Risk assessment and unusual events March 18, 2011Posted by Cameron Shelley in : STV202 , trackback
The recent earthquake off the coast of Japan has, of course, devastated that country and, unfortunately, brutally exposed some of the difficulties of disaster planning and preparedness. In particular, a tremblor of such a high magnitude was not expected at the site where it occurred, according to this article in Science:
“I never thought this kind of [event] could happen” in this region, says Hiroo Kanamori, a seismologist at the California Institute of Technology in Pasadena.
This lack of forewarning is doubly unwelcome in a country whose seismic risks have been so long and minutely studied. How could such a possibility have been missed or dismissed?
(Image courtesy of Tpemg2 via Wikimedia Commons.)
The article provides some insight. First of all, risk assessments rely on past experience. Detailed earthquake measurements have been made and kept for a about a century in the region. In that record, earthquakes of magnitude 7 to 8.3 have occurred in 30 to 50-year intervals. This pattern suggested to seismologists that stress on the crustal plates in the region was being released rather than pent up. Given the apparent lack of stress buildup, who would expect such a powerful event?
Records of earlier earthquakes are less precise. Studies of silt buildup in the region suggest a large flood occurred in 869 AD, caused by an earthquake of magnitude of roughly 8.3. Perhaps earthquakes of such high magnitudes occur in 1000-year cycles, but there is almost no way to tell because they are too infrequent to figure in the data available.
The article also states that Japanese seismologists have been focused on the earthquake danger from the Nankai Trough, a fault zone that runs close to Tokyo. Perhaps this focus comes from the proximity of the trough to Japan’s largest city. Unfortunately, as the article points out, the 11 March earthquake and the 1995 Kobe earthquake, the two worst in Japan’s history, did not occur in this area. Perhaps the obvious impact of a quake near Tokyo led researchers to be more interested in the likelihood of an earthquake in that area than in other ones.
Finally, even if the stress on a fault can be estimated with some accuracy, its release is still difficult to predict. Will it be released in the form of several, smaller quakes, or in the form of one big one?
These points all emphasize the problems that bias and uncertainty represent for risk analysis. Risk assessment naturally tends to focus on available data and assets at stake. The result may be under-representation of risk and, therefore, under-preparation for destructive events. In view of these lessons, Kanamori argues for a change of focus in preparations:
“In view of the inevitable uncertainty, in my opinion it’s better to have a more general approach [to earthquake preparation] than to have a prioritized, focused effort.”
What this means is not clear. A greater margin of safety in building codes? More standby disaster relief assets? These are hard to maintain for extended periods because of the extra expense involved.
A more skeptical possibility is that disasters like this one are hardly possible to prepare for:
For a nation that prides itself on preparedness, a distressing realization might be that some earthquakes are just too big, and too rare, to prepare for.
If true, then this could be bad news for people living near other faults similar to that off Fukushima, Japan. In Canada, this includes citizens of British Columbia, who live near the Cascadia fault. What should Canada do to prepare?