Iceberg capability in ACCESS-OM3/4

COSIMA
1

I’ve had a number of people asking me about whether ACCESS-OM3 will have iceberg capability. (@adfraser, Nicole Hill + Nicole named a handful of others in the BGC community who want them). Currently we have no way planned to simulate icebergs, because at GFDL icebergs are within the SIS2 sea ice model, but ACCESS-OM3 will use CICE6 sea ice. I’m guessing moving forward, the cryosphere community may also be interested in icebergs once we have coupled ocean / ice sheet models.

Anyway, just flagging this issue for those who have raised it with me, so it’s not forgotten. I’m not sure what the best way forward is. Possible options:

  • Work on including icebergs in CICE6?
  • Have a second version of ACCESS-OM3/4 with SIS2 instead of CICE6?
  • Keep running regional PanAntarctic MOM6/SIS2 models where we can study icebergs? But this may have limited capability for future projection simulations.
2

I think that a separate ACCESS-OM3 version with a different sea ice model sounds like a huge deal and let alone consider the maintenance of both of them. But I may be wrong and basing my thinking on struggles of the past — NUOPC might have made plug and play really easy!

But this is a good point raised and should be taken into consideration into the decisions for mode components for ACCESS-OM4?

3

I’m not sure NUOPC would help us here – as a SIS2-based model would use the FMS. I think it depends on the science case – if someone wants to investigate something about icebergs they could just use GFDL-OM5 which has SIS2 inbuilt; this is what we already do for PanAntarctic at the moment. But supporting it as a released configuration for general use would be hard to justify.

Maybe best here is to raise the issue more broadly with the CESM & CICE community to see whether there is alignment there?

4

I’ve not found anyone actively pursuing it for CICE at the moment. It’s not something CESM have on their current priorities. The closest is there is some work happening for MPAS-SeaIce, which whilst similar to CICE, is not CICE.

My suggestion is if we are happy with the science of the icebergs in SIS2, to investigate re-using that iceberg module/scheme in CICE.

5

Thanks Adele for kicking off this discussion. I was pretty excited about it but put it in the “too hard” basket for now – at least for someone with my skill set. And not completely confident to supervise a PhD on this (e.g. implementing the GFDL iceberg module in CICE) without some dedicated input from the GFDL side (which I haven’t explored), or at least someone within the CICE consortium.

To draw an analogy with @dpath2o 's project – for implementing the new fast ice parameterisation using coastal drag (from MITgcm) in CICE, we will collaborate closely with J-F Lemieux and Mathiue Plante from ECCC, who are both very close to the CICE code, and have collaborated with the original authors of the MITgcm code. I think this kind of arrangement is what would be needed for iceberg advection/grounding in CICE, but I haven’t explored or spent much time on it.

Without saying too much, an international collaborator is preparing a proposal on a prognostic iceberg advection/grounding model in a non-CICE (and non-SIS2!) model - basically what I’m saying is there might be good non-SIS2 options “soon”, but still not CICE.

Maybe @anton could be best placed to drum up interest for this within the Consortium? It would seem to be that CICE should have something like this to keep its reputation of being at the cutting edge :slight_smile: