[patio11]

Keep in mind, log scaling makes 9.0 not just a wee bit bigger than 8.3 (and the plants were virtually unaffected by the quake, it was the tsunami that got them, because quakes were the events they spent most resources planning for).

[anigbrowl]

Tsunami is a Japanese word, no? If the language has a particular word for giant seismic waves, and you have a nuclear nuclear plant sitting right on the Pacific coastline, a large tsunami seems like exactly the sort of thing you might expect to occur in combination with an earthquake.

Do you remember that 2004 earthquake in Indonesia, the on followed by a huge tsunami that killed a quarter of a million people? It swamped a coastal reactor in India, ~1000 miles away. Statistical evidence suggests that Japan gets hit with tsunamis every 8 years on average; this is the 3rd in the last 30 years with waves >10m. Fukishima Daiichi is 6m above sea level. I know they're not psychic, but it was rather predictable.

http://www.bloomberg.com/news/2011-03-25/tsunami-risk-well-k...

[patio11]

It is a much more sensible objection to say that the incident happened because the tsunami was 10 meters instead of nine meters, and the possibility of it being 10 meters was suggested by a previous tsunami mentioned once half a decade ago by a single engineer on the basis of an estimation gleaned from historical records that predate England, by not pursued further. But when you phrase it like that, it sounds like the awesomely accurate 20/20 hindsight that it is, and the narrative needs a villain.

[anigbrowl]

That interpretation requires a lack of intellectual curiosity that borders upon the obtuse. From the link above, which it appears you didn't bother to read:

Japan has suffered 195 tsunamis since 400, according to Japan’s Central Research Institute of Electric Power Industry, which produced a report on tsunami threats to nuclear plants on the opposite coast to Dai-Ichi in July 2008. Three in the past three decades had waves of more than 10 meters.

A 7.6-magnitude quake in 1896 off the east coast of Japan created waves as high as 38 meters, while an 8.6- magnitude temblor in 1933 led to a surge as high as 29 meters, according to the U.S. Geological Survey.

Yes, it was a very big earthquake and they were very unlucky, but the country is famously earthquake prone. I mean, nobody expected that Indonesian earthquake to be as big as it was either. they had an earthquake 15 years ago that wiped $100m off their GDP and killed thousands, and which surprised everyone considerably. If they are being astonished by natural phenomena on a regular basis then maybe their intellectual confidence is somewhat misplaced. The concept of >10 meter waves off the coast of Japan is hardly beyond imagination: http://4.bp.blogspot.com/_8Ok7XT4IAwo/TRQYSauk1gI/AAAAAAAAAH...

[nitrogen]

Weren't the reactors that failed designed four decades ago, so that 10 meter waves in the past three decades could not have been considered?

[anigbrowl]

They were, but the actual damage done by the tsunami was to wreck the fuel tanks for the diesel backup generators, which were intended to keep the cooling systems operating if the grid power supply failed. The failure of the backup system led to failure of the cooling system, allowing the heat to build up inside the reactors, causing the pressure explosions around the reactors and seemingly inside one or two of them, if reports about cracked containment chambers are correct.

A seawall was supposed to handle any large waves, but among other problems, some fuel tanks for the diesel generators were immediately behind it - not really the ideal place to put any component of your backup system. There's a before & after photo here: http://everist.org/pics/Fukushima/Fukushima_fuel_tanks.jpg

The reactors themselves would have been hard to move, but would likely have been OK had the backup systems been more robust - and being external and modular, they could have been made more robust, and ought to have been.

Even if it were an all-or-nothing proposition, the additional data gathered in the years since the plants' construction was sufficient to justify a re-evaluation of risk; if the risk were sufficiently probable but impossible to mitigate, then decommissioning would be the appropriate response. Sure, that would be very expensive, but much less so than a post-disaster situation like this. Going ahead with a plan that's known to be flawed because you've already put a lot of money into it is known as a 'sunk costs fallacy'; the amount you've already spent has zero bearing on the future probability of failure, so if the latter is unacceptably high then the size of the former is no excuse for inaction.

[yread]

But you generally build nuclear powerplants so that they withstand events with much lower probability than 99%!

Besides the Sanriku earthquake was estimated 8.6