Science, No.11  May. 20, 2012


Photo : Goto Kazuhisa , Nishimura Yuichi , Shishikura Masanobu

Tsunami sediment research has garnered substantial attention since the tsunami triggered by the 2011 off the Pacific coast of Tohoku Earthquake and Tsunami (“2011 tsunami” hereinafter). This is because the 2011 tsunami was possibly a recurrence of the 869 AD Jogan tsunami that was a topic of sedimentology research for the last 20-plus years1. We know for sure that tsunami sediment (composed of various sizes of particles ranging from clay and sand to large boulder2) is effective for estimating when a past tsunami occurred and its size, and is potentially the only and critical proof for tsunami that occurred from prior to the Edo Period (1603-1868 AD) dating back to prehistory, for which historical records are especially thin. Yet tsunami sedimentology is not a mature field of study and it has numerous issues and technological limitations. We currently have very few research results that are qualitatively and quantitatively sufficient for incorporating into disaster prevention plans.

While we sense Japanese society’s increased expectations on tsunami sedimentology, we realize that unless we honestly explain the difficulties and limitations to the research, society may lose trust in this field of study over the medium-to-long term. We also need to let people understand that under the current state in which we lack experts in Japan, we will not be able to expeditiously meet society’s requests.

This paper discusses the current situation and issues of tsunami sedimentology and what we must do to develop as many case studies as possible that we can incorporate into disaster prevention plans.

2011 tsunami was not unexpected

The 2011 tsunami was not unexpected. Take, for instance, the Sanriku coast in Iwate Prefecture and we see that the tsunami-hit area and the run-up height were not too different from those of the 1896 Meiji Sanriku tsunami. Study the distribution of tsunami deposit and changes in its thickness in the Sendai plain and we see that they so greatly resemble those of the Jogan tsunami that it is hard to find differences3,4. Experts estimate that the 2011 earthquake was possibly an unprecedented type that resulted from joint action of the fault that caused the Meiji Sanriku earthquake and off the Sendai coast fault. While we may not have expected a magnitude 9.0 earthquake, the ensuing tsunami was, historically speaking, not particularly unique at all.

So then why had sediment studies on the Jogan tsunami not been reflected in any disaster prevention plans? We first must candidly admit that tsunami sedimentology itself is not a mature study and before the earthquake it had not reached a level of recognition that researchers and disaster prevention experts in various fields could have sufficiently understood. Another factor is that people were unable to comprehend the scale of time spanning from several hundred to 1,000 years that far exceeds human or architectural life at the pace of common society. We also must point out the fact that it took some experts 20 years to assess the potential tsunami deposit among the Jogan deposit and to bring the results to a level capable of being incorporating in disaster prevention.

Why does the study take so much time? Sediment that greatly resembles tsunami deposit is formed even from storm-induced high waves and from floods. Distinguishing this is not an easy task by merely comparing with that produced from recent events (Figure 1). Moreover, tsunami sediment of the past transforms over a long period of time and we have no direct identification criteria established at the current point. We cannot certify any particular deposit as owing to a tsunami unless we gather indirect evidence and eliminate every possibility of it being produced by an event other than a tsunami. It is because of this difficulty that the interest of many experts today is in knowing the characteristics of sediment produced from recent tsunami. In other words, the present is the key to understanding the past. And the global trend in today’s studies is to be exactingly careful in assessing sediment from the past as that of a tsunami when it is to link directly to disaster prevention.

re birth.JPG
Even under such circumstances, sediment study on the Jogan tsunami was the most advanced case study in the world in both quality and quantity. Minoura6 and Abe et al.7 first identified the Jogan tsunami deposit in 1990, after which the National Institute of Advanced Industrial Science and Technology (AIST) and Tohoku University played central roles in undertaking broader studies. By the time prior to the 2011 tsunami, they had conducted over 600 total drilling surveys and had clarified that on the Sendai Plain the tsunami had run up at least three to four kilometers inland from the coastline at that time8-10. Some factors played to the case’s advantage, such as coastal lowland that suited the survey and revealed deposit only by digging up to about one meter, and the fact that the stratum was immediately under a layer of volcanic ash (915 AD) and was easy to identify. Multiple research groups had all reached almost identical conclusions, including the estimate from numerical calculations that the magnitude of the earthquake was at least around 8.410-12.

Difficulties in identifying a tsunami: Issues and technological limitations

Circumstances concerning tsunami sedimentology changed drastically after the 2011 tsunami, with frequent reports by the media in Japan. The Central Disaster Prevention Council’s Special Research Committee had also touched on its importance in its advisory statement. We highly welcome this recognition. Yet the current state of study has not necessarily improved, and we are concerned that some information still regarding potential tsunami deposit and which has not gone through sufficient assessment is spreading throughout the public via the media, owing to the rapidly heightening expectations among society and local autonomies.

