Antarctica developed a continental-scale ice sheet 34 million years ago while Earth was about 5°C warmer than today — helped by slow-moving mantle waves that had spent more than 100 million years lifting its interior high enough for permanent ice to survive.

Antarctica did not wait for Earth to become as cold as it is today. At the Eocene-Oligocene transition about 34 million years ago, ice expanded across East Antarctica while global temperature was still roughly five degrees Celsius above the modern level. The Arctic remained largely ice-free for almost another 30 million years.

Falling atmospheric carbon dioxide remains the principal trigger for that change. A 2026 study in Science adds a much slower preparation beneath the continent: mantle processes linked to the Jurassic breakup of Antarctica and Africa gradually lifted East Antarctica into a plateau crowned by mountains. The extra height allowed snow to outlast summer and gave ice a place to begin accumulating.

The result is a chain of modelled cause and effect, not a direct film of ancient geology. Researchers integrated geodynamic and topographic reconstructions with energy-balance and ice-sheet models. Their simulations show that uplift could have moved the landscape across a glaciation threshold long before global climate reached modern temperatures.

A warm world crossed into an icehouse

The transition near 34 million years ago was one of the largest climate reorganisations of the Cenozoic Era. Earth moved from the warm greenhouse conditions of the Eocene into the colder Oligocene. A continental-scale East Antarctic Ice Sheet assembled, sea level fell sharply and bright ice began feeding back on the global climate.

The timing needs care. Geological evidence indicates that isolated and unstable glaciers appeared in Antarctica before the main transition. A community reconstruction of Cenozoic carbon dioxide places continent-wide glaciation at about 33.9 million years ago after roughly ten million years of generally declining CO₂. The event did not create Antarctica’s first patch of ice. It established a lasting ice sheet at continental scale.

The five-degree comparison used in the new study team’s GFZ research summary refers to global climate, not the temperature at a single Antarctic site. Estimates of ancient global temperature carry uncertainty, but the central puzzle remains: a huge southern ice sheet appeared in a world warmer than the present.

The delayed northern response sharpens that puzzle. The Arctic did not gain comparable permanent ice until almost 30 million years later, despite experiencing the same broad change in greenhouse forcing. Antarctica had a continental landmass centred on the pole and, in the new reconstruction, a growing high interior. The asymmetry suggests that global temperature alone cannot explain when each pole crossed its glaciation threshold.

Continental breakup launched a very slow mantle signal

The proposed preparation began during the Jurassic, when Antarctica and Africa were separating as Gondwana broke apart. Stretching and rifting altered the thick continental plate and the mantle beneath it. According to the model, dense material at the plate’s underside detached in coordinated drips.

That loss of dense material made the surface above more buoyant, rather as unloading weight allows an object to rise. The detachment pattern moved inward from the rifted margins. The researchers call the propagating process a mantle wave. It is not a seismic wave racing through rock in minutes. It unfolds across tens of millions of years and leaves a sequence of uplift at the surface.

The study’s accepted-manuscript record describes the mechanism as a mantle-surface feedback rooted in continental breakup. That origin matters because the ice sheet’s eventual birth would then depend partly on tectonic events that began more than 100 million years earlier.

The long delay is a feature of the mechanism. Continental breakup did not immediately create an ice-ready plateau. Instead, the disturbance was followed by deep material reorganising and uplift migrating through the interior. Rates that would be imperceptible across a human lifetime can reshape an entire region when they operate for tens of millions of years.

Two kilometres became the critical height

The model reconstructs an East Antarctic coastal escarpment, high interior plateau and the Gamburtsev Mountains now buried beneath the ice. Before about 50 million years ago, most of the Gamburtsev landscape in the reconstruction stood below 1.5 kilometres. By around 45 million years ago, broad areas had risen beyond a critical elevation close to two kilometres.

Height changes the annual snow budget. Air is generally colder higher above sea level. A modest rise can separate snow that disappears during summer from snow that survives into the next winter. Once a persistent white surface forms, it reflects more solar energy than dark ground and cools its surroundings further.

By 34 million years ago, the team estimates that almost half of the Gamburtsev range stood above two kilometres. The resulting ice-albedo feedback lowered global temperature by about one degree Celsius in the simulations. Colder, drier air then reduced the warming supplied by atmospheric water vapour, assisting further expansion.

