Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, NATURAL HAZARD SCIENCE (naturalhazardscience.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 23 September 2017

Palaeotsunamis

Summary and Keywords

How big, how often, and where from? This is almost a mantra for researchers trying to understand tsunami hazard and risk. What we do know is that events such as the 2004 Indian Ocean Tsunami (2004 IOT) caught scientists by surprise, largely because there was no “research memory” of past events for that region, and as such, there was no hazard awareness, no planning, no risk assessment, and no disaster risk reduction. Forewarned is forearmed, but to be in that position, we have to be able to understand the evidence left behind by past events—palaeootsunamis—and to have at least some inkling of what generated them.

While the 2004 IOT was a devastating wake-up call for science, we need to bear in mind that palaeotsunami research was still in its infancy at the time. What we now see is still a comparatively new discipline that is practiced worldwide, but as the “new kid on the block,” there are still many unknowns. What we do know is that in many cases, there is clear evidence of multiple palaeotsunamis generated by a variety of source mechanisms. There is a suite of proxy data—a toolbox, if you will—that can be used to identify a palaeotsunami deposit in the sedimentary record. Things are never quite as simple as they sound, though, and there are strong divisions within the research community as to whether one can really differentiate between a palaeotsunami and a palaeostorm deposit, and whether proxies as such are the way to go. As the discipline matures, though, many of these issues are being resolved, and indeed we have now arrived at a point where we have the potential to detect “invisible deposits” laid down by palaeotsunamis once they have run out of sediment to lay down as they move inland. As such, we are on the brink of being able to better understand the full extent of inundation by past events, a valuable tool in gauging the magnitude of palaeotsunamis.

Palaeotsunami research is multidisciplinary, and as such, it is a melting pot of different scientific perspectives, which leads to rapid innovations. Basically, whatever is associated with modern events may be reflected in prehistory. Also, palaeotsunamis are often part of a landscape response pushed beyond an environmental threshold from which it will never fully recover, but that leaves indelible markers for us to read. In some cases, we do not even need to find a palaeotsunami deposit to know that one happened.

Keywords: palaeootsunamis, prehistory, geology, archeology, anthropology, tsunamigenesis, hazard, risk, magnitude, frequency

What Are Palaeotsunamis?

A palaeotsunami is defined as a tsunami occurring prior to the historical record or for which there are no written observations (IOC, 2013). While it may be a moot point, the line between historical tsunamis and palaeotsunamis is not quite as clear cut as the definition suggests. Internationally, there are many hybrid tsunamis—ones that have a place in both historical and paleotsunami archives. The disjunct between the longevities of national historical databases creates instances where a paleotsunami in one country is an historical event in another. The most well-known hybrid case is probably either the 1700 Cascadia tsunami, which is a paleotsunami in North America and an historical one in Japan; or the 1604 Chilean tsunami, which is a paleotsunami in New Zealand and historical in Chile (Goff et al., 2010a; Goff, Nichol, & Kennedy, 2010b).

Paleotsunami research is most commonly based on the identification, mapping, and dating of deposits found in coastal areas, and their correlation with similar sediments found elsewhere locally, regionally, or across ocean basins. However, as will be seen, this is a somewhat simplistic explanation in the sense that the science has progressed to the point that under the right circumstances, it is not only possible to identify “invisible deposits,” but also to know that an event occurred even though there is no deposit, invisible or otherwise.

It is useful to know not only the definition of a paleotsunami, but also the reason why these deposits are so important. One of the most difficult issues in assessing hazard (and ultimately risk) is having some understanding of the environmental context—how big have events been in the past, and how often have they happened? In the context of the tsunami hazard and risk, it is also helpful—but difficult—to get a sense of the source of past events as well. For example, where do the most frequent big ones come from?

The Evolution of Paleotsunami Research

“… the effect of the great waves, which generally accompany earthquakes, must not be overlooked: few years pass without a severe earthquake occurring on some part of the west coast of South America; and the waves thus caused have great power. At Concepcion, after the shock of 1835, I saw large slabs of sandstone, one of which was six feet long, three in breadth and two in thickness, thrown high up on the beach; and from the nature of the marine animals still adhering to it, it must have been torn up from a considerable depth. On the other hand, at Callao, the recoil-wave of the earthquake of 1746 carried great masses of brickwork, between three and four feet square, some way out seaward. During the course of ages, the effect thus produced at each successive level, cannot have been small; and in some of the tertiary deposits on this line of coast, I observed great boulders of granite and other neighbouring rocks, embedded in fine sedimentary layers, the transportal of which, except by the means of earthquake-waves, always appeared to me inexplicable.”

(Darwin, 1846)

Paleotsunami research is in reality an extremely young discipline, dating to about the mid-1980s, to the works of Moore and Moore (1984), covering an about120-ka submarine landslide-generated megatsunami off the slopes of Hawaii; Atwater (1987), discussing a Cascadia earthquake and tsunami in the Pacific Northwest in 1700; Bourgeois, Hansen, Wiberg, and Kauffman (1988), describing an asteroid impact that generated a megatsunami at about the 65-Ma K-T boundary in Texas; and Dawson, Long, and Smith (1988), covering a Storegga submarine landslide-generated tsunami in Scotland that occurred about 7,000 years BP. However, there are several notable earlier publications that had either speculated or attributed a variety of deposits to a paleotsunami origin, and so in a sense, the mid-1980s represents the final “coming of age” of paleotsunami research. Earlier publications of note include the following:

  • Hall (1815), who proposed that rare elevations of the land could form moderate waves (the term tsunami was not used in the paper), such as the ones that destroyed Lisbon in 1755, or even larger ones of overwhelming magnitude.

  • Bailey and Weir (1933), who inferred a paleotsunami (called a tunami in their paper) origin for some Jurassic sediments in Scotland, reinterpreting the earlier freshwater flood interpretation of Judd (1873).

  • Coleman (1968), who speculated that tsunami action may well explain unusual turbidite deposits in submarine canyons.

  • Wood (1967), who suggested that high-elevation Holocene beach ridges in the Cook Islands may have at least partially been emplaced by tsunamis.

  • Kastens and Cita (1981), who proposed a paleotsunami origin for an unusual deposit in the Mediterranean Sea dating to around 3100–4400 years BP.

What is most intriguing about all the publications listed here is that the earlier ones do not generally identify a source for the proposed paleotsunami, but rather relate to the unusual nature or location of a deposit. The mid-1980s, therefore, represents something of a watershed, with researchers strengthening their arguments by offering proposed sources for their paleotsunami interpretations—and there are many options. Asteroid/bolide impact, landslides (submarine or subaerial into the sea), volcano-related events (e.g., eruptions, pyroclastic flows, and caldera collapse), and fault ruptures. The fact that a variety of sources are introduced early in paleotsunami studies is important because it establishes that this is a distinct discipline in the wider field of tsunami research and not, as many paleoseismologists tend to think, a subdiscipline of earthquake research. Indeed, paleoseismology is the study of past earthquakes, and while it is true that many paleoseismologists study the evidence of past earthquake-generated tsunamis, there is more to paleotsunami research than just earthquakes—as such, we risk greatly underestimating the tsunami hazard if we limit ourselves to this narrow seismological vision (Chamberlin, 1890; Wegman, 1939).

Following the initial consolidation of paleotsunami research in the mid-1980s, there was a gradual global uptake over the ensuing years, although the extent of this growth is not entirely clear, largely because the term paleotsunami does not appear to have been coined in a peer reviewed paper until Dawson (1994) first used it to describe prehistoric tsunami deposits discussed by earlier researchers [Bondevik & Svendsen (1993) used it in their part of a report to the European Union a year earlier]. It might be argued that Aida (1977) first used the term in Japan, but recent citations of his work as “Paleo-tsunami Simulation Along Sanriku-Oki” are incorrect because it was originally called “Simulations of Large Tsunamis Occurring in the Past off the Coast of the Sanriku District.” While this may seem to be a minor issue, it all helps to cloud the early development of paleotsunami research.

It is perhaps easier to search databases of peer-reviewed literature for the terms palaeotsunami (the British spelling of paleotsunami) or paleotsunami. As noted, these searches reveal that they started being used in 1994, with a notable tripling of publications to about 15 per year in 2005 [a rapid research response to the need for more paleotsunami data following the 2004 Indian Ocean Tsunami (2004 IOT)] and an almost doubling again in 2012 [the post-2011 Tohoku-oki Tsunami (2011 TOT) in Japan] (Scopus, 2016). Noting that the mid-1980s was really the decade that put paleotsunamis on the map in research (if not by name), a simple data search for the term tsunami reveals a similar publication growth pattern; however, it not only extends back further in time, but also contains a significantly greater number of publications. According to Scopus (2016) peer-reviewed references to the term first started in 1929, although at least two earlier ones (Anonymous, 1905; Platania, 1909) actually push this back at least two more decades, and then published, non-peer-reviewed material written in English taking it as far back as 1896 (Scidmore, 1896), and in Japanese to 1524 (Goff, Goto, Ebina, & Terry, 2016). The number of publications retained in the database should therefore be considered an underestimate of the total amount, but there is a steady climb in research following events such the 1960 Chilean, 1964 Alaskan, 1976 Philippines, and 1994 Java tsunamis, followed by an almost order-of-magnitude increase after the 2004 IOT and a doubling after the 2011 TOT. In essence, changes in paleotsunami research are reflected in the overall study of tsunamis, with the not-unreasonable linkage between major historical tsunamis and the need to better understand how big and how often such events have occurred in the past.

Paleotsunami Deposits

By far, the greatest number of paleotsunami deposits have been studied in terrestrial as opposed to marine settings. This is not surprising given that the main concerns about hazard and risk relate to coastal infrastructure, residential, and industrial areas, not to sea-floor installations. It is also far more expensive to do the work in a marine environment. However, another important point to consider is that past tsunami inundation of the land would normally have been as a short-term, high-energy, episodic event as opposed to everyday ongoing processes. This means that it is reasonable to conclude that some evidence will be likely be preserved on land since ongoing processes are unlikely to remove it all. In the offshore environment, on the other hand, ongoing processes such as large storms, near-shore currents, and a variable wave climate invariably rework much, if not all, of the evidence of past events, thus complicating site selection in an already-expensive research endeavor.

Paleotsunamis are essentially events that occurred prior to the written record, but by analogy, they potentially carry many of the attributes associated with known historical ones. Notable differences tend to relate to the relative preservation of deposits over time or the “tsunami taphonomy” of a deposit (Goff, Chagué-Goff, Nichol, Jaffe, and Dominey-Howes, 2012a). As mentioned previously, there are hybrids that may be prior to the written record of one country but are historical in another, thus giving the possibility of correlating the two by using a variety of chronological techniques to date the paleotsunami deposit. There are actually two additional types of hybrid. First, there are tsunamis that have occurred during the historical period of a country, but for which there are no written records. Second, there are historical tsunamis where documentation is either rare or sufficiently vague that the nature and extent of the event are poorly understood. In essence, the comprehensive documentation of these hybrid types tends to use a suite of paleotsunami criteria (geological, chemical, biological, archeological, anthropogical, geomorphological, and contextual) that are discussed in Table 1.

Perhaps the key difference between historical tsunamis and paleotsunamis is the reversal of the key parameters used to define an event. In the case of an historical tsunami, the event is known and recognized by the date that it occurred (e.g., December 26, 2004). Details relating to this known event are then reported from specific geographical locations, and information concerning the source and magnitude is normally readily available through instrumental or historical data. For a paleotsunami, an event is generally known and recognized by the geographical location of the original evidence. Chronological data rarely, if ever, define a specific year, and as such, it is not until many sites have been studied that a better understanding of the timing, potential source, and magnitude of a paleotsunami can be estimated.

This reversal of event definition can prove problematic when attempting to determine tsunami hazard and risk for a geographic area. Paleotsunami data do not fit neatly within the adopted conventions for written historical tsunami information, so collating data or comparing them is difficult. For example, a paleotsunami has no measured run-ups, no written eyewitness accounts, and no specific calendar date (Goff et al., 2012a). While the ability to estimate these parameters more accurately will undoubtedly improve as we understand more about how to identify the processes of paleotsunami inundation, these disparities will always exist because estimates will be based upon different methodologies.

Over the past three decades or more, there has been a growing data set of literature reporting useful criteria or proxies that may help identify paleotsunamis (e.g., Atwater, 1987; Pinegina & Bourgeois, 2001; Morton, Gelfenbaum, & Jaffe, 2007). However, there has also been a growing body of literature pointing out the difficulties of differentiating between paleotsunami and paleostorm deposits (e.g., Bridge, 2008; Engel & Brückner, 2011; Shanmugam, 2012). As such, it is probably best to start with a relatively up-to-date list of proxies that can be used and work through the debate from there (Table 1).

Table 1. A List of Paleotsunami Proxies Developed Through a Series of Published Iterations

Geological

1. Particle/grain sizes range from boulders (may be 750 m3 or larger) to fine mud. A tsunami will usually transport whatever size ranges are available—it is dependent on the sediment source.

2. Sediments generally fine inland and upward within the deposit, but can also have a coarsening upward component associated with deposition from a traction carpet. Deposits generally rise in altitude and thin inland and can extend for several kilometers inland and tens or hundreds of kilometers along the shore.

3. Each wave can form a distinct sedimentary unit and/or there may be laminated subunits.

4. Distinct lower and upper subunits representing run-up and backwash can sometimes be identified, unlike with storm or anthropogenic deposits.

5. Lower contact is usually unconformable or erosional.

6. Can contain intraclasts (rip-up clasts) of reworked material.

7. Sometimes associated with loading structures at base of deposit—and can be associated with liquifaction features on the ground surface caused by earthquake ground-shaking.

8. Microscale features can include micro-rip-up clasts, millimeter-scale banding, organic entrainment, fining-up sequences, and erosive contacts that may be visible in thin sections, but not in field stratigraphy.

9. Measurement of magnetic fabric (MF) combined with grain size analysis provides information on hydrodynamic conditions typical during tsunami deposition. Essential when no sedimentary structures are visible. Magnetic properties of minerals (including magnetic susceptibility) provide information about depositional environment.

10. Heavy mineral laminations often present, but they are source-dependent. Normally near the base of unit/subunit, but not always. Composition and vertical distribution of heavy mineral assemblage may change from the bottom to the top of the deposit (e.g., often more micas at the top).

Chemical

11. Increases in elemental concentrations of sodium, sulfur, chlorine (paleosalinity indicators, including element ratios), calcium, strontium, magnesium (shell, shell hash, and coral), titanium, and zirconium (associated with heavy mineral laminas if present) occur in tsunami deposits relative to underlying and overlying sediments. Indicates saltwater inundation, high marine shell/coral content, and/or a high-energy environment (heavy minerals, source-dependent). Preservation issues to be considered, particularly for salt (downward leaching), but uptake and preservation in wetlands/soils.

12. Possible contamination by heavy metals and metalloids (source-dependent, including the water depth source).

13. Geochemical (saltwater signature) and microfossil evidence often extends farther inland than the landward maximum extent of sedimentary deposit.

Biological

14. Individual shells and shell-rich units are often present (shells are often articulated and can be water-worn). Often more intact shells, as opposed to shell hash. A wide range of shell ages is indicative of greater reworking by a tsunami, as opposed to storm or anthropogenic deposits. Small, fragile shells and shellfish can be found at or near the upper surface of more recent paleo-tsunami deposits.

15. Shell, wood, and less dense debris often found rafted at or near the top of the sequence (increase in organic content determined by loss on ignition, and sometimes moisture content).

16. Often associated with buried vascular plant material and/or buried soil and/or skeletal (nonhuman) remains.

17. Generally associated with an increase in abundance of marine to brackish diatoms—often a greater percentage of reworked terrestrial diatoms near the upper part of the deposit. Large number of broken valves often observed, reflecting turbulent flows. Variations in diatom affinities often indicative of source areas and the magnitude of event.

18. Marked changes in foraminifera (and other marine microfossils, such as dinoflagellates and nannoliths) assemblages occur. Deeper water species are introduced—this is unlikely in storm or anthropogenic deposits, and/or increase in foraminifera abundance and breakage of tests. Composition relates to source (near-shore versus offshore). Foraminifera size tends to vary with grain size.

19. Pollen concentrations are often lower (diluted) in the deposit because of the marine origin and/or include a high percentage of coastal pollen (e.g., mangroves). Pollen changes above and below the deposit are often indicative of sustained environmental change and that a critical ecological threshold has been crossed—e.g., infilling or shallowing of coastal wetland.

Archeological

20. Archeological sites—a sediment layer separating, underlying, or overlying anthropogenic deposits/occupation layers.

21. Archeological middens: changes in shellfish species/absence of expected species indicate sudden change in onshore and nearshore paleo-environmental conditions.

22. Archeological structures show structural damage by water to buildings/foundations at a site.

23. Archeological burial sites have been reworked, often recognizable as culturally inappropriate burials.

24. Replication—coastal archeological occupation layers and shell middens are often separated or extensively reworked at several sites along the coastline, giving a regional/national signal of inundation.

Anthropological

25. Traditional environmental knowledge (oral traditions) from the locality/region.

Geomorphological

26. Acquired paleogeomorphology indicates tsunami inundation—a tsunami geomorphology is present that could include evidence of (a) uplift or subsidence/compaction of site/locality, (b) scour/erosion/reworking of sediments at site/locality—altered dune morphology, and (c) sand sheet or other similar deposits, such as gravel deposition/gravel pavements.

27. Paleogeomorphology at the time of inundation indicates low likelihood of storm inundation

Contextual

28. Known local or distant tsunamigenic sources can be postulated or identified.

29. Known local and regional paleoenvironmental drivers indicate low likelihood of storm inundation.

30. Replication—similar contemporaneous coastal deposits are found regionally, giving a regional signal of inundation.

Note: Modified after Goff et al. (2012a).

Research continues to reveal exciting new techniques that can be applied to the study of paleotsunamis. While Table 1 presents a reasonably comprehensive summary of techniques currently available, new studies, such as those of paleo-DNA within purported paleotsunami deposits, continue to add to this “toolbox.” Given the ongoing work being carried out on paleotsunami deposits, it seems reasonable to ask why there is such a significant debate surrounding the ability to differentiate them from paleostorms.

Pinegina and Bourgeois (2001) point out that the tsunami deposits are not uniquely identifiable. Indeed several other types of deposit undoubtedly share one or more of the same characteristics “but will not share all.” This sentiment was echoed by Goff, Nichol, and Kennedy (2010b) when developing a paleotsunami database for New Zealand, in that they stated that the more proxies indicating paleotsunami inundation as opposed to paleostorm, the more valid the interpretation. It is perhaps unfortunate that the paleotsunami-paleostorm debate has become so heated so quickly in the life of the discipline because it distracts somewhat from the ongoing research. The main points made by those advocating the ability to differentiate between these two inundation processes are the following:

  • The local and regional geological and meteorological contexts must be taken into consideration (e.g., the nature of local storm climate/potential tsunamigenic sources).

  • A large tsunami is invariably the result of a significant environmental perturbation or driver. This driver (e.g., an earthquake) is associated with a series of other environmental aftereffects. Large earthquakes can cause regionwide landsliding in the surrounding catchments, disturb vegetation patterns, cause rivers to aggrade and new coastal dunes to form due to increases in sediment, and displace communities. These effects can be contemporaneous across regions, across adjacent countries, and at oceanwide scales (e.g., Goff & McFadgen, 2002). Such a cascade of environmental aftereffects has not been documented for even the largest historical storms.

  • Local factors determine what the most significant differences will be between them—variables such as source material, nearshore bathymetry, coastal geomorphology, and depositional environment play key roles (Fig. 1).

  • There are many proxies, and while in some cases differentiating between the two processes is simple (e.g., deposits found a significant distance inland from the paleoshoreline on a sheltered coast), there are many instances where either many lines of evidence (proxies) are needed to tentatively determine the process involved, or it is impossible for an assortment of reasons to distinguish between events such as a small paleotsunami and a large paleostorm. In all such instances, the onus is on researchers to qualify their interpretations, using terms such as possible, probable, or unusual (e.g., Goto, Miyagi, Kawamata, & Imamura, 2010; Atwater et al., 2012; Yawsangratt et al., 2012).

  • In most paleotsunami research, there is an overwhelming emphasis on the geological and paleontological evidence for paleotsunamis—in other words, the physical evidence available within the deposits. However, we know from historically documented events that tsunamis also do the following:

    • o Have the ability to erode coastal sediments, so possible erosional evidence needs to be investigated.

    • o Are often associated with large earthquakes that can cause uplift or subsidence or the land, in which case the association of a candidate deposit with a subsided or uplifted soil most likely indicates tsunami inundation.

    • o Affect humans and their settlements—in prehistory, this type of catastrophic event is beyond everyday storminess and can be reflected in archeological and anthropological data.

  • The use and/or perceived relevance and importance of individual proxies indicate a strong bias toward the expertise of those involved in the research. While this is understandable, it gets to the heart of the argument concerning paleotsunamis and paleostorms—invariably, an insufficient number of proxies are used to be definitive about the formative process.

PalaeotsunamisClick to view larger

Figure 1. Local factors determine what a paleotsunami deposit might look like, how to identify it, and why it is not related to a paleostorm. (A) Tirua, Chile (river estuary): Known historical events act as analogs for earlier tsunamis. In this case, they are all historical sandy deposits, but the oldest, from 1575, is actually a hybrid. (B) Pigeon Bay, New Zealand (coastal wetland): a 1575 paleotsunami (hybrid) composed of coarse clasts with a shell hash matrix. (C) Saltwater Lagoon, New Zealand: a 900-year-old paleotsunami deposit composed of silt and very fine sand associated with a drowned forest.

Photos: J. Goff.

From Simple to Complex: Paleostorm Versus Paleotsunami

Goff, Chagué-Goff, and Nichol (2001) listed a set of 15 diagnostic criteria for identifying paleotsunami deposits, stating that the identification depended upon the deposit possessing a suite of these criteria, not just a single attribute. This point was made previously, but it does not suffer from being repeated. The use of the term diagnostic criteria was later changed to proxies, and the list has subsequently grown to 30 (Table 1; Goff et al., 2012a). An interesting question is: How could the list of proxies double in a mere 11 years? The answer to this question covers several important observations about paleotsunami research and goes some way toward answering the debate about paleotsunami versus paleostorm deposits.

First, this is a rapidly developing discipline, and as such, considerable strides have been made in the past decade to better understand the nature of individual proxies, while at the same time new ones have been developed. Has this improved our abilities to differentiate between the two processes? Yes. For example, Kortekaas and Dawson (2007) reported differences between the 1755 Lisbon tsunami and earlier storm deposits in Portugal. At the time, they noted that the tsunami contained boulders and rip-up clasts and a higher concentration of foraminifera, although the latter was not considered to be particularly useful. This paper, among others, engendered a brief debate between Bridge (2008) and Jaffe et al. (2008); the skeptical sedimentologist (Bridge) took issue with the evidence provided by Kortekaas and Dawson (2007), essentially suggesting that boulders and rip-up clasts were not exclusively the domain of tsunami deposits. In reply, Jaffe et al. (2008) indicated (quite rightly) that the differences are dependent upon local conditions—in this instance, it was boulders and rip-up clasts that were different, while elsewhere, it might be something else. There is no “silver bullet,” but rather a wider selection of potential differences that need to be identified. Indeed, as we learn more about tsunami deposits, it has become apparent that tsunamis are unlikely to be capable of transporting foraminifera from deeper than about a 900-m water depth (Weiss, 2008). However, this means that a tsunami could transport foraminifera up to about a 900-m water depth, whereas a storm’s wave base (ability to move sediment) is half its wavelength, which is unlikely to exceed 250–300 m at the very most and as such, is unable to access deeper sediments. This is relevant to the debate about paleotsunami versus paleostorm because it provides an example of how one can use the proxies to differentiate between the two types of deposits. Kortekaas and Dawson (2007) noted that the 1755 tsunami deposit had a higher concentration of foraminifera than that of the storm, but they did not consider this particularly helpful. On closer inspection, there are not only more foraminifera, but the assemblage is markedly different. The proposed storm deposit is dominated by intertidal species, whereas the inferred 1755 tsunami sediments contain a fully marine assemblage. The devil is in the details. Assuming that we have good resolution data, in the early 21st century, we now know more about individual proxies and are able to infer more from the data (Goff et al., 2012a).

Second, in Goff et al. (2001), brief mention was made about geochemical signatures and their potential to be used to identify paleotsunami deposits, but there has subsequently been a considerable amount of new research in the area, which has sadly been missed by many northern hemisphere researchers, largely because until recently, the bulk of the innovations had been developed and published by southern hemisphere scientists. This led to somewhat perfunctory dismissals of the technique by sedimentologists as recently as 2012 (Engel & Brückner, 2011; Shanmugam, 2012). Ignoring the importance of site context and local conditions, Engel and Brückner (2011) suggested that a paleosalinity signal in a deposit simply represented marine flooding—it could be a storm or a tsunami. Shanmugam’s (2012) brief dismissal of the technique used only three examples highlighting the differences in elemental data collected from sandy deposits and inferring that postdepositional digenetic alterations made the data meaningless. It is interesting to note that at least a decade earlier, researchers (albeit in the southern hemisphere) had noted the strong affinity between fine sediments/organic matter and key elements indicative of salinity (e.g., Chagué-Goff & Goff, 1999; Chagué-Goff et al., 2002). There had also been a thorough review of all tsunami geochemistry prior to 2011/2012 (Chagué-Goff, 2010). One of the key take-home messages of this work was that it is not usually the coarse sediment of the deposit that provides a useful geochemical signature, but the less glamorous material that tended to be overlooked—namely, muds, organic matter and the underlying soil.

Innovations in both analytical techniques and interpretation have now lifted geochemistry into one of the key proxies to use. Unlike its grain-size and microfossil counterparts, it is no longer necessarily an invasive technique that requires physical samples, destruction of the material being studied, and a relatively coarse resolution-sampling regime. The work can be carried out noninvasively at a submillimeter level using a core scanner that combines x-ray fluorescence (XRF), radiographic x-ray and optical imaging, and magnetic susceptibility measurements to provide detailed geochemical information. This has revolutionized the process of core analysis for paleotsunamis in several ways:

  • It allows the core to remain untouched, with all the sediment then being available for additional coarser resolution analyses such as grain size and microfossils.

  • It can be done first, with core stratigraphy being logged at the same time. The importance of this innovation cannot be overstated—in the past, researchers have based their subsampling regime on what they have “seen” when the core was logged and from field observations. Having a continuous, potentially submillimeter record of core geochemistry not only provides data to guide the subsampling regime of the visible stratigraphy but can also reveal invisible (to the naked eye) units of interest. This significantly enhances the subsampling decision-making process and has the potential to identify previously overlooked event horizons (e.g., Chagué-Goff, Chan, Goff, & Gadd, 2016).

  • Statistical analyses can now be performed on the large quantity of data produced, enabling more subtle differences between marine signals to be identified.

Recent work in the Cook Islands provides an excellent example of the significance of these advances. A 4.3-m core taken from a lake on the island of Mangaia revealed a sequence of dark brown to black peat, with only minor visible variations in texture and color. Core scanning revealed complex geochemical variability, with many discrete, and yet invisible, spikes in elemental data. As a result, a targeted subsampling regime was undertaken for grain size and microfossil analyses to investigate this variability in the context of local conditions. When these data were combined with the results of statistical analysis of the scanned geochemical data, they not only identified essentially invisible deposits, but they also could be identified as either paleotsunami or paleostorm events (Figure 2) (Chagué-Goff et al., 2016).

PalaeotsunamisClick to view larger

Figure 2. Mangaia, Cook Islands: (A) Summary core data—from left to right, this includes: ITRAX core scanner data — core photo showing mainly color variations throughout a moderately ubiquitous peat sequence, x-ray sequence showing slightly more structure associated with varying densities of material, magnetic susceptibility recording the presence of magnetic materials, summary of high resolution counts of seven elements (out of 30), molybdenum incoherent/coherent ratio that is a proxy for organic content; low-resolution summary grain size data; low-resolution diatom data. (B) Results of principal component analysis of all elemental data revealing distinct differences between tsunami and storm (cyclone) deposits. Modified from Chagué-Goff et al. (2016).

Third, as noted previously, the study of paleotsunamis is multidisciplinary and as such, it is not solely the domain of the sedimentologist. Rather, it encompasses a melting pot of different scientific perspectives, which leads to rapid innovations that can be difficult to keep up with. Unfortunately, one of the growing pains of paleotsunami research has been the tension between two groups—paleoseismologists and sedimentologists. The former focus solely on earthquake (predominantly subduction) sources and their associated deposits, and the latter almost exclusively on the physical features of potential paleotsunami deposits. This has led to misconceptions, the most obvious of which are the following:

  • In the absence of either an associated uplifted/subsided soil and an adjacent subduction zone, it is difficult to identify paleotsunami deposits.

