Wednesday, 16 March 2011

People are strange

Suppose this were the lead item in tomorrow's local news roundup:
A local man was driving along a narrow road on the side of a cliff when he drove over a spike strip left there inadvertently by police. The car, a Volvo, went off the side of the cliff, rolling several times and bouncing off the cliff wall repeatedly before coming to rest some two hundred metres below. The driver reported a mild concussion and a broken arm; the car is a write-off. Volvo engineers reported that the car wasn't designed to withstand falls of several hundred metres, has leaked some engine oil, and the radiator may yet spill coolant.
If I read the story, I'd be more inclined to buy a Volvo. But these guys would likely start protesting in front of Volvo headquarters for building unsafe cars.

What's going on in Japan is awful. But, relative to expectations about nuclear accidents - it could be worse. There remains strong chance of meltdown. But so long as the containment unit doesn't blow (yes, that's a huge so-long-as), this isn't a nuclear apocalypse.

The reactor had been tested through a 7.9. Nobody had reckoned on a 9.0 earthquake - recall that this is a logarithmic scale so the increment from eight to nine isn't an eighth on top of design spec. And even with that massive quake, followed by a tsunami that wiped out a whole pile of backup systems, we've had thus far only relatively minor releases of radioactive gases. This is about the worst beyond-spec thing you could throw at those reactors. And, three days later, they've still only released worrying rather than apocalyptic levels of radiation.

Yes, this is a terrible tragedy. And I'd be damned scared if I lived anywhere near these plants now too. I'm on the paranoid side about such things - we bought the Iodine pills post 9-11 living outside DC. But think about how far beyond design spec these reactors have been pushed. As far as updating estimates about the safety of nuclear power, we ought be revising upwards, not downwards. You have to beat on them massively hard - well beyond design specs - before bad things happen.

Germany's pretty geologically stable. I don't see what new information is revealed by the Japanese earthquake --> tsunami --> nuclear accident that ought convince any rational German to revise downwards his estimate of the desirability of nuclear power. It's only increased the salience of downside risk, not the estimate of its probability.

And we oughtn't forget that people die in coal mining accidents, that fine particulate matter from coal-fired generation kills people, and that burning coal also releases radiation. If coal's the most likely substitute for reactors in Europe, that's almost certainly a net negative for the environment.

MIT's Nuclear Information Hub is the best source I've found on the ongoing crisis. Two days ago folks there didn't seem too worried, though that may yet change. Here's more on the importance of preparation and strict emergency drills for this kind of thing.


  1. Just one comment - the quake may have been a 9 on the Richter scale, but it was also at least 150km away from the site of the reactors.

    As this USGS map shows, the shaking wasn't that severe.

    As I understand it, it was the shutdown of the reactor followed by the tsunami taking out the backup power that is causing the problems.

    Link to the Chch shaking magnitude map for comparison:

  2. @Thomas: Ground force acceleration there remained fairly high - just a bit below Kobe. You're right on the tsunami being the biggest problem.

    @David: I did link to the updated version where the MIT folks expurgated the parts that they didn't think held up; the disclaimer at the top remains worth noting though.

  3. What makes you assume any updating on the part of the protesters? The Japanese events provide a Schelling point, raise emotional involvement (which faciliates action) and lend themselves to easy exploitation for party politics given that there are elections coming up in which the incumbent is staunchly pro-nuclear.

    As for the wider public, I am sure the recent events cause an association of nuclear energy with negative emotions, but I don't think Bayesian reasoning is a good model for that.

  4. @Lemmus: No way that the protesters are updating. It's their audience that I'm wondering about. You're right that they're not updating - if they were, would go the other way. That's kinda the problem/point.

  5. The thing about a nuclear catastrophe is that it will affect our children, their children, their children and probably their children too.

  6. @Paul: Yes, but we've known that since, well, at least 1945. We have no new information about how bad the bad thing is. And we haven't much new information about the likelihood of bad things happening - what information we have suggests it's less likely than we'd previously thought. And plants newer and better designed than the older Japanese plants are less likely to fail.