Eldgjá eruption

The eruption broke out along the Eldgjá (“fire gorge”) fissures of the Katla volcanic system, partly beneath the Mýrdalsjökull ice cap. At least 16 explosive phases accompanied huge outpourings of basalt lava: Thordarson and colleagues estimate the magma erupted at about 19.6 km³ (dense-rock equivalent), more than in the Laki eruption of 1783–1784. The lava ran down river valleys to the sea and still covers some 780 km², and tephra fell over about 20,000 km² of Iceland. The date long remained uncertain: estimates from ice cores ranged from 933 to 938, and 934 was the year most often quoted, although 939 had been proposed on the strength of an Irish chronicle and adopted in a revised ice-core chronology. In 2018 Clive Oppenheimer and colleagues, using a Greenland ice core anchored to the precisely dated eruption of Paektu Mountain in late 946, confirmed and refined this date: the eruption began around the spring of 939 (though the first sulphur in the core may not mark its very first phase) and continued, at least episodically, until at least the autumn of 940; some volcanologists allow for activity lasting several more years.

No contemporary Icelandic account survives. The Book of Settlements, written down two centuries later, tells of settlers in the Álftaver district who fled when their land was overrun by an “earth fire”, and 12th-century sources call parts of the area a wasteland. Abroad, the Irish Chronicon Scotorum notes under 939 that the sun was the colour of blood from the beginning of one day to the middle of the next; the Italian Annales Casinates describe, probably in July 939, a sun with “neither light nor heat”; and Widukind of Corvey records blood-red sunlight and an exceptionally severe winter in 939–940 — although linking these entries with Eldgjá is disputed. Tree-ring reconstructions show the Northern Hemisphere summer of 940 cooler than the surrounding decades by about 0.7–1.5 °C, depending on the reconstruction, while in Central Europe, Scandinavia, Alaska, the Canadian Rockies and Central Asia summer temperatures fell as much as 2 °C below the 1961–1990 average; the Nile flood of 939 was unusually low. Harsh winters, locusts and famines in Europe, the Middle East and China in the early 940s may also be connected with the volcanic veil, but how far is uncertain, and some of these effects may belong to other eruptions of the period.

The eruption came within the lifetime of the first two or three generations of settlers; some of those brought to Iceland as children may have witnessed it. Oppenheimer and his co-authors argue that stanzas of Völuspá about the end of the pagan gods — “the sun starts to turn black… steam spurts up… flame flies high against heaven itself” — draw on memories of Eldgjá, and that the poem was meant to prepare minds for the conversion to Christianity, decided by the Althing in 999/1000. This remains a hypothesis: the volcanologist Sigurður Þórarinsson had read the poem’s volcanic imagery as referring to an eruption of Katla around 1000.