  • Sedimentologically, paleotsunami deposits cannot be identified because they are almost indistinguishable from those of paleostorms.

Some of these points have already been addressed, but a wide gulf remains between what paleotsunami research actually entails and what individual researchers consider it to be, of which the perceptions of paleoseismologists and sedimentologists are a prime example. Having said this, since paleotsunami research is such a rapidly developing discipline, it is difficult to know whether all aspects of it are at the fingertips of any single researcher. The best way to understand the nature of the work is to recognize that, as stated previously, almost everything that is associated with a modern tsunami will be reflected in its prehistoric equivalent. A simplistic point perhaps, but it is remarkable how few paleotsunami researchers collaborate with archeologists and anthropologists. Yet the key reason we undertake this type of work is to better understand the hazard and risk to human populations and coastal infrastructure.

The 2011 TOT is a perfect example of what happens in a modern event, with widespread abandonment of coastal settlements, building damage, and regionwide loss of food resources from tsunami erosion/deposition and the salinization of soils (Chagué-Goff et al., 2014). Equally, though, beyond the reconfiguration of the coastal landscape by tsunami erosion and deposition, there was regionwide, earthquake-generated landsliding in the mountainous catchments of central Honshu, which blocked waterways and severed communication routes (Wartman, Dunham, Tiwari, & Pradel, 2013), and there are many tales of lost communities, human tragedy, and bravery that occurred while all this was taking place (e.g., Jiji Press, 2012).

Context

Paleotsunami research is, therefore, more than seismology and more than sedimentology, and this is reflected in the marked increase in proxies related to archeological, anthropological, geomorphological, and contextual information. Once researchers fully recognize this, it engenders several fundamental changes in approach. There is a need for multidisciplinary groups, simply because no single researcher has the necessary skills to encompass all the expertise needed to address the question—was this a paleotsunami or something else? If this question is approached simply from the point of view of context, without any consideration of the deposit (or absence thereof), then we can look to historical events to guide this work. First, let us assume that this hypothetical location is in a tectonically active area (context); as such, if a paleotsunami is preserved, then, like known historical events, it is highly likely that it was associated with a suite of immediate and delayed environmental aftereffects (Figure 3). We have already noted widespread landsliding following the 2011 Mw 9.0 Tohoku-oki earthquake, but it also caused regionwide coseismic subsidence near the coast of the Sendai Plain, eroded coastal dunes, and killed crops through erosion/deposition and/or saltwater inundation (Goto et al., 2011). Recognition of such environmental response linkages is not new. For example, following an earthquake, Moseley, Wagner, and Richardson (1991) charted the formation of a new beach ridge at the mouth of the Santa River catchment in arid coastal Peru using space shuttle and other high-altitude imagery. Here, the suite of responses is in a sense taken from source (landslides) to sink (beach). A relatively small, Mw 7.7 earthquake in 1970 caused numerous landslides that introduced excess sediment in to the river catchment, but it was not until the strong El Niño of 1972–1973 that the river was capable of transporting the bulk of this material downstream. By 1974, a new sand beach ridge had formed. The transfer of material was rapid in this instance, a matter of four years, but this process was enhanced by intense rainfall and the absence of an extensive low-profile floodplain between the mountains and the sea, which would have slowed sediment transport. These same process responses have also been noted in the geological record of prehistoric events, with notable variations in the timing of source-to-sink sediment transfer caused by the width of the floodplain and rainfall. For example, in southwest South Island, New Zealand, the same process has taken 10–70 years in an area of high rainfall and narrow floodplains (Wells & Goff, 2006, 2007). On the east coast, with lower rainfall and wider floodplains, the same process has taken up to 200 years (McFadgen & Goff, 2005). Somewhere between these two rates, sediment transfer on the Sendai Plain in Japan probably takes 20–100 years (Goff & Sugawara, 2014). More important, given the right conditions, this suite of environmental aftereffects can be traced back in time over multiple events through extensive beach ridge systems, river aggradation surfaces, and landslide scars. In essence, what we see in the historically tsunami-prone areas of Japan and Peru can be seen elsewhere in the world and serve as a guide for where paleotsunamis may well have occurred in the past.

PalaeotsunamisClick to view larger

Figure 3. The seismic driving model showing possible immediate and delayed physical environmental and human responses (adapted from Goff & McFadgen, 2002; Goff et al., 2016).

A tsunami is one of the immediate aftereffects in this cascade of responses that can be traced through the landscape over a time period ranging up to hundreds of years (Figure 3). Knowing the chronological sequence of these events can help researchers to place the potential paleotsunami within the context of this cascade (e.g., Goff & Sugawara, 2014). Needless to say, not all the evidence related to every step of this process is necessarily preserved in the landscape, not all tsunamis are generated by earthquakes (as stressed earlier), and this sequence of responses is only potentially relevant to locally or regionally generated events. However, once the concept of context is taken on board, it becomes a useful tool for identifying paleotsunamis generated by other mechanisms such as volcanic eruptions and submarine landslides. For example, if a potential paleotsunami deposit immediately overlies or is mixed with a volcanic ash, then while it is always plausible that there may have been a major storm at the same time as the ash was laid down, a tsunamigenic source seems more reasonable (e.g., Bruins et al., 2008). Equally, a potential paleotsunami deposit that covers only a short length of coastline may well be associated with a submarine landslide source, as was experienced by historically documented events such as the 1998 Papua New Guinea and 2010 Mangaia tsunamis (Tappin, Watts, & Grilli, 2008; Goff, 2011). In such instances, it is important to search for potential tsunamigenic slope failure sources nearby at the very least (e.g., Tappin et al., 2008; Terry & Goff, 2013).

It is harder to apply contextual proxies to potential paleotsunami deposits that may have been from distantly generated sources such as those known from numerous historically documented trans-Pacific tsunamis (e.g., the 1946 tsunami in Hilo, Hawaii, caused by an Aleutian earthquake, described in Macdonald, Shepard, & Cox, 1947). However, there are some useful pointers, not the least of which involves referring to the historical record to identify recurrence intervals of past tsunamis and storms. Furthermore, given the longer historical tsunami records of some countries (e.g., Japan—almost 1,400 years) and the relative hazard of others to waves from that region (e.g., the west coasts of the Hawaiian Islands), a hybrid event may be entirely plausible. Indeed, in many instances, such analogs have been drawn using numerical modeling of potential sources to link a possible paleotsunami deposit to a known historical source (e.g., Butler, Burney, & Walsh, 2014). Again, it is important to recognize that the paleotsunami researcher must draw on much more than simply the sedimentary evidence left behind, and indeed, the ability to ground-truth numerical models effectively adds immense value to both methodological approaches to understanding the tsunami hazard for local, regional, or global coastlines.

