The earthquake that struck Japan may have been the largest since historical records began (and the fourth largest ever recorded), but the tsunami had many precedents – bigger ones – in the historical and geological record. The size of a tsunami is related to the displacement of the seafloor, not necessarily the magnitude of the earthquake, and significantly smaller events than the one on March 11th have generated larger tsunamis. This raises two questions: given the size and devastation of past events, why should this have been a surprise and why were “defences” so woefully inadequate? And, are there, realistically, such things as “tsunami defences” at all?
These are questions that countries beyond Japan should also be asking themselves.
I have put off writing anything about the catastrophe, in part because the geoblogosphere has been doing its typical comprehensive job (see Dana Hunter’s links as a good starting point), but also because I was, I suspect like many of us, simply mesmerised by the surreal and incomprehensible scenes that unfolded, day after day, through the unprecedented documentation of the tragedy. But now I have begun reflecting on one of my earliest posts, “Tsunami forensics - sand helping save lives,” and concluded that we would seem to have a long way to go; if hubris, complacency, and other human foibles led to a failure of thinking and planning in Japan, then where is the hope for less developed (or even equally developed) parts of the world?
First of all, following up on the theme of my post on the forensics of sand deposits documenting past tsunamis in Thailand and Sumatra, I easily located a couple of examples of the same kind of work that has been conducted, hardly surprisingly, in Japan. In their article, “Geological Study of Unusual Tsunami Deposits in the Kuril Subduction Zone for Mitigation of Tsunami Disasters,” Japanese researchers describe meticulous analysis of tsunami sand deposits covering the last 5,000 years from Hokkaido, north of the recent events; the sands are carefully dated, and their internal structures used to shed light on the flow regime of the tsunamis that deposited them. They observe that “Only when an unusual tsunami strikes coastal lowlands does a large-scale sediment transfer occur, leaving a sedimentary record, that is, tsunami deposits, in the geological strata on shore.” And they calculate, from the multiple tsunami sands preserved, the average recurrence interval of “unusual tsunamis”: every 220-379 years over the past 5,000. They conclude that “The geological information obtained by this study is valuable for mitigation of tsunami disasters.” Well, maybe.
Similar recent studies have been done in other areas of Hokkaido. For example, “Aperiodic recurrence of geologically recorded tsunamis during the past 5500 years in eastern Hokkaido, Japan” (only the abstract is publically available) documents that “The sand sheet extents suggest that most of these tsunamis were larger than any generated at Hokkaido in the last 200 years. The intervals between these inferred outsized tsunamis average nearly 400 years but range widely from about 100 to about 800 years.” Bear in mind these intervals – 100 to 200 years are entirely possible – in the context of historical events.
Of course, I couldn’t help but be curious about a paper with the title “A Huge Sand Dome Formed by the 1854 Earthquake Tsunami in Seruga Bay, Central Japan.” This is a geological study of tsunami sand deposition as a result of a historically documented event. Seruga Bay is not far south of Tokyo, and the 1854 tsunami was devastating; here’s the abstract of the paper:
The 1854 Ansei-Tokai earthquake brought massive destruction to Suruga Bay, central Japan. The earthquake triggered a large-scale tsunami, which surged over the Pacific coast of Japan. Waves exceeding 13.2 m in height attacked Iruma, southeastern coast of Suruga Bay, and provoked peculiar types of tsunami sedimentation. On the coast of Iruma, a huge mound of shoreface sand, reaching more than 11.2 m above sea level, appeared after the tsunami run-up. We performed a historical and sedimentological survey to clarify the origin of the mound. Result of a field excavation and submarine
investigation suggests that the sand came from the seafloor with a water depth of 20 to 30 m, and historical data illustrates a dramatic change of the landform by the tsunami run-up. Numerical examination of the tsunami implies that the coastal topography played an important role in excitation of the tsunami, and it induced the characteristic tsunami sedimentation.
