From the Salon: Nuclear Power – Numbers Matter


I was one of the participants – one of the speakers – in a production at the Traverse in Edinburgh, by Untitled Projects. The Salon Project was immersive theatre, evoking a 19th century parisian salon, but with anachronisms an important part of the point, and with the talks intended to be (mild) ‘provocations’. In that spirit, I was talking about Nuclear Power – Numbers Matter. This is what I said.

Actually, it's not. On the three occasions I delivered it (since, ever keen, I volunteered for the dress-rehearsal, too), it surprised me by coming out somewhat differently each time, so this version, after the fact, is... what? The idealised form? The ur-form? The platonic form? The form it would have if I made a version for a blog post? Ah yes, that'll be it.

Nuclear power and numbers

Imagine you're reading a newspaper account of an industrial accident. It's a bad accident, and there are lives lost. That's bad.


Someone asks you: how bad was it? how many people died? But you don't know – the article wasn't clear. OK, they say, roughly – was it one? A thousand? Was it a whole town wiped out, and a million deaths? Oh I don't know you say, I couldn't really tell from the article. It wouldn't be controversial to say that this was a poor bit of journalism.

Unfortunately, this sort of article is common. There were plenty of newspaper accounts of the recent accident in Fukushima which were similarly ambiguous, to the tune of factors of a thousand, in reporting how bad the accident was, and how much radiation could be measured in various places around the reactor.

And this isn't just about knocking journalists (fun though that might be). There's similar ambiguity in a lot of the public debate about nuclear power, which makes a space for vague assertion and weak generalisation. And this is bad, because if we as a society want to come to some sort of rational conclusion about what to do next with nuclear power – a debate in a rational world, rather than a superstitious one in which nuclear power is some sort of black magic – then that debate has to be grounded in facts about the world. Those facts are, mostly, numerical facts.

What this is not

I have a variety of opinions about nuclear power, and about where we should get our energy, but I'm not trying to persuade you of those.

I'm not trying to persuade you that the whole thing comes down to Gradgrindian Facts – there are issues of principle, and of engineering aesthetics, which come in here, too.

I'm not even trying to persuade you that it's universally clear what the important facts are or, where it is clear, that the facts are uncontested – there's plenty of reasonable disagreement about some quite important numbers.

But I am trying to persuade you that facts about the natural do matter somewhere, and that a discussion which doesn't include some backed-up numbers is probably a discussion that it's not worth listening to.

What counts as a big number?

The problem is that it's not always terribly helpful to simply get a number written down, and this is because, in spite the title of this talk, it's not really the numbers that matter, but whether they're big numbers or not.

So, the real point of this talk is to give you some idea of what counts as a Big Number in this discussion. And for that, I will have to resort to Pedagogy.

I'm going to tell you what counts as a big radiation dose; so I'm going to tell you about the Sievert. And I'm going to tell you how many people we collectively think it's OK should die, in order that we can get the power to run our lives; so I'm going to tell you about deaths per Gigawatt year. I'll talk about the second one first.

Deaths per gigawatt year

When you get your electricity bill, you see it's charged in units of kilowatt hours. That's the amount of energy it takes to keep an electric fire going for an hour or so, and it costs about 10p.

A rather bigger amount of energy – about 10 billion times bigger – is the gigawatt year (GWy), which is the amount of energy it takes to keep the UK running for a week or so. That Gigawatt year comes from a mixture of coal, gas, oil, nuclear and other sources including hydro and wind. How dangerous are these sources of energy? How many people die, on average, per GWy generated from coal, oil, gas and so on?

In the UK, coal comes in at about three deaths per GWy, oil about four, gas at 0.4 – roughly one death per two GWy from gas – wind is a bit less than 0.2 deaths per GWy, and nuclear is 0.1. Hydro appears to be somewhere between zero and one, depending on who's counding. Rooftop solar power turns out relatively poorly because – well – it's installed on rooftops, and folk fall off of them.

These deaths are from various types of industrial accidents, plus accidents which affect the public, plus the effects of pollution on shortening people's lives.

You can't say from this anything as bald as nuclear is thirty times better than coal, but the difference in death rate is something you can conjure with and, very crudely, you can reasonably conclude that coal would have to be thirty times better than nuclear at something, simply in order to break even.

Doses and Sieverts

OK, so much for deaths. What about doses?

We can measure radiation doses – the amount of energy your body absorbs from exposure to radiation – in units of sieverts (Sv). Just like the gigawatt year, a sievert is (basically) a unit of energy, though much, much tinier. So, is a sievert a lot or a little?

A sievert is quite a lot. If you're exposed to a sievert within a relatively short period – say hours or days – you'll probably end up in hospital. You will be ill, but you'll probably get better, though your doctor will keep an eye on you long term, as you'll have a few-percent increase in your lifetime cancer risk. If you're exposed to several sieverts, that is a Bad Thing: a dose of about four sieverts gives you only a 50-50 chance of surviving the next two months.

