Meeting Minutes 2025: Atmosphere Working Group

Atmosphere Working Group
9

ACCESS NRI Atmospheric Working Group Meeting minutes - 19/11/2025

21 participants on zoom

Updates from the co-chairs
Charmaine Franklin (BoM):

  • RAL4 project has started, UKMO + partners. Four parts to this project: Ancil dev; Optimisation; Workflows / git; Model development and Scientific Assessment
  • CSET Convective Scale Evaluation Toolkit
    • Visualisation and diagnostics, work on including obs starting next year
    • Working with Jorge Bornemann from NZ - he is happy to talk to the WG meeting next year.

Yi Huang (Uni Melb):

  • 21stCenturyWeather Modelling Science update. Regional coupled atoms-ocean model team has made some progress → a test domain has been setup
  • Also some progress with regional atmosphere only domain. Scoping inclusion of Cable
  • Global atmosphere only model progress
  • Also working on urban model

Heidi Nettelbeck (ACCESS-NRI):

  • ACCESS-NRI - collected feedback from alpha-release of RAM3. Beta-release due on a few weeks. Still open to more feedback (including on OSTIA capability)
  • Progress towards an ACCESS-AM3 release

Data update Kelsey Druken

  • Data storage options at ACCESS-NRI
  • Introduction to Joshua Torrance, helping in data management at ACCESS-NRI
  • Merit allocation: Provide storage and compute resources
  • 3 Categories of data storage:
      1. Working group storage gx60 ~400TB. Most flexible storage option. Best suitable for data that’s changing and evolving. Longer term than /scratch, but not long-term storage.
      1. Data Sandbox - sharing data across the working groups & NCI platform. Flexible lightweight management, added to ACCESS-NRI’s intake catalogue. No DOIs or THREDDS - not what you need for publication
      1. Data Publication - longer-term support with persistent identifier.
  • Get in touch with data.access.nri@anu.edu.au for support for a reference dataset.

CMIP7 updates and timelines Rachel Law

  • DECK - core set of experiments that every model needs to submit as entry card
  • Fast Tracks
  • Community MIPS
    • See GMD Special Issue
  • ACCESS plans for CMIP7:
    • ACCESS-ESM1.6 (update of ESM1.5). Targeting Fast-track MIP
    • ACCESS-ESM3 family - Community MIPS - Tuesday monthly meeting 10:30 (see details on the forum)
  • ACCESS-ESM1.6.
    • Biggest change from ESM1.5 is CICE4 to CICE5.
    • Land - work adding Australian vegetation types
    • Atmosphere - not much change, 38 levels in the atmosphere
    • Changes around pseudo icebergs
    • Big improvements in compute, thanks to ACCESS-NRI
    • New plant functional types for Australia
    • Two types of Australian Trees - Xeric (~Acacia) and Mesic (Eucalypts) → Improved carbon fluxes
  • DECK and Fast track aim to run by the end of next year
  • Preference for emissions-driven with interactive carbon in CMIP7 (rather than prescribed concentration). We have capacity to do this in ESM1.6.
  • Updated CMIP6 / CMIP7 solar constant → Lower solar constant led to cooling in the system, but challenges with stabilisation. Started new runs with reduction in ocean albedo → greater stability
  • Carbon spinup - currently tuning ocean biogeochemistry.
  • Land - currently fairly confident of zero flux (photosynthesis balances respiration).
  • Equilibrium Climate Sensitivity - 3.36, a bit lower than previous 3.88 in ESM1.5.
  • Next steps:
    • Finalise pre-industrial forcings
    • Test historical simulation
    • Start piControl
    • Branch spin-up for emissions-driven pre-industrial
    • Test emissions driven historical runs
    • Test workflow for model output to data publishing.

ACCESS-AM3 release Martin Dix

  • Model configs:
    • ACCESS-AM3 - 21st Century Weather running at n512, new urban scheme in CABLE (Jiachen Lu).
    • ACCESS-CM3 - 100 km and 25km ocean
    • ACCESS-rAM3
  • GC5 Central
  • ACCESS-CM2 had a high ECS of 4.7K (IPCC best estimate range 2.5-4)
  • GAL9 runs > 6K, unacceptably high, perturbed parameter ensemble to find a result with a more reasonable ECS
  • New version GC5 Central, with ECS ~4 K.
  • Testing in CM3 - > getting closer to reasonable ECS.
  • Testing Cable in a version of rAM3.
  • Update using Australian PFTs, CMIP7 forcing
  • Building Cable as a library - > same Cable code for all models
  • Evaluation - data is now in intake catalogue for easy loading into Jupyter notebooks.
  • CM3 evaluation would welcome atmosphere and land evaluations
  • Is the AM3 evaluation part of this? Or does it have its separate evaluation.

Lachlan Whyborn:

  • AM3-alpha release ready some time next week
  • There will be a post on the forum where you can request access
  • Anyone who is interested in testing the alpha version, please keep an eye out for the post.

ESA CCI SST and sea-ice dataset Heidi Nettelbeck

  • Community interest for ACCESS-NRI to potentially support for ESA CCI SST and sea-ice dataset

Compute and storage resources Heidi Nettelbeck

  • Only experiment request this quarter from Chris Chambers.
  • Getting close to the end of the quarter. If you’re thinking of running something, please get it in quickly (next week or so)! → See here for how to apply for Atmosphere WG Resources.

Next WG meeting

  • February next year - avoid AMOS week
1 Like
Announce: Atmosphere Working Group