostics. Not urgent as user-specific and easy to change.
No objections.
Actions:
Dougie will make the switch
25km releases todo
Dougie shared a todo list of things to be done before 25km releases
opened 06:03AM - 29 Jun 26 UTC
m-MC_25km_jra
Things to do before we can (beta) release all of the `MC_25km` configs:
- [x] 1… . Set `DO_RIVERMIX = True` and `RIVERMIX_DEPTH = 40.0` - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1394
- [x] 2. Set `MAX_DELTA_SRESTORE = 2.0` (only needed for IAF) - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1180
- [x] 3. Add new diagnostics - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1046
- [x] 4. Turn off collation of restarts (only needed for WOMBATlite configs) - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1013
- [x] 5. Check/update `metadata.yaml` files
- [x] 6. Update `CITATION.cff` files
- [x] 7. Update version of `model-config-tests` used by CI - see https://github.com/ACCESS-NRI/access-om3-configs/pull/1480
- [ ] 8. Add section on WOMBATlite to the documentation - see https://github.com/ACCESS-NRI/access-om3-configs/pull/1481
- [x] 9. Update [docs page on 25km configs](https://access-om3-configs.access-hive.org.au/latest/configurations/MC_25km/) - see https://github.com/ACCESS-NRI/access-om3-configs/pull/1486
- [x] 10. Add info about `mom6-solo` exe to docs - see https://github.com/ACCESS-NRI/access-om3-configs/pull/1486
- [x] 11. Double check consistency between 25km configs - see https://github.com/ACCESS-NRI/access-om3-configs/issues/1482#issuecomment-4840466786
- [ ] 12. Add docs on iceberg spreading
- [ ] 13. Add docs on fresh water balance
- [ ] 14. Release note - WIP [here](https://forum.access-hive.org.au/t/draft-access-om3-release-post/6580)
- [ ] 15. Sync control experiments to `ol01`
- [ ] 16. Add experiments to experiment table in docs
- [ ] 17. Add new `release-` branches to [config docs home page](https://access-om3-configs.access-hive.org.au/latest/)
Anything missing?
Suggested changes/additions:
Set MAX_DELTA_SRESTORE = 0.5 in RYF
Add 3D and 2D conservative temperature diagnostics
There is not currently enough space in ol01 to store the experiment output associated with the releases. Josh, Adele and others are working on moving ~50TB of data, hopefully soon.
Options for release:
Wait
Remove some old experiments
$ du -hsc /g/data/ol01/outputs/access-om3-25km
4.0T 25km-iaf-test-for-AK-expt-7df5ef4c
4.7T MC_25km_jra_iaf-1.0-beta-5165c0f8
2.0T MC_25km_jra_iaf-1.0-beta-gm1-d968c801
2.2T MC_25km_jra_iaf-1.0-beta-gm2-5dc49da6
2.1T MC_25km_jra_iaf-1.0-beta-gm3-da330542
2.2T MC_25km_jra_iaf-1.0-beta-gm4-9fd08880
2.2T MC_25km_jra_iaf-1.0-beta-gm5-9b5dbfa9
2.1T MC_25km_jra_iaf+wombatlite_bs1-mpudig-backscat1-4c21c151
2.1T MC_25km_jra_iaf+wombatlite_bs2-mpudig-backscat2-0578cc36
11T MC_25km_jra_iaf+wombatlite-test3v2-00532b88
5.2T MC_25km_jra_iaf+wombatlite-test3v2-COARE-12fab415
2.4T MC_25km_jra_ryf-1.0-beta-cdfb3543
6.3T MC_25km_jra_ryf+wombatlite-81ad20e-c4347f5a
48T total
Andrew wants to keep 25km-iaf-test-for-AK-expt-7df5ef4c (CHL-informed SW turned off) and potentially a few others. There was agreement that the following could be removed:
2.0T MC_25km_jra_iaf-1.0-beta-gm1-d968c801
2.2T MC_25km_jra_iaf-1.0-beta-gm2-5dc49da6
2.1T MC_25km_jra_iaf-1.0-beta-gm3-da330542
2.2T MC_25km_jra_iaf-1.0-beta-gm4-9fd08880
2.2T MC_25km_jra_iaf-1.0-beta-gm5-9b5dbfa9
6.3T MC_25km_jra_ryf+wombatlite-81ad20e-c4347f5a
but in the end we decided to wait a bit (I think)
Release without data. “Data will be available at these locations”
May as well wait
Ask for more allocation
Ask for another project explicitly for control experiments
Possibly a good idea regardless of immediate decision
Actions:
Dougie will press on with doing Github release and will only post to the forum once the data is available in ol01.
ACCESS-OM2 experiments repo
Dougie will soon start some ACCESS-OM2 control experiments using the latest releases.
Shall we do a single experiment directory or one for each experiment?
Agreed on single access-om2-experiments repo in the Community organisation
Actions:
Dougie will set up repo.
Dougie will make Forum post inviting people to suggest new diagnostics to include in experiments (promised here ).
MOM6 dev meeting
PR from dev/ncar nearly merged. Large PR from dev/gfdl up next
Next meeting (4 weeks) will be annual default parameter change meeting
Ed Yang has a PR demonstrating a new blocking code structure (for the continuity solver, for now) that allows dynamically switching between CPU/GPU optimised layout. Performance is looking good