On March 11, 2011, a magnitude 9.0 earthquake struck off the east coast of northern Honshu Island in Japan, causing devastating tsunamis and killing thousands. The quake and ensuing tsunami resulted in the catastrophic failure of many of the reactors and associated facilities at the Fukushima Daiichi nuclear power plant. The crisis at the Fukushima plant continues to unfold at this moment.
Was this earthquake unexpected? Yes and no. We live on a dynamic Earth. Our active Earth actually surprises many people who live calm lives in areas unaffected by the processes of a planet constantly in the course of reshaping its surface. The reason Japan even exists is due to the large subduction trench that lies to its east, the source of the recent earthquake. So, though earthquakes in that zone did not exceed magnitude 7.8 in the twentieth century, it is not totally unexpected that such a high-magnitude quake would occur there.
Herein lies the problem: Geologic knowledge is incomplete and imperfect. And we rely on it perhaps too heavily when making policy decisions about siting hazardous technologies such as nuclear power plants. These facilities need geologically stable, physically secure environments in which to operate, and sometimes we push the envelope too far. Clearly that is the case with Fukushima.
Nuclear reactors are built to withstand what is termed "design basis accidents." What happened at Fukushima is a "beyond design basis accident." The nuclear industry bases its design basis accidents on "credible events," which are determined by a probabilistic analysis. Unfortunately, the Fukushima crisis is clear evidence that they got the probabilistic analysis wrong. Why was it wrong? In part because the "credible event" -- one that was "reasonably expected to occur" -- was thought to be an earthquake of magnitude 7.9 at most and a tsunami of 6.7 meters at most. The actual quake released more than 1,000 times the energy and the actual tsunami was almost 9 meters high.
How could we not know that larger quakes and tsunamis could occur there? Actually, geologists are constantly being surprised by the Earth. The magnitude 6.3 earthquake that occurred last month in Christchurch, New Zealand, occurred on an unknown and unexposed fault. Unknown faults that cause damaging earthquakes are fairly common, it turns out.
Moreover, geologists cannot predict much with any accuracy. Though they know there will be large earthquakes associated with movement on the San Andreas Fault in California, they cannot say when they will occur. Geologists know that the volcanoes in the Oregon and Washington Cascade Range are active, but they can't tell you when they'll blow. They know there will be large earthquakes in eastern Japan, but they simply cannot predict to two significant figures how large they will be.
If a nuclear reactor's safety margin is based on specific predictions of geologic phenomena, as it is, then there is a problem, because geologists are not always correct in their bounding assessments. In addition, there are a number of geologic processes that fall into Donald Rumsfeld's category of "unknown unknowns." Geologists just don't know what these processes are, let alone when they will occur. And if they don't know all the processes and the bounding conditions for them, they cannot make accurate predictions about "credible events."
If the "credible event" is viewed as a risk assessment, as the crisis grows at Fukushima, we now face the question of whether nuclear power in unstable areas was worth the risk. At the very least, it appears that they have lost three reactors, perhaps four, and may lose all six. According to Congressional testimony by Nuclear Regulatory Commission Chair Gregory Jaczko, the spent fuel pool at one of the reactors has lost all coolant and may release large amounts of radioactivity. It will take time to really know the full extent of the accident and answer this question.
I think this situation calls for a reassessment of the portion of "design basis accidents" that depend on geologic information, especially in tectonically active regions. We need to do this reassessment for existing reactors around the world.
Countries that are considering acquiring nuclear power also need to soberly reassess their plans, especially if they are located in tectonically active regions. These include China, India, Turkey, Indonesia, the Philippines, and others.
A final note. Many of the issues described here can be easily applied to the issue of physical security. Spent fuel pools packed with hot, radioactive fuel need water to cool them. They are vulnerable to attack; perhaps even more vulnerable than the cores of the reactors. The pools of the type at Fukushima -- General Electric Boiling Water Reactor Mark I plants -- are located stories above ground level. If the pools were damaged in a sabotage attack, and water was lost, the scenario would be the same as what is currently occurring at Fukushima.
Nuclear power requires stability -- geologic stability and political stability -- to perform well. We can no longer ignore these risks, but must clearly account for them.