A vault in the permafrost

Between the airport at Longyearbyen and the open Arctic there is a concrete wedge poking out of a mountainside, lit at night by a single artwork of mirrors and prisms. Behind its door a tunnel runs about a hundred and twenty metres into the sandstone, to three chambers kept at around minus eighteen degrees. The Svalbard Global Seed Vault opened in 2008, and it holds duplicates of seed samples from gene banks all over the world: well over a million of them, from rice and wheat to beans and sorghum. It has been nicknamed the Doomsday Vault, a label its staff do not much like.

The idea is more modest than the name suggests. Gene banks, the places that keep the world's crop varieties alive, are vulnerable. Some are in war zones, some are poorly funded, and many have been wrecked by floods, fires or mismanagement. The vault is an insurance policy. Each country or institution that deposits seeds keeps ownership of them, and only the depositor can ask for them back. Norway, which pays for the building, has no claim on a single seed.

In 2015 the insurance was called in for the first time. A research centre in Aleppo that held one of the world's most important collections of seeds adapted to dry climates had been cut off by the war in Syria, and its scientists asked for their duplicates. The boxes travelled from the Arctic to Morocco and Lebanon, where the seeds were planted and harvested, and new copies were later sent back to Svalbard. It was a small, almost dull piece of logistics, and a convincing proof that the scheme works.

The sceptics' favourite objection came true in a way nobody expected. In the unusually warm winter of 2016 and 2017, meltwater and heavy rain seeped into the entrance tunnel. The water never reached the seed chambers, and no sample was damaged, but the incident was embarrassing for a facility designed to survive almost anything, and the permafrost it relied on had proved less permanent than the planners had assumed. Norway spent millions on waterproofing, drainage and a new access tunnel.

Visiting it is, by all accounts, an anticlimax. The public cannot go in; the door opens only a few times a year, when new deposits arrive, and the seeds sit in sealed foil packets on plain metal shelving that could belong to any warehouse. Staff describe the place as cold and strangely quiet. What gives the vault its pull is not the sight of it but what it implies: that a species with a short memory has, once, thought several centuries ahead.

It also raises an awkward question about what exactly is being saved. Seeds kept in the cold are frozen in time, while the pests and diseases outside go on evolving; a wheat variety that resisted a fungus in 1980 may be defenceless against its descendants. This is why the vault is only half of the plan. Gene banks must keep growing and swapping their collections, and farmers must keep planting old varieties alongside new ones, so that the diversity in the freezer is matched by diversity in the field. A vault cannot keep a crop alive; it can only keep the option open.

The episode is instructive precisely because it was undramatic. The vault was never meant to be a fortress against the end of the world, still less a monument to human cleverness. Its purpose is simpler: to ensure that when something goes wrong, as it will, a copy exists somewhere else. The Arctic merely makes the copy cheap to keep, and, as the engineers have since learned, no substitute for being checked.