“Waves exceeding 13.2 meters.” Here’s the town of Iruma today, completely rebuilt on top of this gigantic pile of sand:
Mass transfer of sediment by a tsunami causes wholesale changes to coastal topography, and when the shape of the coast resonates with, and amplifies, the character of the tsunami wave, the effects can be dramatic. Among the compelling images of the last ten days are the series of “before and after” images, of which fascinating (and depressing) examples are from the New York Times and Picasaweb. Any number of these illustrate coastal change and the shifting of huge amounts of sand. Here, for example, the coast near Ueda, south of Sendai:
But this is only part of the story of coastal change. The wholesale physical displacement of Japan as a result of the earthquake is, even to a geologist, astonishing. Callan Bentley reported GPS data showing a lateral movement of 4.4 meters and he points out the less widely reported, but in many ways more significant vertical subsidence of up to 75 cm along some parts of the coast. But now, as if these figures weren’t dramatic enough, Japan’s Geospatial Information Authority in Tsukuba are reporting that the Oshika Peninsula that lies closest to the epicentre has moved 5.3 meters and dropped 1.2 meters since the earthquake – records for Japan. Scrutinise the before and after images, and evidence of coastal subsidence as well as erosion, becomes apparent. Consider these images of the Yamamoto coast, not far from Sendai:
Look closely and compare – the entire coastline appears to have sunk (even allowing for the fact that the relative state of the normal tide in the two images is unknown). The town of Otomo is at the left in each image; a seawall is apparent across the bay “before,” but has completely disappeared – and would seem to be buried in sand – in the post-tsunami image. The devastation in these towns was almost total and rescuers are still struggling to even access them. Here’s another pair of images – are we seeing an enormous sand deposition event similar to that at Iruma in 1854? Was there also a resonance between the shape of the coastline and the tsunami that amplified the potential for catastrophe?
That potential for catastrophe certainly becomes greater if the subsidence of the coast as a result of the earthquake is taken into account. A seawall, such as that supposedly protecting Otomo, instantly drops by a meter, its height rendered even less effective against an unusual tsunami. And it’s not as if this stretch of coast had not experienced “unusual tsunamis” before – you don’t even have to dig into the geological record. In 1896, an earthquake whose magnitude is estimated as 7.2 – in other words, not at the worst end of the scale – caused a reported 25 meter tsunami. And the 1933 8.4 magnitude earthquake caused a tsunami described as 28 meters high. The scale of a tsunami is very much dependent on the shape of the coast, not necessarily the magnitude of the earthquake.
The recorded history of Japan’s coast is filled with accounts of “unusual tsunamis” – David Bressan has an excellent summary. Look into the sedimentary record, and the evidence continues: around Sendai, five tsunami sands are found up to four kilometres inland; along the coast near Ofunato and Otomo, up to 113 sand layers record major tsunamis – some of them from elsewhere in the Pacific. These figures come from the presciently titled book by Edward Bryant, Tsunami – The Underrated Hazard (published in 2001 and now extremely expensive, but available on Google books). As Bryant writes, “the stratigraphic evidence designates this coastline as the most frequently threatened inhabited coastline in the world.” Indeed.
So how, when the historical and geological evidence shows so clearly that devastating tsunamis hit a sinking coastline with a frequency that hardly labels them as “black swans,” could such a catastrophic loss of life – not to mention a nuclear emergency - as we have been witnessing possibly occur?
There is, clearly, no simple answer, but part of the problem would seem to be a level of complacency that is not, unfortunately, atypical of human nature. Take the seawalls, for a major example. As a piece in the New York Times reported, “At least 40 percent of Japan’s 22,000-mile coastline is lined with concrete seawalls, breakwaters or other structures meant to protect the country against high waves, typhoons or even tsunamis. They are as much a part of Japan’s coastal scenery as beaches or fishing boats, especially in areas where the government estimates the possibility of a major earthquake occurring in the next three decades at more than 90 percent, like the northern stretch that was devastated by Friday’s earthquake and tsunami.” But the article goes on to say:
A fuller picture of how seawalls protected or failed to protect areas beyond the nuclear plants will not emerge for at least a few more days. But reports from affected areas indicate that waves simply washed over seawalls, some of which collapsed. Even in the two cities with seawalls built specifically to withstand tsunamis, Ofunato and Kamaishi, the tsunami crashed over before moving a few miles inland, carrying houses and cars with it.
