Meeting Minutes
Date: February 4, 2025
Attendees: @cbull, @clairecarouge, @micael, @paulleopardi, @MartinDix, @inh599, @spencerwong, @dougiesquire, @pearseb, @dhb599, @harshula, @Aidan, @tiloz, @tammasloughran, @Jhan, @manodeep
Minutes and chair: @cbull
Apologies: @RachelLaw, everyone else…
ACCESS-ESM1.6 project board review
The focus was on reviewing project components, identifying blockers, and discussing any issues or updates.
- Chris Bull invited attendees to mention any blockers or issues missing from the project board.
- Micael noted the need to add issues related to the software transformation team’s work. Claire Carouge encouraged adding more components and organizing them as needed.
- Paul is documenting ancillary work for esm 1.6 for cmip7 and was interested in this issue. Martin can share a suite to look into it.
Land Team Updates
- Ian: still working on resolving energy balance issues. The fix looks to have an instability in AM3 that they’d prefer not to inherit into ESM. Discussions around updating vegetation distribution.
- Claire update: Jhan has been working on a large PR that Claire has been reviewing. Vegetation/soil parameters (all the cable namelists) are being read in by the UM rather than being hardcoded making it easier to change them at runtime. Spencer asked if the new namelists files should be on a payu update or vk83. Ian: cautioned that perhaps some parameters shouldn’t be available to be changeable to users. Claire: countered suggesting that we can make them available and it is up to users to change with care.
- Ian: thinking about making greater use of soil ancillary information in esm 1.6. Martin: would you want to move to the new data sources? Ian: need to think carefully about the best workflow. Paul/Ian: CMIP7 land use cover forcing is indeed part of this. Tilo: what is the status of the restart file such that we can use the new vegetation maps? Ian: Lachlan has been working on this, see the forum topic on it – sounds like he’d appreciate input. Claire: Rachel has provided an update on this, this morning.
Ocean Team Updates
- Dougie and Parse: no major changes from last week. Pearse: in year 352 in ESM1.6 has modified the deep alkalinity, dic, nitrate and oxygen back to the initialised restart state. This was done due to the bug that we found ~100 year. Model now seems to be more stable 200-250 only required one restart, whereas in the first few hundred years there were quite a few more restarts required (atmosphere was slightly perturbed).
- Dave Bi: cice5. Still working on cice5 update for ESM1.5 system. Initialisation has gone well, reading in initial files/model set-up looks to run. Had some issues with model code, is getting a seg fault.
Release Team Updates
- Harshula discussed improvements in SPACK develop instructions and stricter rules from upstream SPACK developers.
- Aidan Heerdegen emphasized the importance of merging pull requests and ensuring the process works smoothly.
- Harshula: looking to improve the spack develop process. Been getting some input from the spack developers and they’ve made some suggestions and found a bug which has led to a change (stricter rules) in the spec, this means updating
spack.yamlfiles (example). Related changes have been propagated. ESM 1.6 is affected in minor ways, in this instance where a branch rebase or similar is required a comment should have appeared in PRs (like other infrastructure changes). The release team will try to minimise the impact of this. - Aidan: question to Pearse about the bug: is Pearse happy with the branch-develop workflow he’s using? Yes, those fixes are inside wombatlite-sinking branch. Need to eventually get those into main branch. Aidan: Could be a good time to fix the above Spack bug/update.
Closing Remarks
- Chris Bull highlighted software transformation team are looking to improve performance of esm1.6 over the next few months.
- We have a meeting next week Claire will be chairing the next meeting.