A dark layer inside a sediment core from the seabed off West Antarctica contained something that should feel impossible: a dense network of fossil roots, along with pollen and spores from a temperate rainforest that grew near the South Pole roughly 90 million years ago.
This was not Antarctica briefly drifting through the tropics. The continent was already centred near the pole during the mid-Cretaceous. The forest endured a polar night lasting about four months, yet its reconstructed climate was mild enough for swampy ground, conifers, ferns and other vegetation.
I find the root network more persuasive than any artist’s reconstruction. It is not merely evidence that leaves or pollen blew in from somewhere warmer. The soil still holds the architecture of plants that grew there.
The forest was found under the sea
The evidence came from core PS104_20-2, recovered during a 2017 expedition aboard the German research icebreaker Polarstern. A seabed drill worked in the Amundsen Sea near Pine Island and Thwaites glaciers, in water almost a kilometre deep.
When the researchers examined one unusual section with X-ray computed tomography, they saw an intact network of roots running through fine-grained soil. Microscopic analysis then revealed abundant pollen and spores. Chemical indicators preserved in the sediment supplied additional clues about temperature and rainfall.
The team reported the result in a 2020 paper in Nature. The core was dated to the Turonian-Santonian interval, between about 92 and 83 million years ago, and its plant remains were consistent with a temperate lowland rainforest at a palaeolatitude near 82 degrees south.
This is one exceptional core, not a complete map of Cretaceous Antarctica. It records one location and interval, and the climate values are reconstructions rather than thermometer readings. What makes it valuable is the combination of independent evidence preserved together: roots, soil, pollen, spores and organic temperature markers.
A South Pole without an ice sheet
The reconstructed mean annual air temperature was about 12 degrees Celsius. Summer months averaged around 19 degrees, while river and swamp water reached roughly 20 degrees. Estimated precipitation resembled modern Wales.
Those numbers do not describe a tropical Amazon beside the pole. “Temperate rainforest” is the more accurate term. The forest was likely dense, wet and rich in conifers and ferns, but it grew within a light cycle no modern rainforest experiences.
For a third of every year, the Sun did not rise. Plants would have had a productive summer followed by a long dark season. The core cannot show every biological strategy that made this possible, but Cretaceous polar forests elsewhere suggest that high-latitude plants were adapted to extreme seasonality.
Climate modelling in the 2020 study found that reproducing the reconstructed Antarctic warmth required a vegetated continent without a major ice sheet and atmospheric carbon dioxide between about 1,120 and 1,680 parts per million. That range is model-dependent, and estimates of Cretaceous carbon dioxide carry uncertainty. It nevertheless shows how difficult the forest is to reconcile with a cold, ice-covered South Pole.
A grain of amber added another piece
The same core later yielded Antarctica’s first reported amber. In a 2024 paper in Antarctic Science, researchers described tiny fragments extracted from a layer of lignite. The material was named Pine Island amber after the region where the core was drilled.
Amber begins as tree resin. Its presence therefore offers direct evidence that resin-producing trees lived in this environment. The fragments were minute, about a millimetre before further preparation, and they did not reveal a spectacular trapped insect. Some appeared to retain possible remnants of tree bark and signs of resin flow associated with damage.
That modest result is useful. It strengthens the forest reconstruction without pretending the core is a perfect time capsule. The amber may eventually preserve microscopic traces of the ecosystem, but researchers are still working out what is inside it.
The word “tree” hides how different the forest was
I recently wrote that a tree is not one biological family but a growth form evolved by many lineages. That distinction matters here. Calling this an Antarctic forest can make us picture familiar modern woods relocated south. Its species and ecological relationships belonged to another world.
Flowering plants were diversifying during the Cretaceous, but conifers and ferns remained important. Antarctica was connected differently to the fragments of Gondwana, ocean circulation was different, and global sea level was far higher. The forest was not a modern ecosystem waiting beneath the ice. It was an ecosystem built under Cretaceous conditions.
A core that connects land, ocean and climate
The route by which the evidence reached us is almost as interesting as the forest. Soil formed on land, was buried and preserved through enormous geological change, and now lies beneath the Antarctic continental shelf. A ship drilled through the seabed to recover a narrow cylinder containing that history.
In an earlier article, I described the named oceans as one connected system moving heat around Earth. The Antarctic rainforest is another reminder that land, ocean, atmosphere and ice cannot be understood as separate stories. Change one strongly enough and the geography of life changes with it.
The most surprising thing about this vanished forest is not simply that Antarctica was warm. It is that a recognisable soil surface once sat close to the South Pole, threaded with roots and rained on beneath a sky that disappeared for months each year.
Today that landscape is inaccessible beneath ice, rock and sea. Yet a dark strip of core kept enough of it intact for the forest to become visible again.
