CM3 Dev-Eval Working group: Meeting minutes 2025

[spencerwong]

CM3 Dev-Eval Working group meeting minutes

Meeting link
Date: Tuesday 3 March at 10:30 am AEDT
Participants: @peterdobro, @kieranricardo, @JulieA, @RachelLaw, @wghuneke, @paulleopardi, @aekiss, @sofarrell, @ezhilsabareesh8, @zoegillett27, @DeepashreeDutta, @Harun_Rashid, @spencerwong, @Dietmar_Dommenget, @ShayneM
Chair: @kieranricardo
Minutes: @spencerwong

Subscribe to the ESM Working Group calendar as follows:

1. ESM Working Group Meeting Calendar feed (right click on the link and copy the URL)
2. Use calendar feed URL to subscribe to the calendar

Agenda:

• Description of new simulation. (Output at /g/data/zv30/non-cmip/ACCESS-CM3/cm3-run-20-01-2026-om3-update/cm3-datastore/cm3-datastore.json)
  • The new simulation updates the ocean to the latest OM3 release, including parameter changes.
  • Remapping wind stresses via the patch method removes artefacts in the ocean currents.
  • Land ice fluxes improve the water balance and drift in sea surface hight, however increase the sea ice extent especially in the southern hemisphere.
  • Work on the GC5 configuration is continuing to improve the energy balance.
• shout-out props for Figures from: @ctychung, @Harun_Rashid, @ezhilsabareesh8
• CM3 technical meeting updates (Jan 28, Feb 11, Feb 25)

GitHub issue links:

• @ctychung (new run) Monthly mean precipitation and surface temperature. GHLINK
  • Mean precipitation figures show a double ITCZ bias. Flux correction runs were suggested, where the coupled heat exchange would be manipulated to bring the SSTs closer to the observed mean state, as a way to address this bias, or as a way to investigate the sources of model biases. It may be possible to set up a flux correction run via the MOM settings.
  • An AM3 evaluation repository is being set up, which will make it simpler to compare biases in coupled and atmosphere only simulations.
• ENSO and IOD variability Evaluation: ENSO and IOD power spectra & seasonal cycle · Issue #15 · ACCESS-Community-Hub/access-cm3-paper-1 · GitHub

GitHub issue links to OM3 figures for CM3:

• @ezhilsabareesh8 Meridional transects of mean planetary geostrophic potential vorticity, potential density anomaly and max/min monthly mean MLD

  • The mixed layer depth is improved compared to CM2, which may be related to the swap from the KPP to EPBL mixing scheme.

• @ezhilsabareesh8 Mean zonally integrated overturning circulation

  • CM3 shows water mass changes which may be related to the GM parameterisation. @AndyHoggANU has been testing parameters which may improve the circulation.

• @ezhilsabareesh8 Time series of AMOC. Compared with observational estimate of McCarthy et al. (2015)

  • The trend in AABW suggests that the model is still spinning up. The issue with the excessive ice extent may feed into this. AMOC strength at 55N is almost double CM2, which may be related to convection activity in the Labrador Sea.
  • Future plots of upwelling could use 34S rather than 30S, as there are observations that could be added to the figure.

• @ezhilsabareesh8 Drake passage transport in CM3 new run

  • The Drake passage transport is slightly lower than the previous CM3 run, but still higher than the Donohue value which may itself be too high. Both of the two latest OM3 versions have a transport close to the Donohue value, suggesting the bias arises from the coupling.

Community figures

@Harun_Rashid presented plots on:

• Sea surface height: The sea surface height drift seen in the previous CM3 run no longer occurs. This would be due to the addition of land ice fluxes which close the water budget.
• Biases in SAT: Compared to ERA5, CM3 SAT is much cooler, especially in the Arctic. This is similar to the biases in CM2 and other CMIP6 models.
• Linear trends in SAT: The CM3 run (using present day forcings) is currently cooling. The model contains an initial cool bias, and is expected to warm as it spins up. @wghuneke’s figures show similar behaviour in CM2 and 0.25 degree CM2, where an initial period of cooling was followed by a long period of warming.
• Regional means, standard deviations, and trends: SSTs in several regions show higher variability than observations, especially in the Niño3/3.4 regions of the Pacific.
• Power spectra: Niño power spectra shows a spectral peak at ~3.3 years, however the period is too regular and the overall variability is too high.
• ENSO metrics: Comparing several ENSO metrics against several CMIP6 models shows that the current CM3 run performs quite well. Biases in most metrics are small, though larger biases exist in the thermocline feedback, zonal wind forcing, sst standard deviation and sst skewness.

Please paste here and include a caption!