Ice core detectives

[727Sky]

https://scienceblog.com/what-antarctic-i...ion-years/

Three Million Years
March 18, 2026

Key Takeaways

• Ice cores from Allan Hills provide insights into atmospheric conditions 3.1 million years ago, revealing minimal changes in CO2 and methane during critical climate transitions.
  
• The research shows that ocean heat and circulation patterns also changed, indicating factors beyond CO2 impacted ice age intensification.
  
• Key hypotheses suggest ice sheet reflectivity, ocean circulation shifts, and vegetation changes contributed to climate shifts, narrowing the focus beyond greenhouse gases.
  
• While the Allan Hills cores are fragmented and complex, they represent a significant advance in understanding ancient climates.
  
• Researchers aim to extend the records at Allan Hills even further, potentially revealing insights from the last warm period before ice ages began.
  

The ice at Allan Hills, a wind-scoured margin of the East Antarctic ice sheet, is not where it started. Glacial flow has dragged it from the interior over millions of years and deposited it close to the surface, where katabatic winds strip away fresh snow almost as quickly as it falls. The result is a natural archive, one that researchers have now read to answer a question that has nagged at climate scientists for decades: what was the atmosphere doing before the ice age intensified?
Two papers published today in Nature, both drawing on shallow ice cores drilled at Allan Hills by the NSF-funded Center for Oldest Ice Exploration (COLDEX), extend the direct record of atmospheric greenhouse gases and ocean heat content back to 3.1 million years ago. The combined picture they present is, in a specific and somewhat uncomfortable sense, surprising.
The first paper, led by Julia Marks-Peterson, a doctoral student at Oregon State University, measured carbon dioxide and methane concentrations in air trapped in ice spanning 3.1 to 0.5 million years before present. Over that period, Earth’s climate underwent dramatic transformation: the Northern Hemisphere glaciations intensified, ice sheets expanded, and glacial cycles shifted from roughly 41,000-year rhythms to much more prolonged 100,000-year cycles in what is known as the Mid-Pleistocene Transition. The expectation, reasonable on first principles, was that CO2 must have fallen substantially to drive all that cooling. It had not. CO2 dropped by only around 20 parts per million between 2.9 and 1.2 million years ago, and then stabilised. Methane barely changed at all.
The oceans told a similar story.
The second paper, led by Sarah Shackleton of Woods Hole Oceanographic Institution, reconstructed mean ocean temperature from noble gas ratios, specifically xenon to krypton, in the same ice cores, a technique that works because the solubility of these gases in seawater is exquisitely temperature-dependent. Ocean heat did fall markedly around 2.7 million years ago, coinciding with the Plio-Pleistocene Transition, and then held broadly steady through the Mid-Pleistocene Transition itself. Compared against global surface temperature records, the new data reveal important differences in how heat was distributed between the ocean surface and the deep, pointing toward significant reorganisation of deep water formation and ocean circulation as glaciations intensified.
The implication is that CO2 alone cannot be the whole story of ice age intensification. Something else was changing too, plausibly the reflectivity of growing ice sheets, shifts in vegetation patterns, alterations in ocean circulation. Understanding which of these factors dominated, and how they interacted, is now a live and pressing research question.
“Our hope is that this work will refine our view of past warmer climates and sharpen our understanding of how different elements of the Earth system interact,” said Marks-Peterson.
The Allan Hills cores are not continuous archives. Their stratigraphy is complex, the layers folded and disordered by millions of years of ice flow, and what the measurements capture are weighted averages rather than precise reconstructions of individual glacial cycles. That is a genuine limitation. But for the period between 800,000 and 3 million years ago, a span that was previously almost entirely opaque, even averages are a significant advance.
COLDEX researchers recently found ice at Allan Hills as old as 6 million years and are now working toward extending these records further, potentially deep enough to reach the last warm period before the ice ages began in earnest.
DOI: 10.1038/s41586-025-10032-y / 10.1038/s41586-026-10116-3
Frequently Asked Questions
[b]If CO2 barely changed, what caused the ice ages to intensify? [/b]
That is now precisely the question the data opens up. The current leading hypotheses involve changes in Earth’s reflectivity as ice sheets expanded (a positive feedback that doesn’t require more CO2), shifts in ocean circulation that redistributed heat between the surface and deep ocean, and changes in vegetation and dust. The new ice core data rule out a large CO2 decline as the primary driver, which narrows the field considerably.
[b]Why is Allan Hills special for finding ancient ice? [/b]
Most Antarctic ice cores require drilling to depths of over two kilometres to reach ice hundreds of thousands of years old. At Allan Hills, peculiarities of ice flow bring very old layers close to the surface, where they can be sampled at depths of only 100-200 metres. The trade-off is that the record is fragmented rather than continuous, discrete time snapshots rather than a year-by-year archive.
[b]How do scientists measure ancient atmospheric CO2 from ice? [/b]
Air bubbles become trapped in snow as it compresses into ice, preserving samples of the atmosphere from when the snow fell. By drilling out ice cores and carefully melting them under controlled conditions, researchers can extract and analyse the gas mixtures inside. The Allan Hills samples required additional correction using carbon isotope ratios because microbial activity had partially altered the oldest samples.