A millisievert, on the other hand (mSv), is not very much at all. A millisievert is to a sievert what a millimetre is to a metre.

You're exposed to a few mSv over the course of a year, purely from natural sources (around double that if you live in Aberdeenshire or Cornwall). This might come from our houses, our food, medicine, coal power generation, and about half a mSv comes from space. The Health and Safety Executive (HSE) have 1 mSv as the limit on the extra dose members of the public can receive from industrial processes.

The paperwork event horizon

It's important to realise that this isn't a safety limit – the limit isn't saying above this lies danger – instead it's the point below which the HSE loses interest in you or, put another way, the point at which paperwork starts.

If your job exposes you to more than 1 mSv a year – perhaps you work in a hospital radiology department, or you're an air steward – your employer will have to fill out a form saying they've thought about how much radiation you're expsed to. If your exposure is above 6 mSv (which generally rules out the aircrew, but includes the radiologist), you become a Classified Person, which means more paperwork, a personal monitoring badge, and a limit of 20 mSv per year. Emergency personnel may be exposed to higher doses (a few times that) in emergencies, or more in exceptional cases where lives are at stake.

Just for scale, it appears that it's at doses of around 100 mSv per year that it becomes possible to detect an increased cancer risk. This depends, however, on how you receive the dose: getting it all in a few days or weeks has a more severe effect than getting it spread out over the whole year.

Depending on what you're having done, a medical X-ray will give you a dose of anything between a thousandth of a mSv, and a few mSv. That's unless it's a CT scan, where the dose is from 1 mSv to somewhere in the region of 20 mSv. If you're getting radiotherapy, you'll get a big dose to a small area of your body: bad for the cancer, but a relatively mild bout of radiation sickness for the rest of you.

The microsievert

Heading downwards, the next unit down is the microsievert (µSv) – a thousandth of a mSv and so a millionth of a Sv. A µSv really is a negligible dose, but the place where you're likely to see the unit mentioned is when someone is talking about a number of microsieverts per hour (µSv/hr) – the sort of dose rate that was one of the higher ones mentioned in the context of Fukushima, outside the accident site itself. Since there are about 10 000 hours in a year, a dose rate of 10 µSv/hr turns into about 100 000 µSv/year, or 100 mSv/year, so a dose worth paying attention to if you stay in place for 24 hours a day, for a year.


So there you have it – you know what's a big number and what's not, in terms of the deaths we're collectively willing to accept for our energy, and in terms of what doses we should and shouldn't worry about. You're therefore a sophisticated critic of nuclear arguments, for or against. But you're also a more demanding critic, since any argument which claims that this or that is ‘dangerous’ or ‘safe’, without backing this up with a number you can put on this scale, is an argument you should probably ignore.

Additional details

This wasn't part of the original talk.

But that's not all there is to it!

Of course not.

There are all sorts of other complications about the medical effects of radiation: how much is too much, and will lower doses always result in a lower risk? I think it's significant, however, that one reason why it's hard to find out about the medical effects of low doses of radiation, is because the effects are so small that they're hard to detect. That's not an excuse to become complacent, but on the grounds that one should worry about the biggest hazards first, it is an argument that radiation risks should perhaps go a little further down the list of Things To Worry About Today.

There are also lots and lots and lots of other complications about the nuclear industry. There are multiple families of nuclear reactor designs, with various trade-offs, only one of which has been thoroughly developed (partly because it happened to mesh, in the past, with military nuclear objectives). The huge scale of nuclear investments, with a correspondingly small number of possible manufacturers, means that nuclear regulators must be permanently vigilant against regulatory capture. The ‘shape’ of nuclear accidents and waste disposal is different from that of coil (mine collapses and slagheaps) or oil (oilspills and particulates) or, say, solar panels (roof falls and the exotic pollutants found in panels and batteries).

One of the biggest complications is that radiation-related risks tend to loom larger, in people's minds, than risks associated with more risky, but more familar, hazards associated with fossil fuels. There is more to our acceptance or rejection of risks than our objective chance of injury from them. But that number, and the numbers associated with it, are surely good places to start thinking.

Where the numbers come from

This wasn't intended to be a precise talk, partly because it's a bit tedious trying to mime footnotes whilst trying to be at least a little entertaining. Also, I'm not any deep expert when it comes to energy policy – but this wasn't a talk about energy policy, but a talk about how that policy should be talked about.

All the numbers are real, though. Quite a few come from the excellent book David J C MacKay, Sustainable Energy -- Without the Hot Air, UIT, 2008 isbn:978-0954452933, or online at

Norman, 2011 October 20