Life › The age of mammals › Pleistocene ice ages
Antarctica freezes over
c. 34 million years ago· Event· Earth, climate & extinctions
Well establishedStrong fossil, genetic or geological evidence.
About 34 million years ago, a permanent ice sheet spread across Antarctica, shifting Earth into a colder state.
See it on the Life map →What happened
For most of the dinosaur age and early mammal age, Antarctica had forests and little or no permanent ice. About 34 million years ago, within a few hundred thousand years, a continent-wide ice sheet formed. Falling carbon dioxide was probably the main trigger, helped as ocean gateways opened to South America and Australia, letting a cold current circle the continent. Global temperatures dropped, and many European mammals were replaced in a turnover called the Grande Coupure ('great break').
Why it matters
It moved Earth from a greenhouse world into the icehouse state that continues today, setting up later ice ages and drier, more seasonal climates.
Think of it like
Like a thermostat crossing a set point: once carbon dioxide fell below a threshold, ice spread fast and sustained itself by reflecting sunlight.
The evidence (fossils, genes, rocks)
A sudden jump in heavy oxygen in the shells of tiny deep-sea organisms (ice locks up the light form), iceberg-dropped debris in ocean sediments, and fossil leaves from earlier Antarctic forests.
Still debated
How much the ocean gateways mattered compared with falling carbon dioxide is debated, as are the causes of that carbon dioxide decline.
Inside this (1)
Connections (3)
▸Antarctica freezes over was an early step toward Pleistocene ice agesWell established
The connection
When a permanent ice sheet formed on Antarctica about 34 million years ago, Earth left the ice-free hothouse of the dinosaurs and early mammals. Cooling continued in steps as carbon dioxide fell: Greenland's ice grew, and by about 2.7 million years ago ice sheets could spread across the northern continents, starting the Pleistocene cycles.
When and how
Antarctica, at the Eocene–Oligocene boundary about 34 million years ago; northern ice sheets in North America and Europe from about 2.7 million years ago.
The evidence
Deep-sea cores show a jump in oxygen isotopes at about 34 million years and a stepwise trend toward the ice ages; rock debris dropped by icebergs marks when ice reached the sea.
Debates and open questions
How much ocean gateways opening around Antarctica mattered compared with falling carbon dioxide, and why large northern ice sheets took some 30 million years longer to form.
▸India hits Asia: the Himalayas rise may have cooled the climate toward Antarctica freezes overDebated
The connection
In 1992 Maureen Raymo and William Ruddiman proposed that the rising Himalayas and Tibet exposed vast amounts of fresh rock. Rain and chemical weathering of that rock pull carbon dioxide from the air and lock it into ocean sediments. Over millions of years, that drawdown may have helped cool Earth enough for Antarctic ice.
When and how
Himalayas and Tibet from about 50 million years ago, with Antarctic ice forming about 34 million years ago; hypothesis published in Nature, 1992.
How sure are we?
Ocean records show falling carbon dioxide and strontium isotope changes consistent with more weathering, but the timing and size of the effect are uncertain.
Debates and open questions
Critics note Earth's natural carbon thermostat should offset extra weathering, and others credit ocean gateways or volcanic changes. The timing of uplift itself is debated.
▸Antarctica freezes over helped dry the world for Grasslands spreadStrong evidence
The connection
Growing ice locked up water and cooled the oceans, which then evaporated less and supplied less rain. Over the following tens of millions of years, cooler, drier and more seasonal climates spread across the continents, especially after further cooling about 14 million years ago. Forests thinned, and grasses took over the drier ground.
When and how
Global, from about 34 million years ago, with a further cooling step about 14 million years ago; grasslands spread across the Americas, Africa and Asia.
How sure are we?
Ocean-core records of cooling line up broadly with fossil soils, pollen and phytoliths showing open habitats spreading, though not in lockstep.
Debates and open questions
Grasslands spread at different times on different continents, sometimes long after cooling steps, so the link is indirect and its details are debated.