Thoughts on the eternal headache that is nuclear power – part 5 of 6: Nukes in the real world

How radioactivity damages your health, or maybe doesn’t:
Why can’t we agree on how many people were killed by the accident at the Chernobyl reactor? We’re not arguing over a narrow range here, we’re arguing about widely disparate numbers. The United Nations Scientific Committee on the Effects of Atomic Radiation states the number of deaths from the radiation to be 62, the International Atomic Energy Agency predicts the number at 4000, Greenpeace put the number at between 100,000-300,000, and there’s a in-depth review by Russian scientists that puts the number at a million. So we’re not arguing plus or minus a few here, we’re talking four orders of magnitude here. So what gives?

An accident like Chernobyl or Fukushima will spread radiation over a large area. If you are close to the accident, you’ll receive lots of radiation, if you’re far away, then less. We know that lots of radiation will kill you. Less will make you very sick, and still less will make you slightly sick. At what point does that radiation present no risk?

It turns out that we don’t actually know. Radiation damages the body and the body has ways to repair itself, so it may be that under a certain threshold, a small radiation dose won’t have any health effect. Then again, the damage to the body is localised and it might not be possible for the body to heal such damage, in which case there is no threshold – any radiation might cause cancer. The latter is the current assumption; there is no scientific consensus that is the “no threshold” model is wrong.

Can we work out which model is true? After Chernobyll, if the “no threshold” model is true, then we could well be looking at hundreds of thousands of deaths across Europe in the next fifty years. However, 40% of people in the West die of cancer anyway. Over fifty million Europeans will die of cancer over the next fifty years, plus or minus quite a bit through variations in how healthy those Europeans are to start with, variations in health care, and variations in natural background radiation. If Chernobyl kills an extra ten thousand per year, then those deaths are impossible to detect on top of about a million a year. Those people will still be dead, but you cannot connect their deaths with the radiation release from Chernobyl.

To quote from the Wikipedia discussion page for the Linear no-threshold model article, “This is hands-down the most controversial topic in radiation protection.” This is simply not a question that science can answer. We don’t know and currently, we can’t find out.

Why do we care? When the radiation from an accident spreads out across a continent, it potentially gives a hundred million people a very low dose. A small increase in the chance of cancer for a hundred million people might add up to a large number of deaths. So, how much of a low dose is low enough? We cannot say. How safe should nuclear plants be designed to be? We cannot say.

If it turns out that low levels of nuclear radiation are as bad for you as we think they are, then we can justify spending a zillion dollars to make each plant safer, to shift the risk of catastrophe from one-in-a-million to one-in-ten-million, or to one-in-a-billion. Or if it turns out that low levels of radiation aren’t that bad, then we’ve just spend zillions to avoid an accident that might have only killed the fifty people next door to the plant and if we’d known better, then we could instead have spent the money on zebra crossings or anti-smoking adverts and saved many more lives.

Why none of this matters to New Zealand:
We’ve got loads of cheap renewables. Loads. Have a supply curve:

Each chunk is how much renewable electricity is available at a given price. Start from the left (the cheapest), and work your way along. For reference, New Zealand’s electricity use is growing at about 500-800 gigaWatt-hours per year. For less than 10 cents/kWh, we can cover at least twenth-five years worth of growing demand, with current technology. And, as I mentioned earlier, renewable energy technologies are getting cheaper faster than any of the alternatives.

Now, I bitched earlier about the price of nuclear power being pretty indeterminate. However, let’s assume we can use the levelised range of numbers from 45-110 US$/MWh for the levelised prices. Right now, the NZ dollar is kicking arse and taking names, but a nuclear plant is an investment over at least a decade. So, taking an average of the last decade’s exchange rates and those estimates come out at between 7 and 17 cents per kWh. I’m being generous to pick a price level of 10 cents per kWh and at that price, there’s a metric fucktonne of hydro, wind, and geothermal that we could build instead for less money than a nuke.

Oh and yes, nuclear power is horribly unpopular here, we’d have problems fitting a big lump of nuclear power into NZ’s electricity system, and we’d have to import not just the plant but all the people to run it and to regulate it, but fundamentally it’s just not cost-competitive here and that’s why we’re not having a debate about whether to even think about building nukes in New Zealand.

The rest of the world isn’t so lucky, but I’ll get onto that tomorrow.

7 thoughts on “Thoughts on the eternal headache that is nuclear power – part 5 of 6: Nukes in the real world”

  1. The biggest single power-generation kill ever recorded was a renewable source, hydro. The Banquaio dam let loose in China in 1975 and the death toll exceeded 200,000 — the figures are inexact because when a wall of water wipes out several dozen towns counting the bodies is kinda difficult. There have been smaller dam disasters over the years, of course — even the recent big earthquake in Japan caused an irrigation dam collapse that killed a bunch of folks; early reports said four died at the dam itself but it also washed away some houses downstream and a number of people were reported as missing. Since the world’s news has focussed on the nuclear plants at Fukushima Daiichi I’ve not seen any updates in the English-language press about this dam break or what the final death toll was.

    The Three Gorges dam complex in China (19GW of renewable hydro power plus flood control) is just about complete; if that ever lets rip the death toll downstream could be in the millions.

    1. Yes, but… the Chinese dam failure killed more than any other dam failure. In fact, it killed more than all other dam failures put together, by at least an order of magnitude. The only other failures that have killed more than a few hundred were the RAF Dambusters attack (killed 1650) and the Italian Vajont Dam (killed 2000). NZ’s own Opuha Dam collapse killed about a thousand sheep.

      With dams, we’re talking about localised and definable risks downstream. Either a house (or a village) is washed away, or it isn’t, so the impact is easy to measure and there are no insidious long-term effects. It’s easy (from an engineering point of view) to design out the risk: design to hold the water, make sure excess water can be spilled safely, make sure the reservoir slopes are stable, and limit building downstream if there’s a risk that you can’t stop the shit hitting the fan.

      The moral for me from the major hydro disasters is that good engineering isn’t enough, you also need to have a society with a sane approach to safety. The Chinese dams that failed were built in the height of the Great Leap Forward and the Cultural Revolution, where ideological ideas trumped rational risk assessments. For Italy, the warnings of experts were dismissed by politicians, and Italian politicians were no better then than they are now. For the German dams, it helps if people are not trying to blow up your generating plant. (We haven’t seen a wartime attack on a running nuclear plant, god hep us if we ever do.)

      With simple risks like that, the chance of an accident and the likely number of deaths can be reduced by good engineering practice and clear, strong regulation.

    1. Please specify the expected marginal impact, per dollar, on quality-adjusted life years, then we can start to talk about rationally allocating funding on this basis.

      (Or, you know, we can admit that these kinds of things span the range from guaranteed to save a life for $XXXX to pretty speculative claims, and proceed on that basis.)

  2. Erm, is it possible to add to that Supply Curve, just about where NuZild is on it at present? It would then make an interesting comparison to the approx 25c/unit I’m paying for my power (not to mention the 38c/day line charge)…

    1. National average about 15 c/kWh, apparently:

      Residential prices differ from the national average, but if you’re a generating company, then the national average is what matters.

Leave a Reply

Your email address will not be published. Required fields are marked *