Lost Communities and Human Tragedy

Historically, one can draw on numerous instances where large tsunamis have caused widespread destruction of coastal communities, salinization of soils, loss of usable agricultural land for multiple seasons, the abandonment of low-lying areas, horrific death tolls, and the birth of tales of bravery in the face of adversity. It is also evident that other immediate environmental aftereffects of the tsunamigenic mechanism have disrupted communities with landslides that block communication routes and overlying productive land, subsidence that causes waterlogging of soils, uplift destroying intertidal shellfish beds, ground-shaking, and liquefaction that causes buildings to collapse and flooding by diverted rivers. The 2011 TOT is the most recent example, and even with the benefits of modern technology, most of the inundated low-lying coastline is still deserted, and it was only through monumental efforts that agricultural land was brought back into use within two to three years (Chagué-Goff et al., 2014).

Prehistorically, the same evidence lies within archeology and anthropology. There are numerous instances of paleotsunami deposits overlying past coastal occupation sites, many with reworked artifacts incorporated within the sediments (e.g., Futuna, Wallis, and Futuna, discussed in Goff, Lamarche, Pelletier, Chagué-Goff, & Strotz, 2011; and Lavericks Bay, New Zealand, shown in Figure 4). Not surprisingly, in most (but not all) cases, these low-lying coastal sites were permanently abandoned (Goff et al., 2011). In other cases, there has been a notable movement of settlements, both inland and uphill (e.g., Sendai Plain, Japan, discussed in Saino, 2015), regionwide changes in the shell species of midden assemblages (e.g., Washpool, New Zealand, discussed in Goff & McFadgen, 2001), with numerous oral traditions highlighting elements of the event (e.g., Pacific Northwest, as discussed by Heaton & Snavely, 1985; McMillan & Hutchinson, 2002).

PalaeotsunamisClick to view larger

Figure 4. Lavericks Bay, New Zealand. Archaic Māori (pre-1450) occupation site overlain by a paleotsunami deposit. The paleotsunami deposit contains shell and bone fragments from reworked midden material, together with rare oven stones.

Photo: J. Goff.

What is most remarkable about much of the archeological and anthropological evidence is that in many cases of the former, it was well known prior to recent large tsunamis, and for the latter, indigenous peoples at the very least were well aware of the information, but for various reasons, the transfer of this knowledge has generally been ineffective. On a regionwide scale, the Sendai Plain in Japan provides a remarkably good example of “What if?” This was the region affected by the 2011 TOT and its historic precursor, the 869 Jogan tsunami. Archeologists working at the Kutsukata archeology site (2.5 km inland from the paleoshoreline) in 2006 in the northern part of the Sendai Plain uncovered paleotsunami deposits overlying paddy fields dated to around 2,000 years ago, during the middle Yayoi Period (Matsumoto, 2007).

Sadly, there was little interest from other scholars in making this information more widely known (Okamura et al., 2013), and the 2011 TOT has now left its mark on history. Perhaps even more disturbing, however, is not just the fact that a 2,000-year-old paleotsunami deposit was found overlying a past occupation site, but also that archeologists charted the abrupt movement of settlements on the Sendai Plain in the late Yayoi Period inland and uphill (Saino, 2015)—a regionwide response to paleotsunami inundation. Equally fascinating, though, was that 400 years later, during the Kohun Period, prior to the next major tsunami inundation in 869, settlements had once again moved back down from the hills and onto the low-lying lands of the Sendai Plain (Saino, 2015).

Similar patterns of settlement movement are evident throughout the South Pacific, where Polynesian settlement commenced around 3,500 years ago in the Western Pacific and finished with the last large landmass, New Zealand, in around 1300. This was most notably reported by McFadgen (2007), who identified an almost nationwide pattern in the mid-15th century of movement of settlements inland and uphill from low-lying coastal sites. Initial archeological theory related this to a resource crisis resulting from the climatic effects of the Little Ice Age, but subsequent studies revealed a concentrated period of paleoseismicity and paleotsunamis that led inevitably to loss of key resources (e.g., Goff & McFadgen, 2001). This was by no means an isolated incident within the South Pacific; and indeed, once noted as a recognizable societal change within the archeological record, it drew attention to widespread contemporaneous geological evidence for possible paleotsunami deposits throughout the region.

This has ultimately highlighted at least two key South Pacific paleotsunamis that had fundamental influences on the seafaring Polynesian society. First, around 2,800–3,000 years ago, Polynesian settlement of the Pacific halted abruptly adjacent to the Tonga Trench around Tonga/Samoa/Fiji, coincidentally at the same time as a major earthquake and paleotsunami occurred (Goff, McFadgen, Chagué-Goff, and Nichol, 2012b). Later in the 15th century, we see a similar event, but with widespread disruption to the now well established canoe trading routes, loss of resources, the rise of warfare, and the movement inland and uphill to fortified sites (Goff et al., 2012b; Goff & Nunn, 2016).

While there is widespread evidence throughout numerous coastal cultures of the effects of devastating tsunamis, there is still some resistance to combining geological and archeological evidence of past events, with some researchers considering rapid-onset natural hazards such as earthquakes and tsunamis to be too localized in space and time to be important (Anderson, 2009). In their view, such events are recognized as adding a layer of complexity to Pacific human history, but not to have the almost ubiquitous and long-term impacts associated with climate change and anthropogenic modification. What these researchers fail to appreciate is that the archeological evidence points to abrupt and significant change—change that is embedded deeply within the culture, as well as the physical environment.

The extent to which such abrupt events as paleotsunamis have been embedded in the culture can also be recognized through religion and oral traditions. Hasegawa (2016) noted that 65%–80% of the Shinto shrines in the Fukushima and Miyagi Prefectures, respectively, along the exposed eastern shores of Japan were situated inland of the inundation from the 2011 TOT, with some having been historically moved following past events, thus suggesting a deliberate response. Oral traditions abound, and in many instances, they are often highly descriptive of tsunami processes. As with all evidence pertaining to paleotsunamis, though, there is an issue of dating. Some researchers see this as a major stumbling block in the use of oral traditions for corroborating a particular event, but there are invariably two ways of recognizing the value of the data. First, it is not necessary to try and relate the tradition to a specific event, but rather to recognize that the area in question has experienced a culturally significant event. Second, depending upon the indigenous group, oral traditions are often associated with a key character who can be genealogically linked to living relatives, thus providing a reasonable chronology. In all cases, though, when dealing with oral traditions, it is imperative for the researcher to engage directly with the relevant indigenous group and to respect their rights to tell their own histories, not to view them through the interpretive lens of Western society (King & Goff, 2010). Invariably, there is much more knowledge to be gained concerning oral traditions through these engagements than can be inferred in the absence of such actions. Three examples of oral traditions are given here:

  • Heaton and Snavely (1985, p. 1456) reported about a native American tradition of a great sea level disturbance along the coast of Washington which most likely relates to the 1700 Cascadia tsunami:

A long time ago but not at a very remote period, the water of the Pacific flowed through what is now the swamp and prairie between Waatch village and Neeah Bay, making an island of Cape Flattery. The water suddenly receded, leaving Neeah Bay perfectly dry. It was four days reaching its lowest ebb, and then rose again without any waves or breakers, till it had submerged the Cape, and in fact the whole country, excepting the tops of the mountains at Clyoquot. The water on its rise became very warm, and as it came up to the houses, those who had canoes put their effects into them, and floated off with the current, which set very strongly to the north. Some drifted one way, some another; and when the waters assumed their accustomed level, a portion of the tribe found themselves beyond Nootka, where their descendants now reside, and are known by the same name as the Makahs in Classett, or wenaitchechat. Many canoes came down in the trees and were destroyed, and numerous lives were lost. The water was four days regaining its accustomed level.

