Of earthquakes, tsunamis and seawalls

More than a year ago I wrote about the issue of seawalls in the aftermath of the Great East Japan Earthquake, and mentioned we were working on evaluating the effect of seawall on evacuation behavior. This time I would like to talk about the results of a study we conducted precisely to investigate this effect. We published this results in the Journal Injury Prevention and wrote a blog article about it which I reproduce below.

The blog post was originally published on the Injury Prevention Blog under a CC-BY-NC licence​.

***Click here for the original academic article published in Injury Prevention.

High concrete wall at seaside for protect tsunami big wave
High concrete wall at seaside for protection from tsunami [Sakarin Sawasdinaka © 123RF.com]
If you have ever lived in Japan or other countries in the Ring of Fire, then you are probably used to some shaking, to earthquake alarms in the middle of the night, and maybe tsunami warnings. And perhaps you might have seen the seawalls, these huge concrete structures in the coast, designed to protect coastal towns from tsunami.

Seawalls have been an intrinsic part of the Japanese disaster mitigation strategy, and with good reason. Japan accounts for 20 percent of the world’s earthquakes of magnitude 6 or greater. But while seawalls might help protect cities from tsunami damages, a prompt evacuation is the best way to protect human life. This idea is embodied in the concept of Tendenko, a traditional maxim from the Tohoku region that dictates that one should not stay and help others, but run and save your own life first (Check the link for a discussion on the morals of Tendenko). Tendenko has been identified as the reason for the so-called Kamaishi Miracle, where almost all elementary and junior high school students survived.

But I digress, what I want to talk about today is the possibility that people might get a false sense of security from having these big seawalls in their town, and that this might result in people delaying their evacuation, or even worse, not evacuating at all. Some critics have observed as much and cited testimonial evidence of people speculating that loved ones had not evacuated in the belief that the seawall would protect them. However, this was just anecdotal evidence and we wanted to test this quantitatively.

To do so, we used data from a survey conducted by Japan’s Ministry of Land Infrastructure Transportation and Tourism (MLIT) on survivors of the disaster. In the survey, people were asked about where they were when the earthquake occurred, whether they heard the tsunami warnings, and what they did right after the earthquake. We then used that data to compare the evacuation patterns of those who were near a wall that was higher than the forecast tsunami (we called this an “effective wall”) wave and those who had no wall nearby or were near a wall that was lower than the forecast. We found that people in areas near an “effective wall” were 30% less likely to evacuate promptly than those who did not have an “effective wall” nearby, suggesting some sort of false sense of security in relation to the wall.

Staying put might be a rational decision if you are 100% sure that the tsunami forecast is accurate, and the wave will not overcome the wall. However, forecasts of rare events that happen once in a thousand years are not very reliable even with the latest technologies, and the larger the underestimation of the tsunami height the more you expose to risk if deciding when to (or whether to) evacuate based on the tsunami forecast. In fact, the tsunami height estimates were considerably underestimated in the first warnings and revised later on.

These findings underscore the importance of disaster education programs such as the one in Kamaishi City that emphasize proactive evacuation decisions over reliance on predefined risk assumptions, and the validity of the traditional wisdom of a region historically ravaged by large tsunamis, which can be put simply in two words: “evacuate now!”


Published by

%d bloggers like this: