When the trees came back to Europe after the Ice Age, town by town

14,000 years ago

Ice over Scandinavia, and a Europe with few trees#

The ice sheet still covered Scandinavia. The sea stood so low that Britain was joined to the continent across land that is now the floor of the North Sea.

Around the cores that reach back this far, trees covered a median 22% of the land, according to a reconstruction from the same kind of pollen by Sweeney, Harrison and Vander Linden (2025). At the height of the forest, 6.6 thousand years ago, they covered 68%.

On the map, that tree cover is the speckled canopy. Its colour says how many of the nine trees grew in the region: pale where one or two, deep green where all nine. Each small dot is one core. Where few cores lie near, the land is hatched. All dates are calendar years before 1950.

14,000 to 11,000 years ago

Birch and pine came first#

Birch and pine seeds travel far on the wind, and both trees stand the cold. Of the 422 cores with a sample from 11,000 years ago, 72% show birch growing in their region and 72% show pine.

The trees that need warmer summers were still in the south: in Italy, the Balkans and the Carpathians, where the speckle is already greener.

11,500 to 9,000 years ago

Hazel and elm spread fastest#

In about two thousand years the land where hazel grew went from 1,541,000 km² to 4,302,000 km² on this map, out to Ireland and southern Sweden; elm's went from 1,411,000 to 4,242,000 km².

Thomas Giesecke and colleagues, mapping the European Pollen Database in 2017, measured spread of at least 1,000 metres a year for hazel and elm, against about 400 for lime, oak and beech.

10,000 to 6,500 years ago

Elm, oak and lime: the forest at its fullest#

Elm, oak and lime followed close behind hazel. By 6,000 years ago each of the three grew across four to five million square kilometres of the map, and the cores show the most tree cover of the whole period: a median 68% about 6.6 thousand years ago.

This is the forest that most of Europe had when the first farmers arrived.

Britain, 11,000 to 8,000 years ago

Did hazel, elm and oak reach Britain before the sea cut it off?#

This site's map of Doggerland has Britain cut off between 9,250 and 9,000 years ago; the models of Sturt, Garrow and Bradley (2013) put the break later, between 8,000 and 7,500.

Around Canterbury the pollen map has hazel from about 10,750 years ago, elm from 10,250 and oak from 10,000, all before the earlier date for the break. Lime comes about 8,500 years ago, between the two dates.

How sure is that order? For British cores, one date in five misses by more than 1,150 to 1,600 years. Hazel's lead of 1,500 years is larger than that; elm's (1,000) and oak's (750) are not, so for those two the order is uncertain.

Beech came much later. In the English records its pollen passes 0.5% only from about 3,750 years ago, and 1% from 2,500. A 2005 study of the New Forest proposed a few beech trees in southern Britain around 8,000 years ago. Spruce and fir never came back on their own.

Spruce

Spruce probably waited out the Ice Age in the Carpathians#

Around Cluj the pollen shows spruce in the region from about 13,500 years ago, around Sibiu from 13,000, while ice still covered Sweden. Enikő Magyari and colleagues (2011), who found spruce leaf pores in a Retezat lake from about 13,900 years ago, think spruce was probably common at lower elevations through the last glacial, though absent from that lake's shore for part of it.

It reached Scandinavia late, and from the north-east. Around Oslo the map has it only from about 1,000 years ago. In central Sweden, a spruce needle in the mud of Klotjärnen lake is 2,590 years old, 340 years before the map's date there.

Beech

Beech came last across most of Europe#

UNESCO, listing the ancient beech forests of Europe, says beech "started expanding its range" from southern refuges around 11,000 years ago. The pollen map has it then on 132,000 km², mostly in Italy and the Balkans.

In 72% of the 4,847 towns where the map dates beech, it is the last of the nine to appear. With every millennium the map shows it in more places. Around Cluj it arrives about 4,750 years ago, roughly when the Great Pyramid of Giza was built, about 4,500 years ago; around Brașov about 3,750. Around Paris and Brussels beech comes about 2,750 years ago, around Copenhagen 1,750. Around Brașov and Sibiu, fir came later still: the map has it from about 2,000 and 2,750 years ago.

By 2,000 years ago the land where beech grew was 22 times what it had been 11,000 years ago. For Carpathian records the map misses beech's arrival by a median 250 years.

6,500 years ago to 250 years ago

After 6,000 years ago the forest thinned#

Tree cover around the cores fell from a median 68% to 48%, over the same millennia in which farming spread across Europe. Lime and elm changed most: on the map, the land where lime passes its threshold shrank by 86% after 6,000 years ago, and elm's by 75%. That can mean fewer trees, including a decline large enough for the pollen to fall below the threshold; Giesecke and colleagues note this for lime.

Elm pollen fell sharply across Britain and Ireland; a review of the dated sites (Parker and others, 2002) puts the fall between about 6,350 and 5,300 years ago, its pooled range at one standard deviation. The causes discussed are a disease like today's Dutch elm disease, the first farmers, and climate.

How far to trust it

Where this map is wrong#

Pollen flies, so a date here says when a tree's pollen became common enough to suggest it grew in the region around the cores, not that it stood by your door. To test it: 44 lakes and bogs in Europe where needles, seeds or wood of these trees were found and dated. In the 48 cases that can be compared (the map has an arrival there, and the oldest remains are not simply the bottom of the record), only 5 put the tree more than 1,000 years before the map's date; in 18 the two agreed within 1,000 years; in 25 the remains came later, by a median 1,500 years. Such remains are rare finds, so a late find says little; an early one suggests the map may be late there, and among these comparisons that was rare.

Each core was also left out in turn and its date estimated from cores more than 25 km away. The median difference from the core's own date was 500 to 750 years, depending on the tree; four differences in five were under 1,500 to 2,250.

Where few cores lie near, the land is hatched and the town search gives no date. That is most of the Romanian Plain: within 150 km of Bucharest there is 1 record, Mire Garvan, 74 km away.