Where nuclear waste will finally go

The engineering has not been the hard part. The Finnish and Swedish KBS-3 method seals fuel into copper canisters, surrounds them with bentonite clay that swells when wet, and places them in tunnels hundreds of metres down in rock that has been geologically stable for a billion years. It has been studied for four decades and no serious technical objection has survived.

What has failed is consent. Repository siting has collapsed politically in more countries than it has succeeded in. Yucca Mountain absorbed two decades and billions of dollars, was selected by an act of Congress rather than by a search, never received a single fuel assembly, and lost its funding in 2010. Germany spent thirty years on Gorleben before abandoning it and restarting the search from a blank map.

The programmes that worked did the opposite. Finland and Sweden spent decades building local agreement before they moved rock, and both host municipalities volunteered. Canada's process ran fifteen years and ended with a willing community and a willing First Nation. What those three have in common is not geology, which is unremarkable in each case, but that nobody was told where the waste was going.

Most waste is not spent fuel, and most of it already has somewhere to go. Drigg in Cumbria has been taking low level waste since 1959, Finland has been putting it underground since 1992, and Sweden's SFR sits fifty metres below the Baltic seabed. These places work, and they are routinely counted alongside spent fuel repositories in a way that makes the problem look closer to solved than it is. The Waste Isolation Pilot Plant in New Mexico is the only deep geological repository in the world receiving waste, and it takes transuranic waste from the weapons programme. It is proof the concept works, not proof the harder problem is finished.

Germany is the record of what failure costs. Asse II took 47,000 cubic metres of waste into a salt mine on the assumption that salt stays dry. Water entered, and the decision now is to retrieve all of it, at an estimated 4.7 billion euros merely to be ready to begin in 2033. Morsleben, 37,000 cubic metres, is being sealed with its contents in place, the opposite decision. Konrad, licensed in 2002, is still being built.