SaFT first! – Ocean constraints on winter sea ice cover
Will Hobbs
1. Winter Antarctic sea ice extent has historically shown relatively low variability compared to warmer seasons, but recent winters have shown very extreme low ice cover. As a result, there is significant interest in understanding the physical constraints on Antarctic sea ice cover.
In this work the ocean’s role in constraining maximum extent is explored, using ACCESS-OM2 and subsurface ocean observations.Using basic physical principles, the concept of Stability at Freezing Temperature (SaFT) is introduced, and then applied to calculate the total heat that must be lost to the atmosphere before sea ice can begin to form.
The northern limit of this thermal barrier shows a very close colocation with the winter sea ice edge in both ACCESS-OM2 and observations and proves that the vertical salinity gradient is the dominant control on winter sea ice extent. This method shows considerable skill in defining the winter sea ice extent, including the extreme low sea ice covers in 2023 and 2024, and demonstrates that those events were driven by ocean heat anomalies.
This research is relevant to polar Southern Ocean change (especially sea ice), and to model evaluations of Antarctic sea ice.
Please use this thread for discussion about this talk.
Aidan
(Aidan Heerdegen, ACCESS-NRI Release Team Lead)
2
Meant to say @willrhobbs this was a great talk! Loved it. Tthe model under-performance seems to have led to some genuinely interesting scientific insight.
@anton this is the talk I was referencing when I asked the question about the ice-berg transport scheme/parameterisation added to ESM1.6 by @dhb599. That this might help to pre-condition the ocean for sea-ice formation hence the reduced Weddell Sea polynya, but I defer to @willrhobbs on that.
Aww, thanks @Aidan, you say the nicest things! I didn’t say too much in the talk about OM2 performance (time constraints) but the paper does have a bit of comparison (and of course the fact that OM2 AIF doesn;t chapter recent sea ice loss is an indicator of the ocean as a driver.
There is a definite difference between the model and the obs. The model has stronger salinity stratification than obs, so that SaFT (the shaded area) is possible much further north than the ice edge; it’s heat content that stops the ice growing. In the obs/reanalysis, the limit is much more on whether SaFT is possible at all. I haven’t done that analysis for a coupled model, but happy to share the code library,
So, surface salinity biases are important, but I would have to dig a bit more to know whether iceberg melt distribution is enough to change that.
Aidan
(Aidan Heerdegen, ACCESS-NRI Release Team Lead)
4
My memory (and again @anton would know better) is that the iceberg melt scheme affects salinity and temperature, so it may be SaFT relevant.
Another point that I thought was interesting: the ESM1.6 1 degree configuration bottom water transports were largely unaffected, even though the massive polynya was reduced and I assume the associated open ocean convection? So it may be forming bottom water in a more sensible location which is a result.
Also the location of the iceberg melt is prescribed, so if there were better places to put it I am sure they would be interested to know.
This is all development that is planned to go into ACCESS-ESM1.6 CMIP7 fast-track, so it would be good to get it working as well as possible before the CMIP7 train gathers steam …
