Linus Vogt
linusvogt.bsky.social
Linus Vogt
@linusvogt.bsky.social
Postdoc at NYU Courant Institute 🌱 Ocean heat, carbon, oxygen
📈 As CMIP6 models underestimate historical Antarctic sea ice extent on average, this leads to an upward correction of climate projections by 2100 under all considered scenarios:

- OHU +3–14%
- Cloud feedback +19–32%
- Global surface warming +6–7%
October 2, 2025 at 10:49 AM
Thus, the state of Antarctic sea ice today quantifies the potential for global climate change in the future.

More sea ice now → more warming later

This is evident as a strong correlation across CMIP6 models (and it also holds in CMIP5)
October 2, 2025 at 10:49 AM
The constraint is based on Antarctic sea ice extent 🧊

Models with more sea ice today also simulate colder Southern Ocean SSTs, a colder deep ocean, and more mid-latitude cloud cover.

In the future, these models lose more sea ice and clouds, amplifying SW cloud feedback, surface warming, and OHU.
October 2, 2025 at 10:49 AM
We further find that the previously assumed importance of the AMOC for OHU efficiency can be explained by an inter-hemispheric connection in stratification strength between these two regions.

As CMIP6 models are too stratified in precisely these regions on average, they tend to underestimate OHUE.
June 23, 2025 at 1:35 PM
We show that upper-ocean stratification inhibits OHU efficiency in the Southern Ocean and the subpolar North Atlantic.
This is because stratification in these regions is linked to large overturning cells: the Southern Ocean upper cell and the AMOC.
June 23, 2025 at 1:35 PM
In observations, the OHU efficiency in recent decades was around 0.6 W/m² per °C of global warming (Cael 2022, doi.org/10.1029/2022...).

However, current-generation climate models simulate a large range of values for this quantity, almost spanning a factor of 2:
June 23, 2025 at 1:35 PM