Â鶹ÒùÔº

A study published today in Earth System Dynamics provides a critical and previously underestimated connection between Antarctic sea ice, cloud cover, and global warming. This research is important because it shows that a greater extent of Antarctic sea ice today, compared to climate model predictions, means we can expect more significant global warming in the coming decades.

The study, led by Linus Vogt from Sorbonne University, utilized an emergent constraint based on data from 28 Earth system models and satellite observations from 1980 to 2020. This constraint allowed the team to reduce uncertainty in climate projections and provide improved estimates of key climate variables.

Their findings indicate that ocean heat uptake and the resulting thermal sea level rise by the year 2100 are projected to be 3–14% higher than the average from CMIP6, a leading collection of climate models. Furthermore, the projected cloud feedback is 19–31% stronger, which enhances climate sensitivity, and global surface warming is estimated to be 3–7% greater than previously thought.

The study found that the extent of Antarctic summer sea ice, which has been considered stable and only weakly connected to human-caused climate change, is a crucial indicator of the Southern Hemisphere's climate. Models that start with a higher, more accurate representation of pre-industrial sea ice levels simulate colder surface waters, colder deep ocean temperatures, and thicker cloud cover in the mid-latitudes.

These initial conditions then amplify warming responses under greenhouse gas forcing, meaning they lead to a more severe and accelerated warming effect than what was previously estimated. Essentially, the climate system's starting point makes it more sensitive to the impact of greenhouse gases.

"When we initially discovered this link between historical Antarctic sea ice and future global ocean heat uptake, we were surprised by the strength of the relationship. Antarctic sea ice covers less than 4% of the ocean's surface, so how could it be so strongly associated with global ocean warming?" says Vogt, who led the study at Sorbonne University in Paris, and is now based at New York University. "Only after a lot of analysis did we understand the full implications of the sea ice-ocean-atmosphere coupling which is responsible for these global changes."

This relationship isn't merely correlative: it is mechanistically explained through ocean-atmosphere feedback. Higher sea ice extent enhances cloud cover, which has a cooling effect overall by reducing incoming solar radiation. Greater sea ice loss in the coming decades is thus linked to larger reductions of clouds, stronger surface warming, and enhanced ocean heat uptake. As a result, the baseline state of sea ice and deep ocean temperatures in models effectively preconditions the magnitude of warming, cloud feedback, and heat uptake in the future.

"While it has long been known that accurately representing clouds is crucial for climate projections, our study highlights that it is equally important to also accurately simulate the surface and deep ocean circulation and its interaction with sea ice," says Jens Terhaar, a senior scientist at the Division of Climate and Environmental Â鶹ÒùÔºics at the University of Bern, who initiated the study at the Woods Hole Oceanographic Institution in the U.S..

Under future climate change scenarios, models with greater historical sea ice tend to lose more sea ice by 2100, contributing to stronger radiative feedback. This stronger feedback leads to a stronger atmospheric and oceanic warming, especially across the Southern Hemisphere.

Implications for policy and science

This study provides evidence that current models may be underestimating future warming and ocean heat storage. It shows that models tend to simulate a too warm Southern Ocean in the preindustrial state and therefore have too little warming potential. The findings also stress the importance of continued satellite monitoring and improved modeling of cloud processes and deep ocean hydrography, both of which significantly shape global climate projections.

The study warns that previous approaches, which relied on observed trends over limited timeframes, may have underestimated future warming due to their inability to capture systemic changes ("regime shifts") that are now becoming more evident, such as the record-low Antarctic sea ice extent in 2023. Furthermore, these older constraint methods relied on trends over short historical windows (e.g. 1980–2015), which are sensitive to internal natural variability and may thus not be representative of future climate change.

"Several high-profile studies have used temperature trends over recent decades in an attempt to constrain future warming," says Vogt. "However, we have now found that this approach can give misleading results. Accounting for the sea ice-related mechanism we identified leads to increased estimates of future ocean and atmospheric warming.

"This likely stronger warming calls for urgent action to reduce greenhouse gas emissions in order to avoid the increased heat waves, floods and ecosystem impacts associated with ocean warming."