(WSJ links may not last)
One glaring oversight, they say, is the measurement used to determine the test levels for a potentially crippling tidal wave. In at least one case, the perceived maximum level was set at 38 feet—even though Tokyo Electric Power Co.'s Fukushima Daiichi was hit by a 46-foot-high tsunami in the March earthquake and some towns saw waves of up to 132 feet.
They say the tests don't explore other types of risks, such as a deliberate attack, and don't look at multiple causes of failure, like what happened at Fukushima Daiichi, which was first hit by a major earthquake and then a tsunami.
The first paragraph is a big issue I have with Japan. The max height at the reactor was on open ground! This is not the height of run-up, which is what happens if you put a big seawall in front of it. That's 132 feet! The difference is the momentum or kinetic energy of the water.
The second point applies to all nuclear operators, who, as I have said, run a very low-class operation. Multiple failure does not mean earthquake+tsunami! It means things like small loss of coolant with grid power loss, and backup failure. It means failure over 4 reactors. All those sorts of things. Poor operation is the reason that the probability of the next Japan is as common as the chance of a major earthquake hitting another nuclear plant. For Toronto and New Madrid, we are talking worse than 1 in 1000, maybe as common as 1 in 500. Enough for a nuclear disaster anywhere in the world every 10 years.