The mountains were therefore not a substitute for cooling. They were a high-elevation seedbed that allowed the first durable ice to appear under conditions that would have melted summer snow across lower terrain.

The team connected models from the mantle to the ice

No drill can recover a continuous record of East Antarctic elevation extending back through the Jurassic. The interior is hidden beneath kilometres of moving ice, and erosion has modified the old landscape. The researchers instead tested whether several independent model components could produce a consistent history.

Geodynamic simulations reconstructed the breakup-driven mantle processes. Landscape models translated deep uplift into changing surface relief. Energy-balance calculations estimated how latitude, elevation, snow and reflected sunlight affected temperature. An ice-sheet model then tested where ice could nucleate and how far it could spread.

The key comparison is counterfactual. A low version of East Antarctica did not support the same early glaciation under relatively mild climates. The uplifted version created cold highlands where mountain glaciers could persist and later merge as global cooling intensified.

Agreement within this modelling chain does not make every ancient contour certain. Each component carries assumptions about mantle viscosity, erosion, past atmospheric conditions and the way ice responds to bedrock. The value of the experiment is narrower but important: the breakup-driven uplift is physically capable of producing the height, location and timing needed to help seed the ice sheet.

Modern maps provide a check on the broad geometry rather than a photograph of the ancient surface. A NASA visualisation of Antarctic ice and bedrock shows why subglacial relief matters to present ice flow. Radar, gravity, seismic and satellite observations reveal mountains, basins and ridges, but their ancient elevations must still be reconstructed.

Carbon dioxide remains the primary trigger

The new mechanism does not demote greenhouse gases. The paper explicitly starts from the evidence that a critical fall in atmospheric CO₂ was primarily responsible for Antarctic glaciation. Uplift changed how much cooling was needed and where the first stable ice could form.

SpaceDaily previously reported the proxy and modelling evidence that declining carbon dioxide drove the greenhouse-to-icehouse transition. Only the simulations including lower CO₂ reproduced the cooling found in terrestrial temperature records. The new mantle study complements that result rather than overturning it.

Ocean gateways and circulation also evolved. A 2021 ocean-model study found that widening passages and weakening gyres could cool the Southern Ocean, altering the regional threshold for glaciation. Orbital variations determined how sunlight was distributed. Ice elevation, albedo and water vapour supplied feedbacks after growth began.

The ice sheet emerged from this combination. Falling CO₂ provided the global push toward colder climate. Tectonic uplift created unusually favourable ground in the south. Ocean and atmospheric feedbacks helped turn mountain ice into a continental sheet.

Ancient oxygen isotopes do not reduce the story to ice volume

Much of Antarctic climate history is inferred from oxygen isotopes in the shells of marine microorganisms. Those records respond both to the amount of water stored in land ice and to ocean temperature, which makes a large isotope change possible without an equally large change in ice volume.

A recent SpaceDaily report described evidence that later Oligocene isotope swings largely reflected deep-ocean temperature. That result concerns climate variability after the first great glaciation, but it illustrates why multiple kinds of evidence are needed. Neither one proxy nor one model can reconstruct the whole transition.

The mantle-wave hypothesis earns attention because it connects the surviving topography, the timing of continental breakup, modelled uplift and the physical requirements of permanent snow. It remains a reconstruction with uncertain mantle properties, erosion histories and ancient boundary conditions.

Formation is not a forecast of modern survival

A warmer ancient Earth supporting Antarctic ice does not mean today’s ice sheet is safe under present warming. The East Antarctic interior is high, but much of Antarctica’s vulnerable ice meets the ocean or rests on bedrock below sea level. Warm water can attack floating shelves and grounding zones without waiting for summer air to melt the high plateau.

The rates are also incomparable. Mantle uplift prepared East Antarctica over more than 100 million years. Modern greenhouse forcing is changing climate over centuries. A slowly raised landscape can determine where an ice sheet first becomes possible without protecting every part of that sheet from rapid later warming.

The study’s deeper lesson is that climate thresholds are partly written into geography. Processes far below Antarctica lifted the stage. Falling carbon dioxide supplied the cooling. Once snow could survive on the new high ground, ice and atmospheric feedbacks helped transform a mountain refuge into a continent-scale sheet.