  • Sakata (2011, pp. 179–180) related an Ainu tradition from Hokkaido, Japan, a region exposed to Kuril Trench tsunamis that have historically inundated its eastern shores:

There is a story called Umi ni ukabu yama o oyoide hippatta otasut-jin no hanashi (“The story of Otasut-lad who pulled an island”). A boy is raised by an old woman. One day his uncle visits him to take him to trade with the Wajin. The old woman gives him a charm, a belt. The uncle leaves him on an island on the way. The boy gets angry, passes the belt around the mountain island, and pulls it; this action results in a big tsunami that overturns his uncle’s boat and kills him. After the boy’s return, the old woman explains that she is a goddess of water, that his uncle is not a real relative of his but the person who killed his parents.

  • Smith (1910) provided a detailed translation of a Māori oral tradition (pūrākau) from New Zealand entitled the “Coming of the Sand.” This comes from an area where previously there had been no paleotsunami research, but a recent publication indicated that this oral tradition forms part of a suite of anthropological, archeological, geomorphological, and geological evidence pointing to a probable locally generated paleotsunami dated to between 1470–1510 (Goff & Chagué-Goff, 2015). Smith (1910) used genealogical estimates to place this event around 1500:

The tradition centres on a place called Potiki-taua, just to the south of Cape Taranaki where Potiki and his group settled. Mango-huruhuru, the old priest, built a large house on low land near the sea while Potiki-roa and his wife put theirs on higher ground further inland. Mango-huruhuru’s house had a rocky beach in front of it that was unsuitable for landing canoes and so he decided to use his powers to bring sand from Hawaiki. After sunset he sat on his roof and recited a karakia (prayer/chant). On conclusion a dark cloud with its burden of sand reached the shore. The women called out “A! The sea rises; the waves and the sand will overwhelm us”. The people fell where they stood and were buried in the sand along with the house and cultivations and all the surrounding country, and with them, the old priest and his youngest daughter (memorialised and turned into a rock which stands there today). Potiki-roa and his wife escaped the disaster because their home was further inland and on higher ground.”

What Is Use of This Information?

A small, but persistent, debate about whether one can tell the difference between a paleotsunami and a paleostorm deposit has not served to derail the usefulness of the information produced. Furthermore, the rapid development of the discipline and occasional quality assurance of earlier work are helping to improve interpretations and ultimately the rigor of the work.

The original questions were how big, how often, and where from? Perhaps more to the point is simply trying to gain an understanding of the backstory of tsunamis at any one point. Yes, we know that places such as Japan’s east coast get big ones and have had quite a few of them over the past few thousand years. Work is still ongoing there to get a true sense of how big they have been in the past and just how far inland they penetrated (Goto, Chagué-Goff, Goff, & Jaffe, 2012). This is extremely important because prior to the 2011 tsunami, evacuation plans, and seawall defenses were based upon a poor data set and as such, walls were overtopped, evacuation centers were inundated, and many people who had evacuated were killed. Interestingly, geochemistry is at the forefront of estimating the maximum inundation distances in Japan, and this in turn can be fed back into both “inverse” models (which work backward from the details of the maximum inland extent of inundation to determine the details of the source) and “forward” models (which work from a given source condition to try and approximate the inundation) to try and work out how big the generating mechanism was, and to say something about where the waves came from as well. In this way, you not only find out about past tsunamis but can start to get a sense of the key paleosources.

Paleotsunami research has proven to be of great value in different parts of the world and for a variety of reasons, but if we return to one of the original papers that saw the consolidation of paleotsunami research as a discipline, we can see the value of this work. Discovery of the physical evidence for the 1700 Cascadia event along the coast of the Pacific Northwest in the mid-1980s (Atwater, 1987) has focused considerable research in the region. Peters, Jaffe, and Gelfenbaum (2007) summarized the knowledge at that time by noting that 60 sites had been recorded along the coast, with recurrence intervals for events averaging 500–600 years. Information gleaned from these findings have been used to help evaluate and mitigate tsunami hazards in the region, with amendments made to coastal planning, tsunami evacuation plans, community awareness, and education. Including paleotsunami information on inundation maps and working with results from inundation models has allowed more accurate assessments to be made of areas subject to inundation. Various models have been used to estimate flow velocities and wave heights—parameters that are necessary to help establish evacuation routes and plan development in tsunami-prone areas.

It has not always been the discovery of paleotsunami evidence that has been the key to increased interest. The 2004 IOT came as a surprise to many researchers, mainly because most of the large, historically documented, oceanwide events had come from the Pacific, with particularly notable ones in 1960 (Chile), 1946 (Alaska), 1896 (Japan), and 1868 (Chile). The 2004 IOT generated a surge of interest in past events around the Indian Ocean, which in hindsight was perhaps a little unfortunate in that it drew attention away from the Pacific Ocean, which proceeded to produce three catastrophic tsunamis in 2009 (Samoa) and 2010 (Chile), as well as the 2011 TOT. However, we now know considerably more about paleotsunamis in the Indian Ocean (e.g., Sri Lanka, as described in Jackson et al., 2014), which once again is serving to inform and guide actions of entities ranging from central governments to individual communities.

In recognition of the importance of contextual data, there is a growing interest in indigenous knowledge, as well as recognition by a growing number of Western scientists of the importance of this information for not only informing interpretations of potential paleotsunami evidence, but also in guiding where further work needs to be done. This is particularly evident in Polynesian cultures, where oral traditions from locations not previously considered as being exposed to a tsunami threat have helped scientists discover a previously understated or even ignored tsunamigenic source. As noted, the “Coming of the Sand” tradition from the west coast of North Island, New Zealand, is on the nonsubduction (trailing edge) side of the country, and therefore considered to be at little risk from tsunamis. Subsequent research, however, has revealed that the western shores of New Zealand, while not experiencing as many tsunamis as the more exposed eastern side, still has a significant risk based on paleotsunami records dating back nearly 1,000 years (Goff & Chagué-Goff, 2015).

Suffice it to say that paleotsunami research is alive and well. It will doubtless continue to engender debate, as it should do, and it will continue to grow in importance because human populations are moving to increasingly exposed coastal locations for which planning and disaster risk reduction initiatives will be needed.

