At a recent meeting on ‘Coastal Processes’ at Sizewell held at Snape Maltings on the of 29^th February Colin Taylor (Environment Manager (Marine), New Nuclear Build, EDF Energy) stated that they had 150 years of data for the coastal processes at Sizewell.

He also said that their ‘models’ can predict what has happened in the past. I shall not go into this any more in this blog, although it was a bit disconcerting that he said that their models had lots of problems with storm surges and that ‘waves were a bit tricky’.

He did assure me that they were confident about the 1/10,000 flood risk – again he called this 10^-5pa risk when it is in fact 10^-4pa risk.

What is the probability that we would see a 10^-4pa event in 150 years. For details of how to work out this sort of problem, see my previous post. So the probability of seeing this sort of event in 150 years is:

p₁₅₀ = 1-(1-p_yr)¹⁵⁰

This works out as a 1.5% chance of seeing such an event. So it is extremely unlikely that their 150 years of data would have covered such an event.

How many years of data would we need to have a reasonable probability of seeing such an event. Since Sizewell C (if built) will contain massive amounts of radioactive spent fuel (see here) we want to have quite a good chance of seeing such an event to test any model. Lets say a probability of 0.95 (95%). Let us call that probability p_t. Then rewriting the formula above

p_t = 1- (1-p_yr)^t

where t is the period of time that we are looking for. So we rearrange the above equation:

(1-p_t)=(1-p_yr)^t

take the logs of both sides

ln(1-p_t)=t ln(1-p_yr)

t = ln(1-p_t)/ln(1-p_yr)

If we put the numbers in this means that we would need nearly 30,000 years of data to have a 95% probability of seeing a 10^-4pa event.

As pointed out by Colin Taylor, the coast of the UK has changed quite considerably over that period. In fact, 10,000 years ago the UK was connected to the continent via a land bridge. Incidentally, that land bridge may have been destroyed by a massive tsunami¹.

I am sure that Colin Taylor and others would argue that having their models tested over such time periods is impossible and unreasonable. I agree it is impossible but considering the risks involved it is not unreasonable. The only conclusion is that there should not be a nuclear power plant at Sizewell.

¹ Storegga Slide, Wikipedia (http://en.wikipedia.org/wiki/Storegga_Slide)