Meeting Minutes 2026: Atmosphere Working Group

Atmosphere Working Group
,
7

Atmospheric WG meeting 24 June 2026

15 participants on Zoom

Co-chair changes

  • Introducing Andrew Brown as new working group co-chair
  • Yi will be replaced by Bethan White as co-chair in the coming weeks. Bethan is currently away but will be back for next month’s meeting

Talk from Ben Schroeter, ACCESS-NRI - Transitioning from SVN to Git

  • Ben and the land group have recently transitioned from SVN to Git for CABLE development in 2023
  • Primarily used the git-svn tool to convert existing svn directories to git repos, retaining the change history etc. However, there are some issues with git-svn when applied to MOSRS repos
  • Ben provided a nice demo of how to convert a checked out rose directory to a git repo and push it to GitHub
  • Q from Sonya: What about converting an existing suite that has been modified? Ben: May need to use git-svn for that
  • The UKMO is transitioning to git. Dates to know:
    • Oct 2026: ACCESS-NRI transitioning to git
    • Nov 2026: MOSRS code becomes read-only
    • Feb 2027: MOSRS suites become read-only
  • Uncertainty about the status of MOSRS post-Feb 2027. Paul L notes that it may exist until the end of 2027 but not supported/updated.
  • Q from Alanah: Can we still run rose suites after Feb 2027? Ben: Yes, but can’t push changes back onto MOSRS.
  • Changing source code: Will need to be done with spack, soon. Some details on the hive but more to come
  • SVN-Git training from Ben available on YouTube (see agenda)
  • Q from Paul L: Does the community need regular git check-in sessions like they do at UKMO? Sonya: Maybe we could have a separate place on the hive for git support. Training on git and interacting with the UKMO GitHub and source code would also be good.
  • Ben notes tutorials around AM3 development are coming up soon.

Talk from Alanah Chapman, UniMelb - Aerosols and DMS in AM2 over the Southern Ocean

  • This is preliminary/ongoing work using AM2.
  • Motivated by GCM radiation biases in the Southern Ocean that relates to equilibrium climate sensitivity in models.
  • A low bias in aerosol concentrations over the Southern Ocean has been found previously.
  • Alanah is testing if including DMS emissions from sea ice (not currently in the model) can improve these biases.
  • When comparing the DMS concentrations against stations and voyage data, the control run (no DMS from sea ice) performs well, while including DMS from sea ice leads to too much DMS compared with observations.
  • However, the aerosol concentrations are significantly improved in the runs with too much DMS. This suggests that there are issues in the model with how DMS is converted to sulphates/cloud condensation nuclei.
  • Q from Martin D: Do we know exactly where the issue in the model? Alanah: Only that it relates to the processes where DMS is converted to sulphates.

Martin Dix - ACCESS-AM3-chem development

  • What does the community want us to focus on? A global (CM2-like) configuration is straightforward within 1 year. Regional (rAM3 with chem) would be more challenging
    • Mat: Having chemistry and stratospheric aerosols in AM3 would be nice, but rAM3 with chemistry would be good for some use cases (e.g. like Alanah’s work). Potentially a global configuration first then regional?
    • Q from Paul L: How is this different from UKCA? A: Not just aerosols like UKCA, but has full chemical pathways, ozone, stratospheric chemistry/aerosols, etc.
    • Martin: A challenge for regional development is getting high-res emission files

Updates from Heidi

  • The CoE for 21st Century Weather is looking for people to help evaluate a new set of N512 runs (see agenda for details and links). Pao: More runs coming soon, stay tuned. Q from Sonya: Can they be compared with observations directly? A: They are fairly short runs from 2007 onwards, so not really.
  • ML workshop: Please register and also have a think about what we should use our Atmosphere WG meeting time for (45 mins)
  • ACCESS-NRI docs transitioning to Read the Docs. Be aware there may be some ads on the new site
  • Compute resources: Note there are still 1.3 MSU available to burn this Q, otherwise they will be given to other WGs. Qinggang: I can burn a lot of SUs if needed.

Next WG meeting: July 2026 TBC