Thoughts on Fukushima

Whatever the outcome, I don’t think anyone should be surprised by the situation at the Fukushima nuclear plant.  Like virtually all nuclear plants, they’ve been safe and quiet for decades.  But they’re not the kind of thing you can walk away from.  And sometimes, you need to walk away.  Volcanoes erupt.  The Earth trembles beneath your feet.  There are floods, and famines, epidemics and wars.  We do a good job of ignoring these things when they aren’t pressing concerns.  It makes life simpler and more enjoyable, especially since historically, we’ve had little power to do anything about infrequent, terrifying events.

I’m not categorically against nuclear power.  If we can do it in a responsible, scalable way, then great.   Making 10,000-100,000 year commitments is not responsible.  We can’t keep those promises.  Extracting only a couple of percent of the fuel’s energy isn’t scalable to tens of terawatts for centuries or millennia.  So any scalable, responsible nuclear power will involve breeding fissile fuel, and re-processing spent fuel to remove fission products that inhibit the chain reaction.  Additionally, to be responsible in my mind, a nuclear power station should be something you can walk away from at a moment’s notice, with no fear of catastrophe.  It should be something that an invading (or perhaps more likely, retreating) army cannot use as part of a scorched earth campaign without a major engineering effort that would take months of work.

But obviously, that’s not where we are today.

Over the first week or so of reports, I mostly thought “No matter what they say, we really have no idea what’s going on.”  Japanese management and face-saving culture, and the far-reaching economic consequences of a nuclear disaster now, when the industry had been hoping to step in and save us from the Yeti’s coal-based terraforming project, made it seem likely that no matter what was going on, TEPCO et al. would try and minimize the PR consequences.  “It couldn’t possibly be another Chernobyl.” had no information content, one way or the other.  Then the west coast of the US started freaking out about trans-pacific contamination.  Which, even if it were another Chernobyl, wasn’t going to be a real problem.  There’s something about The Atom that makes most people incapable of thinking straight (both the manic atomic booster club, and the reflexively anti-nuclear parties).  If all our Pacific nuclear testing in the 1950s and 1960s didn’t turn the west coast into mutants, then one reactor in Japan certainly isn’t going to either:

All this reminds me of a great Feynman quote, made in relation to the Challenger accident:

For a successful technology, reality must take precedence over public relations, for nature cannot be fooled.

Fortunately, there are some data on the radiological situation near Fukushima.  Unfortunately, they don’t jive with the public story.  In at least one location (the village of Namie-machi, 30 km the plant), the radiation level has been found to be 170 µSv/hr.  Normal background radiation exposure is about 0.3 µSv/hr.  100 mSv/yr is the maximum allowed for radiation workers worldwide, and is equivalent to about 12 µSv/hr.  That measurement was taken on March 17th.  However, radioactive material has continued to escape since then.  Until the reactors are effectively isolated from the environment, we ought to expect that the levels of persistent radioactivity in the area will increase.

The radionuclide I-131, is dangerous in the short term because it’s highly radioactive, and iodine is concentrated in the thyroid gland.  But it has a short half life (8 days), and decays into stable Xe-131, so long term it’s not a serious threat.  The radionuclides that make nuclear sites uninhabitable for long periods are Cs-137 and Sr-90.  Both are common thermal neutron fission products, making up several percent each of the spent mass (i.e. nuclear reactors used in power plants make a lot of them), and both have half lives of decades, not days.  Caesium is very water soluble, and behaves like potassium in your body.  This means it’s very mobile in the environment (bad) but it can be flushed from your system relatively quickly (good).  Strontium on the other hand behaves chemically like calcium, and can be concentrated in your bones and retained for long periods.  Because one of the troubled reactors was running mixed uranium-plutonium oxide (MOX) fuel, there’s also the potential for Pu-239 contamination.  Fissile plutonium, with its 24,100 year half life, actually isn’t very radioactive.  You can safely hold a (subcritical) lump of it in your (gloved) hand.  However, it is spectacularly toxic in a purely chemical sense.

So, we know there are some places 30 km away from the reactors that a couple of weeks ago had already accumulated enough fallout to give you 30 times the legal annual dose for radiation workers.  Since then radioactive material has continued to escape.  Namie-machi has the worst radiation measurements so far.  Hopefully most places will be nowhere near as contaminated.  We won’t know the real magnitude of the accident until the reactors are finally sealed, and the region’s radioactivity can be mapped in detail.  Anybody saying they’ve been convinced by this accident that nuclear is safe is jumping the gun.  So far most of the fallout has gone east, into the Pacific, where it will be diluted to safe concentrations, but the amount I-131 and Cs-137 released so far is comparable to Chernobyl.  Strontium isn’t volatile, so it’s not yet escaping as readily.

Now, just for scale, let’s generalize from small numbers (dangerous, I know).  It’s been 25 years since Chernobyl.  According to the BP Statistical Review of Energy for 2010, we have about 0.8 terawatts of nuclear power generation worldwide, from about 450 reactors.  Nuclear boosters have suggested that we ought to scale up nuclear generation to avoid climate change, and provide North American per-capita quantities of energy for those in developing economies.  If for every 0.8 TW of nuclear power we get one major accident every 25 years, how many accidents might we expect to see if we converted our entire grid to nuclear?  On a 2050 world with 9 billion inhabitants, striving to live the way we do, that would be something like 10 or 20 TW of nuclear power.  That’s roughly 10,000 nuclear power plants, spread all over the world, in nations of every description, exposed to geologic and political turmoil of all types.  A big accident every 25 years for 0.8 TW of power is 0.05 accidents per year per TW.  At 20 TW, that’s one major accident per year on average.  If each one were to render 1,000 sq km uninhabitable for 500 years on average (an exclusion zone within ~20 km from the accident) that’s half a million sq km uninhabitable at steady-state.  That’s roughly the area of the UK, or California.  Or Japan.

Much has been made of the scale of the earthquake and tsunami.  That they were beyond the scope of reasonable design constraints.  A once in a thousand years kind of event.  But if you have a thousand sites (let alone 10,000), you ought to expect roughly one of them to get some kind of once a millennia event each year.

Are these acceptable trade-offs?  If our only choices were terraforming the world with coal, or rendering a California sized chunk of it radioactive, I’d go with radioactive.  It affects less of the world, and it doesn’t last as long as our new atmosphere will.  But then there are other choices.  Efficiency (living well while wasting less, not shivering in the dark) and renewables are safe, reversible, and cheaper than nukes anyway.  So while the events of the last few weeks in NE Japan are utterly tragic, some part of me is thankful that the next nuclear accident happened now, rather than a decade or two from now, after another round of nuclear investment and construction.  It’ll be interesting to see how it affects energy planning in the near future.  It’ll be sad if it means we fall back on coal.  Really they’re both unacceptable.