The Arctic could lose around one million square kilometres more winter sea ice than previously projected according to a new study that uses evidence from a warmer period thousands of years ago to improve the accuracy of climate models.
The findings published in Nature Geoscience could have implications for weather patterns across the United Kingdom, northern Europe and North America, as Arctic winter sea ice helps regulate the movement of heat and moisture between the ocean and atmosphere.
While summer sea ice decline has received considerable attention, winter ice loss remains less understood. Climate models have shown substantial differences in their winter projections, making it difficult to determine the likely scale and location of future ice retreat.
Researchers from the University of Cambridge and Mexico’s Centro de Investigación en Matemáticas analysed evidence from the mid Holocene around 6,000 years ago, when changes in Earth’s orbit contributed to warmer Arctic conditions and reduced sea ice cover.
Ancient algae provide clues
The research team compiled a large database of a biomarker known as IP₂₅, produced by algae associated with sea ice. By measuring this substance alongside other biomarkers from organisms living in open water, scientists estimated historical sea ice conditions using ocean sediment samples.
They compared these estimates with simulations from 27 climate models representing the same period. Using a statistical technique called Bayesian weighting, the researchers assigned greater weight to models that more accurately reproduced ancient sea ice conditions.
The adjusted model weightings were then applied to simulations of the recent past and future, helping researchers refine their projections of Arctic winter sea ice.
The analysis found that the models collectively tended to overestimate the amount of winter sea ice. After adjusting their weightings, the researchers eliminated this bias in historical simulations and lowered future estimates of winter ice coverage by approximately one million square kilometres, or about 386,000 square miles, compared with previous projections.
The approach also reduced the range of uncertainty in future projections by as much as 85 per cent, according to the researchers.
Crystal Fu of the University of Cambridge’s Scott Polar Research Institute, the study’s lead author, said evidence from past climates could help scientists produce more reliable estimates of where winter sea ice is likely to disappear in the coming decades.
Potential effects beyond the Arctic
Arctic winter sea ice influences the exchange of heat and moisture between the ocean and atmosphere, affecting weather conditions far beyond the polar region. Changes in ice cover have been associated with shifts in winter weather, including colder conditions and heavier snowfall in parts of northern Europe and North America.
The Arctic recorded its lowest winter maximum sea ice extent since modern satellite measurements began in March 2025, with the record reportedly matched again in 2026.
The researchers also identified indications that some wider climate impacts could be underestimated. Applying their revised model weightings to precipitation projections suggested that snowfall over southeast Greenland and rainfall in the Mediterranean region may be greater than earlier estimates indicated.
These potential changes could have consequences for the stability of Greenland’s ice sheet and flood risks in southern Europe, although the findings point to possible effects rather than establishing specific future outcomes.
The researchers acknowledged that the mid Holocene warming occurred under different conditions from the warming experienced today. However, they argued that evidence from the past remains useful because sea ice responds to warming regardless of whether its cause is natural variability or human activity.
The study demonstrates how ancient climate records can help improve modern climate projections. By combining evidence preserved in ocean sediments with climate model simulations, scientists hope to better understand the scale of future Arctic sea ice loss and its possible consequences for weather and climate around the world.