Science: An entirely credible catastrophe: Peter Taylor considers the potentially disastrous consequences of a power failure at Sellafield

The fate of British Nuclear Fuels' controversial Thermal Oxide Reprocessing Plant (Thorp) at Sellafield in Cumbria will be decided by the Government before the end of the year.

The pounds 2.85bn plant has been standing idle since it was completed in March 1992, while the Government conducted two rounds of public consultations over the economics of reprocessing, the need for Thorp, its impact on the proliferation of nuclear weapons and its effect on health and the environment. But surprisingly little attention has focused on the possibility of a catastrophic release of radioactivity from the plant. This category of Large Consequence Low Probability Accident was once referred to by the nuclear industry as beyond the 'maximum credible'. Chernobyl made such big events credible, but they have been adequately described in public only for PWR power plants, not for reprocessing of spent nuclear fuel.

But there is a potential for a catastrophic release of radioactivity associated with Thorp. It lies with the storage of hot liquid waste. This material is highly radioactive, dissolved in nitric acid, and kept in huge stainless steel tanks the size of a semi-detached house. The tanks are surrounded by several metres of concrete to attenuate the intense radiation. The concrete acts as an insulating blanket while radioactive decay within the waste generates several megawatts of heat, which is removed by massive cooling coils.

But the system is not fail-safe. If electricity or water were lost, the heat would build up, the water would boil and intense radioactivity would be vented - so much so that the site would have to be evacuated. Once the water had evaporated, the remaining radioactive salt would reach temperatures that would melt steel and vaporise concrete. To reach the point of no return - when the system starts venting - could take as little as 12 hours or as long as two days, depending on the age of the liquors. If one tank were to blow, forcing an evacuation of the site, they could all fail. Venting of radioactivity would then persist for days as the winds carried radioactive material for several hundred miles.

This potential was discussed at the Windscale Inquiry in 1977, when Justice Parker agreed with BNFL that, because there were no 'credible' circumstances under which the cooling could be irreparably lost, he needn't go into the consequences. Dr Gordon Thompson, a nuclear engineer, now with his own institute in the United States, warned that the system was not proof against terrorist attack, aircraft impact or severe earthquake. He also warned that electrical faults could lead to a general system failure. In 1979 he was proved right when similar tanks at Cap de la Hague in France had all their electrical circuits blown, and it took four hours to restore cooling.

Also in 1979, Dr Thompson acted as consultant to the West German government, which was considering its own reprocessing plant. On seeing the scale of the consequences for Central Europe of just one tank failure running its full course (and there are about a dozen tanks: if one goes they all go), the German government ruled that the technology would not be licensed until it could be made fail-safe.

The German plant was abandoned on the grounds of cost. But, had it been built, the Wackersdorf plant would have used 'passive cooling' systems - which would have continued to function even if all power to the site had failed. Thorp's tanks do not.

BNFL knew about the passive cooling option, but had decided on active cooling because the alternative was vastly more expensive - 'several tens of millions per tank'. Of course, the active cooling systems are not fail-safe, but, it is argued, they meet the design criteria for safe installations (redundancy in the cooling coils, several sources of power, water, and so on).

A further check with Her Majesty's Chief Inspector of Nuclear Installations revealed that he did not immediately know if the NII had ever asked BNFL about passive systems. One member of the Government's Advisory Committee on the Safety of Nuclear Installations admitted he knew nothing of the highly active waste tanks, their safety systems, or the consequences if they went wrong. The NII later pointed out that the technology for the tanks had been decided nearly 20 years ago and that, following the safety assessment done at that time, the issue had not been revisited.

During the recent public consultations over Thorp, Her Majesty's Pollution Inspectorate (the body responsible for granting the final licence to permit Thorp to operate) referred the question on to the NII, mentioning the vulnerability of the tanks to terrorist attack. The NII responded: 'This was a matter for the President of the Board of Trade.'

One hopes that someone somewhere has studied the possibility of such a release and measured its consequences - and that Mr Heseltine has considered them.

This country is about to license an installation that has the capacity, were it to fail irreparably, of paralysing half of Britain, causing billions of pounds' worth of damage to agriculture and lost industrial production and necessitating the evacuation of millions of people.

The risk could be avoided, but at a cost to BNFL probably of the same order as its projected profit of pounds 500m. The German Thorp had been designed to avoid that risk, but then, the German parliament had had the advantage of full and open analysis.

The writer is a member of Terramares, a group of independent scientists on terrestrial and marine ecosystems.

(Photograph omitted)