First author

Much of Earth's crust has been formed over time by volcanoes at mid-ocean ridges on the sea floor. Geologists thought that the occurrence of explosive volcanic eruptions — such as that seen in 1980 at Mount St Helens in Oregon — would be rare, if not non-existent, along great tracts of these ridges owing to the extreme water pressure at depths of more than 3,000 metres. Such pressure both counteracts magma expansion and prevents the formation of steam. But while mapping and imaging parts of the Gakkel Ridge, which stretches across the eastern Arctic Basin, geophysicist Robert Sohn and an international team found multiple cratered volcanoes and evidence of an explosive eruption that may have thrown debris 1–2 kilometres up in the water (see page 1236). Sohn, based at the Woods Hole Oceanographic Institution in Massachusetts, tells Nature how such an event could occur.

How did you study the ocean floor beneath the Arctic ice?

We spent two weeks on the Swedish icebreaker Oden, drifting around in the ice floes above an area of a suspected eruption, where a swarm of magnitude 4–6 earthquakes was recorded in 1999. We passed over the same points 15–30 times, making low-noise sonar measurements. We also deployed a tethered vehicle, CAMPER, to take samples and video footage of the site.

How did you know an explosion had occurred?

On the video of the sea floor, we saw a coarse, black sediment covering absolutely everything. The only explanation for this that made sense was that it was pyroclastic material — bits of molten rock ejected during a volcanic explosion. We sampled the deposits and found fragments that provided conclusive evidence of explosive activity.

What mechanism would account for an explosion at that depth?

Nowhere near the amount of carbon dioxide gas that would be needed for an explosion at those depths has ever been found dissolved in basaltic magma. But CO2 can 'exsolve', or come out of solution, in lower concentrations and build up over time — and if it did so would probably become trapped at the top of the magma chamber. If a very large volume of CO2 built up, then an earthquake swarm could crack the chamber roof, which would be akin to tapping a keg, releasing the gas in a catastrophic explosion.

What would it have been like?

Tremendously energetic. There was probably a massive kill-off of deep-sea animals in the immediate area.