News of the biggest demonstration ever against nuclear power in Germany is coming in as I write. Chancellor Merkel’s reversal of her previous policy of extending the lives of nuclear power stations is seen more as a desperate attempt to appease voters than a change of heart. However, most commentators are not surprised by the move. In the UK the government is remaining issuing the usual vague reassurances and energy minister Chris Huhne has said that “knee-jerk reactions” seen in other countries are “not the right basis for British policy”. He went on to say that the coalition government will continue to “envisage a role for new nuclear” but that they will “have to put an emphasis on safety."

However, assessing safety or the risk of something going wrong is not easy, especially in a complex system composed of many parts and numerous operations subject to uncontrollable external forces. Generally, risk is associated with two factors – the likelihood of an unwanted event and the severity of the outcome. So, even if events are unlikely if the effects are very large then they are risky. However, no specific method exists for estimating the vulnerabilities of nuclear power stations to hardware failures and human faults in design and operation. The Japanese government clearly thought that the risks of building nuclear power stations in an earthquake zone could be managed by good design and adherence to procedures and they accepted the risk. However, the more complex the system the more chances there are of something going wrong and the unlikely or the unforeseeable does happen.

Other uncertainties concern people and arouse suspicion. The lack of information at the time of nuclear incidents causes distress and generates a feeling of helplessness and the huge differences in estimates of the resulting casualties cause confusion, anger and suffering for many years after the event. Consider, for example, the explosion at Chernobyl that resulted in the spread of radioactive material across Europe twenty five years ago next month. Calculations of the impact on human health made some 20 years later ranged from the Chernobyl Forum’s initial estimate of 4,000 deaths (later revised to 9,000) to the 16,000 obtained from modelling at the University of South Carolina, to 30-60,000 from an analysis made by nuclear scientists in the UK, to Greenpeace’s estimate of 100,000. An even more recent report published by the New York Academy of Sciences and based on the work of three Russian doctors, has concluded that, up to 2004, 985,000 people died as a result of the catastrophe. We will never know the exact number of fatalities because they nearly all occur sometime after the event and cannot easily be directly connected to it.

The fatalities from Fukushima are unlikely to be as high as for Chernobyl but there will be some- perhaps many. However, governments committed to new nuclear power programmes, such as the UK, will reassure us, saying that things were not that bad, that there were problems with the specific design and positioning of Fukushima, and that such a thing can’t happen again. It is up to us though to decide whether we really are prepared to take the continued risk of going through a nuclear nightmare. Can we rely on an industry that is so complex that there is no one person that knows how it all works? Can we continue to live with the potential risk of occasional nuclear catastrophe? And should we be imposing this risk on future generations?