Source: Nikkei Asian Review
Haiti on Jan. 12 marked the fourth anniversary of the magnitude-7.0 earthquake that left between 100,000 and 200,000 dead. The nation’s government estimated that 250,000 residences and 30,000 commercial buildings collapsed or were severely damaged. The Haitian people have barely recovered.
Four years and one day after that disaster — on Jan. 13, 2014 — a magnitude-6.4 earthquake occurred about 40km north of Puerto Rico. Although some buildings were damaged, no one was killed.
Neither of these events should have been too surprising given the history of quakes in those areas. Yet it seems both Puerto Rico and Haiti were unprepared.
This brings us to efforts to predict and prepare for Japan’s next big shake. Just before New Year’s, the Japanese government’s Central Disaster Prevention Council said there is a 70% likelihood that an earthquake of magnitude 7.3 will hit below south-central Tokyo during the next 30 years. The council said the worst property damage and loss of life would occur if the quake comes on a winter evening with winds blowing at 8 meters per second.
What, exactly, does a 30-year period mean? And should we believe such predictions?
To start with the first question, a “return period” of 30 years is often used in the insurance industry, mostly because the lifetime of certain insurable properties is 30 years. In any given 30-year span, an event may occur once, twice, more often or not at all.
To annualize a 30-year return period you simply divide the likelihood by 30. In this case, a 70% chance in 30 years means a likelihood of about 2% per year. Certainly, 70% in 30 years sounds scarier than 2% per year.
Moreover, given odds of 70% in 30 years, what is the likelihood of the worst-case scenario? Since earthquakes happen independently of winter evenings and wind speed, we must factor these in. When you do that, the result is a likelihood of about 0.06% per year, or roughly 2% in the next 30 years. That is a far cry from 70%.
In maps we trust?
Tokyo sits on a very complex system of tectonic plates. Therefore, it is difficult to predict which part of the Tokyo metropolitan area will be most affected by the next major earthquake.
The CDPC task force created hazard maps for 19 separate quakes of three different types, all with intensities in the magnitude-7 class. The 7.3 earthquake under south-central Tokyo is a worst-case scenario because the areas near the epicenter are packed with industrial facilities and districts full of wooden houses. Massive fires could erupt.
But how accurate are the maps? Dr. Robert Geller of the University of Tokyo’s Department of Earth and Planetary Science points out that the good news is that we have methods for making earthquake hazard maps. The bad news is that the methods have not been verified. And the worst news is that the hazard maps do not agree with the data.
“Science works by formulating hypotheses and testing them against observed or experimental data,” Geller goes on to say. “Most fail and are rejected. The few that appear to succeed are further tested. Most of these, too, are eventually rejected. … Problems can arise when unvalidated hypotheses are adopted as the basis for public policy without the recognition that they may be on shaky ground.”
Does giving predictions, which may or may not be based on solid science, help or hinder the public?
Naoshi Hirata, a professor at the University of Tokyo’s Earthquake Research Institute, seems to think it helps. “I want people to take disaster management measures,” he said, “aware that any part of Tokyo could be directly hit by an earthquake.” Hirata took part in selecting quakes for the task force’s projections.
As the years pass without a major earthquake, the public’s vigilance will lessen. This is called ‘safety drift’. Imagining a huge disaster can dangerously lower our guard for smaller disasters that can have cascading consequences.
Geller again: “It is time to tell the public frankly that earthquakes cannot be predicted. … All of Japan is at risk from earthquakes, and the present state of seismological science does not allow us to reliably differentiate the risk level in particular geographic areas. We should instead tell the public and the government to ‘prepare for the unexpected’ and do our best to communicate both what we know and what we do not.”