People must be warned that most case studies that are currently reported in Japan as of today are still at the “potential” stage and do not compare with the Jogan tsunami deposit studies in quality or quantity. The only other case that has gone through a level of research similar to that of the Jogan tsunami and is incorporated in disaster prevention measures would be the “500-year-span earthquake and tsunami” of (Eastern) Hokkaido13. Case studies on many other regions come with problems such as (1) limited numbers of survey points, (2) not yet having been evaluated by multiple research groups or (3) not yet discussed in any refereed journals (i.e., have not received objective evaluation of experts), and we still have very few cases today that experts widely approve as those of tsunami deposit.

We believe that one of the reasons studies are over-evaluated is that no one has correctly explained what can and cannot be done using tsunami deposit studies. When we discover tsunami deposit from a stratum, we know when in the past a tsunami had reached that point. It tells that water had run up to at least that particular point, and thus defines the minimum point of tsunami reach. In these terms it is obvious that tsunami deposit studies are extremely useful as physical evidence of historical or prehistoric tsunami.

Yet on the other hand it is difficult to discuss the recurrence span of earthquakes or tsunami based on coring surveys in several locations. This is because tsunami deposit does not necessarily accumulate continuously over lowlands and lakes/marshes due to the effects of microtopography. (Figure 2) We therefore need to run a synthetic judgment based on as many coring or trenching results as we can possibly obtain. And we can never discuss scale if we cannot reconstruct topography of the past, because between past and present, coastal topography and geological environments may have changed such that a dune or lagoon could have existed, or the location of the coastline may have moved significantly, and these factors influence the extent of a tsunami’s reach. We know, for example, that at the time of the Jogan tsunami, the Sendai plain coastline was located about one kilometer more inland than its current position9, and without considering this factor we could never compare its scale to the 2011 tsunami.

re birth2.JPG

The height of a tsunami changes depending on whether it comes at high or low tide, so if, for example, two tsunami of exactly the same scale (height and recurrence interval) had hit a particular point on land in the past and present, the resulting amount of earth and the size of sand particles washed up onto land would differ. The amount accumulated and particle size of sand in shallow waters and on the shore may have changed from past to present, so the total amount or particle size of sand washed up onto land does not necessarily indicate a tsunami’s scale. The activity or scale of a tsunami estimated from its deposit always comes with a margin of error owing to such uncertain factors.

Also, if two huge tsunamis had hit over a very short timespan, we may not be able to distinguish them by their deposit. For example, we know as an observed fact that the 1944 Tonankai Earthquake/tsunami and the 1946 Nankai Earthquake/tsunami occurred separately. But we also know from analyzing historical documents that joint action of the two faults responsible for each earthquake had caused several huge earthquakes. If we relied only on tsunami deposit, we could discover deposit dating around the 1940s spanning a broad area from the Tokai to Shikoku regions, but we would not be able to accurately tell whether the two earthquakes/tsunami occurred at the same time from joint action of the faults or separately in a span of anywhere from several hours to several decades, since modern dating technology has a margin of error of several decades. A tsunami also erodes soil on land, and when such tsunami-caused mud accumulates over sandy tsunami deposit, we might not be able to clearly distinguish how much of it is tsunami deposit, and this makes dating estimates difficult. These are issues we would face with every case regardless of its region or era. Even with the Jogan tsunami, it is likely that the fault area responsible for the Meiji Sanriku Earthquake was part of it, but even if tsunami deposit from around the ninth century was found around the Sanriku area, it still would not be easy to identify that the Meiji Sanriku fault area moved together with the fault off the coast of Miyagi and Fukushima Prefectures to cause the 869 AD Jogan Earthquake.

Moreover, even if we found deposit of multiple tsunami while sampling a certain point, it would not necessarily have been caused by the same epicenter (wave source) fault. Around the Pacific coast of the Tohoku region for instance, deposit could be formed from tsunami from far-field areas beyond the periphery of Japan, such as from the 1960 Chile tsunami. Even if we wanted to know the recurrence span of earthquakes/tsunami caused along the Japan Trench, tsunami deposit would contain that from far-field tsunami and we would incorrectly estimate the recurrence span of the earthquake/tsunami we were targeting. To eliminate the possibility of a particular type of deposit being that of a far-field tsunami, we have to reconstruct a detailed land level change of the time by analyzing diatoms, and prove that the tsunami occurred together with crustal movement. This simultaneity of tsunami and land subsidence had been proved in the Jogan tsunami case, in which we confirmed that the fault at the immediate plate border had moved9.

We highly welcome more ideas of putting “potential” tsunami deposit to use. If we notice the possibility of a major tsunami strike in the past, we will let the public know even if it is only at the “potential” stage; and it would be a significant warning. On the other hand, this could get out of hand when we consider the difficulties of certifying and interpreting tsunami deposit if we jump the gun and start incorporating potential-level deposit and ideas that could later be corrected into disaster prevention plans. What we need to do today is have multiple research groups run studies and obtain as many cases as possible that would widely be recognized as tsunami deposit, and to conduct fundamental research, and technological developments that will enable us to overcome the above issues. Researchers and disaster prevention experts must also discuss the process under which we would incorporate the studies into disaster prevention plans.