In Kamaishi, 14-foot waves surmounted the seawall — the world’s largest, erected a few years ago in the city’s harbor at a depth of 209 feet, a length of 1.2 miles and a cost of $1.5 billion — and eventually submerged the city center.
“This is going to force us to rethink our strategy,” said Yoshiaki Kawata, a specialist on disaster management at Kansai University in Osaka and the director of a disaster prevention center in Kobe. “This kind of hardware just isn’t effective.”
We are all now numbed by the videos of the tsunami not only crashing over seawalls, but destroying them – videos taken by residents who may well have assumed that they could remain, protected by those defences, residents who were lucky to have survived. And then look at the before and after images of Otomo, above: a seawall not only rendered useless, but consumed.
Again, from the New York Times article:
The height of seawalls varies according to the predictions of the highest waves in a region. Critics say that no matter how high the seawalls are raised, there will eventually be a higher wave. Indeed, the waves from Friday’s tsunami far exceeded predictions for Japan’s northern region.
And, for a mind-numbing example of tragic complacency, watch this BBC video of the Fire Chief of the town of Rikuzentakata, just SW of Ofunato, describing how a tsunami two years ago damaged the seawall. His warnings went unheeded, and, on the 11th March, 45 of his crew were lost trying to close the harbour gates by hand.
I simply don’t understand how, given even the sampling of historical and geological records above, what the “predictions” that were "far exceeded" could possibly have been based on, and how there could have been the total lack of imagination that allowed them to be regarded as definitive. And then there’s the reliance on the – tragically inadequate – seawalls protecting the nuclear power plants. Couldn’t someone, even years after the construction of the plants, have said, in the spirit of over-engineering where failure could have catastrophic consequences, “There is a scenario, not impossible to conceive, in which the seawalls would fail – just to be safe, why don’t we put our generators and pumps, the switchboards and backup systems, on top of an impregnable fifty-foot reinforced concrete structure?” Or something to that effect.
Maybe I’m over-simplifying things, maybe not….
Update: since I wrote this post, I have been in touch with my geologist friend, Rovicky (who writes his own blog on these kinds of things and many other matters geological - in Indonesian) and he provided me with yet another paper on reconstructing ancient Japanese tsunamis from sand deposits. This is from studies of the Sendai area, one focus of recent devastation. Titled "The 869 Jogan tsunami deposit and recurrence interval of large-scale tsunami on the Pacific coast of northeast Japan," the abstract is as follows:
The fore-arc region of northeast Japan is an area of extensive seismic activity and tsunami generation. On July 13, 869 a tsunami triggered by a large-scale earthquake invaded its coastal zones, causing extensive deposition of well-sorted fine sand over the coastal plains of Sendai and S ma. Sediment analysis and hydrodynamic simulation indicate that the tsunami inferred to be triggered by a magnitude 8.3 earthquake spread more than 4 km inland then coast. We postulate that the sand layer was developed by the tsunami’s first wave. Traces of largescale invasion by old tsunami as recorded in the coastal sequences of the Sendai plain show about a 1000-year
reoccurrence interval. We suggest that the J gan tsunami was much larger than tsunami generated by normal earthquakes in the subduction interface.
The paper would normally be available online from the Japanese National Institute of Informatics, but at the moment, the site sadly simply contains the following note: "NII services is currently unavailable due to a scheduled blackout in connection with the Japan earthquake. Sorry for your inconvenience." If anyone is interested, I can email a pdf, thanks to Rovicky.
[For an interestingly different approach to tsunami shelters, and other approaches, see this report from the New Scientist. And for further reading on earthquake and tsunami - and general - hazard assessment, see http://blogs.agu.org/landslideblog/2011/03/15/from-a-geological-perspective-what-is-surprising-about-the-sendai-earthquake/. For a mesmerising display of the - at the time of writing - more than 600 aftershocks, visit http://www.japanquakemap.com/]