Concept of tsunami prediction

In the previous section, we explained the difficulties and technological limitations in identifying tsunami events from deposit and estimating their scale, source and earthquake magnitude. We expect it would take a significant amount of time for us to solve these issues. But even without solving all of them, if we obtain more cases of certified tsunami deposit, we should be able to put them to use in preventing tsunami-related disasters.

Considering, for example, that the 2011 case was a major disaster caused by a tsunami, we should consider predicting tsunami rather than insisting on predicting earthquakes. Particularly when we have yet to reach a seismological consensus on how we should set an upper limit for the magnitude scale, we should predict tsunami based on physical evidence. The magnitude of the Jogan Earthquake was significantly under-evaluated at 8.4, yet we knew that the ensuing tsunami reached three to four kilometers inland. We could have considered measures that predicted a tsunami of this level even without knowing the precise magnitude. Of course, we still have the problem of under-estimating the area of tsunami reach when we simply see the distribution of sandy tsunami deposit, so we need to undertake further research and technological innovation on how we would estimate the area of tsunami reach based on tsunami deposit.

We should also know that tsunami are not caused by earthquakes alone. In 1792, for example, Mount Mayu in Shimabara, Nagasaki Prefecture, collapsed due to the eruption of Mount Unzen’s Fugendake peak, sending massive amounts of earth into the Ariake Sea that caused a tsunami which killed 15,000 people on the opposite coast of Kumamoto Prefecture. By studying tsunami deposit, we could estimate causes, time period and inland reach of tsunami owing to non-earthquake causes to be used in planning safety measures.

We need to conduct surveys throughout Japan and immediate risk assessments on less-frequent tsunami of major scale. Considering future progress in the field, we now need to discuss assessment criteria and usage, and work on resolving issues. Tsunami sedimentology had lacked an organizing body for comprehensive discussion, with discussion taking place unsystematically among academic societies. We plan to hold a session called “Tsunami deposit” at the FY2012 conference of the Japan Geoscience Union, under the consent of the Volcanological Society, Society for Active Fault Studies, Seismological Society, Sedimentological Society, Association for Quaternary Research, Geological Society, and the Association of Japanese Geographers (listed in no particular order). We hope that this session will serve as an opportunity for experts to engage in deeper discussions and consider social contribution and advice on disaster prevention plans.

And if we are to conduct nationwide surveys in the future, we will need to conduct numerous coring surveys, yet the nation has only a handful of researchers specializing in tsunami sedimentology. We therefore require cooperation of researchers and technicians in relevant fields. And in the medium- to long-term run, we urgently need to train geological and geomorphological researchers and technicians specializing in tsunami sedimentology in order to accurately understand information on occurrences and the scale of past tsunami found in tsunami deposit in order to perform risk assessments, and take appropriate procedures to contribute the results to government administrations and the general public.


1Masanobu Shishikura, Tsugi no kyodai jishin wa dokoka! [Where will the next major earthquake be?], Miyaobi Publishing (2011)

2Yuichi Nishimura, Tsunami no jiten [Dictionary of tsunami], Asakura Publishing, 47 (2007) Chapter 2.1 Geological Survey

3Kazuhisa Goto et al., “Preliminary Papers to the Report Session on the Joint Survey Concerning the 2011 off the Pacific coast of Tohoku Earthquake and Tsunami,” 57 (2011)

Masanobu Shishikura et al., “Preliminary Papers to the Report Session on the Joint Survey Concerning the 2011 off the Pacific coast of Tohoku Earthquake and Tsunami,” 74 (2011)

5Kenji Satake et al., Kagaku [Science], 81, 407 (2011)

6Koji Minoura, Rekishi jishin [Historical earthquakes], 6, 61 (1990)

7Hisashi Abe et al., Jishin (dai 2 shu) [Earthquakes – Volume 2], 43, 513 (1990)

8Masanobu Shishikura et al., Katsudansou/kojishin kenkyuu houkoku [Report on studies on active faults and old earthquakes], 7, 31 (2007)

9Yuuki Sawai et al., Katsudansou/kojishin kenkyuu houkoku [Report on studies on active faults and old earthquakes], 8, 17 (2008)

10Daisuke Sugawara et al., Shizen saigai kagaku [The science of natural disasters], 29-4, 501 (2011)

11Kenji Satake et al., Katsudansou/kojishin kenkyuu houkoku [Report on studies on active faults and old earthquakes], 8, 71 (2008)

12Yuichi Namegaya et al., Katsudansou/Kojishin Kenkyuu Houkoku [Report on studies on active faults and old earthquakes], 10, 1 (2010)

13Kenji Satake, Futoshi Nanayama, Hokkaido taiheiyogan no tsunami shinsui rirekizu [Records of tsunami reach on the Hokkaido pacific coast], Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (2004)

14R.A. Morton et al.; Sedimentary Geology, 200, 184 (2007)

Translated from “Chishitsu kiroku wo tsunami bosai ni ikasu – tsunami taisekibutsu kenkyu no genjo to kadai,” Kagaku, February 2012 Vol. 82 No.2, pp. 0215-0219, ?2011 by Goto Kazuhisa, Nishimura Yuichi and Shishikura Masanobu. Reprinted by permission of the authors c/o Iwanami Shoten, Publishers.