References

Aida, I. (1977). Simulations of large tsunamis occurring in the past off the coast of the Sanriku district. Tokyo-Daigaku-Jishin-Kenkyusho-Ihou (Bulletin of the Earthquake Research Institute, University of Tokyo), 52, 71–101.Find this resource:

Anderson, A. (2009). Epilogue: Changing archaeological perspectives upon historical ecology in the Pacific Islands. Pacific Science, 63, 747–757.Find this resource:

Anonymous. (1905). Seismology in Japan. Nature, 71, 224–225.Find this resource:

Atwater, B. F. (1987). Evidence for great Holocene earthquakes along the outer coast of Washington State. Science, 236, 942–944.Find this resource:

Atwater, B. F., Uri, S., Buckley, M., Halley, R. S., Jaffe, B. E., López-Venegas, et al. (2012). Geomorphic and stratigraphic evidence for an unusual tsunami or storm a few centuries ago at Anegada, British Virgin Islands. Natural Hazards, 63(1), 51–84.Find this resource:

Bailey, E. B., & Weir, J. (1933). XIV—Submarine faulting in Kimmeridgian times: East Sutherland. Transactions of the Royal Society of Edinburgh, 57, 429–467.Find this resource:

Bondevik, S., & Svendsen, J. I. (1993). Palaeotsunamis in the Norwegian Sea and the North Sea. In S. Tinti (Ed.), Genesis and impact of tsunamis on the European coasts—GITEC. Mid-Term Scientific Report, Contract EV5F-CT92-0175, University of Bologna, Italy, 7.2–7.27.Find this resource:

Bourgeois, J., Hansen, T. A., Wiberg, P. L., & Kauffman, E. G. (1988). A tsunami deposit at the Cretaceous–Tertiary boundary in Texas. Science, 241, 567–570.Find this resource:

Bridge, J. S. (2008). Discussion of articles in “Sedimentary features of tsunami deposits.” Sedimentary Geology, 211(3), 94.Find this resource:

Bruins, H., MacGillivray, J. A., Synolakis, C., Benjamini, C., Hanan, J. K., Kisch, J., et al. (2008). Geoarchaeological tsunami deposits at Palaikastro (Crete) and the Late Minoan IA eruption of Santorini. Journal of Archaeological Science, 35, 191–212.Find this resource:

Butler, R., Burney, D., & Walsh, D. (2014). Paleotsunami evidence on Kaua’i and numerical modeling of a great Aleutian tsunami. Geophysical Research Letters, 41(19), 6795–6802.Find this resource:

Chagué-Goff, C. (2010). Chemical signatures of palaeotsunamis: A forgotten proxy? Marine Geology, 271, 67–71.Find this resource:

Chagué-Goff, C., Chan, J., Goff, J., & Gadd, P. (2016). Late Holocene record of environmental changes, cyclones, and tsunamis in a coastal lake, Mangaia, Cook Islands. Island Arc, 25, 333–349.Find this resource:

Chagué-Goff, C., Dawson, S., Goff, J. R., Zachariasen, J., Berryman, K. R., Garnett, D. L., et al. (2002). A tsunami (c.6300 years BP) and other Holocene environmental changes, northern Hawke’s Bay, New Zealand. Sedimentary Geology, 150, 89–102.Find this resource:

Chagué-Goff, C., & Goff, J. R. (1999). Geochemical and sedimentological signature of catastrophic saltwater inundations (tsunami), New Zealand. Quaternary Australasia, 17(1), 38–48.Find this resource:

Chagué-Goff, C., Schneider, J.-L., Goff, J. R., Dominey-Howes, D., & Strotz, L. (2011). Expanding the proxy toolkit to help identify past events: Lessons from the 2004 Indian Ocean tsunami and the 2009 South Pacific tsunami. Earth-Science Reviews, 107, 107–122.Find this resource:

Chagué-Goff, C., Wong, H. K. Y., Sugawara, D., Goff, J., Nishimura, Y., Beer, J., et al. (2014). Impact of tsunami inundation on soil salinisation—up to one year after the 2011 Tohoku-oki tsunami. Chapter 10. In Y. A. Kontar, V. Santiago-Fandiño, & T. Takahashi (Eds.), Tsunami events and lessons learned: Environmental and societal significance (pp. 193–214). Dordrecht, the Netherlands: Springer Science + Business Media.Find this resource:

Chamberlin, T. C. (1890). The method of multiple working hypotheses. Science, 15, 92–96.Find this resource:

Coleman, P. J. (1968). Tsunamis as geological agents. Journal of the Geological Society of Australia, 15(2), 267–273.Find this resource:

Darwin, C. (1846). Geological observations on South America. London: Smith, Elder.Find this resource:

Dawson, A. G. (1994). Geomorphological effects of tsunami run-up and backwash. Geomorphology, 10(1), 83–94.Find this resource:

Dawson, A. G., Long, D., & Smith, D. E. (1988). The Storegga slides: Evidence from eastern Scotland for a possible tsunami. Marine Geology, 82, 271–276.Find this resource:

Engel, M., & Brückner, H. (2011). The identification of palaeo-tsunami deposits–a major challenge in coastal sedimentary research. Coastline Reports, 17, 65–80.Find this resource:

Goff, J. (2011). Evidence of a previously unrecorded local tsunami, 13 April 2010, Cook Islands: Implications for Pacific Island countries. Natural Hazards and Earth System Sciences, 11, 1371–1379.Find this resource:

Goff, J., & Chagué-Goff, C. (2015). Three large tsunamis on the non-subduction, western side of New Zealand over the past 700 years. Marine Geology, 363, 243–260.Find this resource:

Goff, J., Chagué-Goff, C., & Nichol, S. (2001). Palaeotsunami deposits: A New Zealand perspective. Sedimentary Geology, 143, 1–6.Find this resource:

Goff, J., Chagué-Goff, C., Nichol, S. L., Jaffe, B., & Dominey-Howes, D. (2012a). Progress in palaeotsunami research. Sedimentary Geology, 243–244, 70–88.Find this resource:

Goff, J., Goto, K., Ebina, Y., & Terry, J. (2016). Defining tsunamis: Yoda strikes back? Earth-Science Reviews, 159, 271–274.Find this resource:

Goff, J., Lamarche, G., Pelletier, B., Chagué-Goff, C., & Strotz L. (2011). Palaeotsunami precursors to the 2009 South Pacific tsunami in the Wallis and Futuna archipelago. Earth-Science Reviews, 107, 91–106.Find this resource:

Goff, J., & McFadgen, B. G. (2001). Catastrophic seismic-related events and their impact on prehistoric human occupation in coastal New Zealand. Antiquity, 74, 155–162.Find this resource:

Goff, J., & McFadgen, B. G. (2002). Seismic driving of nationwide changes in geomorphology and prehistoric settlement—a 15th-century New Zealand example. Quaternary Science Reviews, 21, 2229–2236.Find this resource:

Goff, J., McFadgen, B. G., Chagué-Goff, C., & Nichol, S. L. (2012b). Palaeotsunamis and their influence on Polynesian settlement. The Holocene, 22, 1061–1063.Find this resource:

Goff, J., Nichol, S. L., Chagué-Goff, C., Horrocks, M., McFadgen, B., & Cisternas, M. (2010a). Predecessor to New Zealand’s largest historic trans–South Pacific tsunami of 1868AD. Marine Geology, 275, 155–165.Find this resource:

Goff, J., Nichol, S. L., & Kennedy, D. (2010b). Development of a palaeotsunami database for New Zealand. Natural Hazards, 54, 193–208.Find this resource:

Goff, J., & Nunn, P. (2016). Rapid societal change as a proxy for regional environmental forcing: Evidence and explanations for Pacific Island societies in the 14th–15th centuries Island Arc, 25, 305–315.Find this resource:

Goff, J., & Sugawara, D. (2014). Seismic driving of sand beach ridge formation in northern Honshu, Japan? Marine Geology, 358, 138–149.Find this resource:

Goto, K., Chagué-Goff, C., Fujino, S., Goff, J., Jaffe, B., Nishimura, Y., et al. (2011). New insights into tsunami risk from the 2011 Tohoku-oki event. Marine Geology, 290, 46–50.Find this resource:

Goto, K., Chagué-Goff, C., Goff, J., & Jaffe, B. (2012). The future of tsunami research following the 2011 Tohoku-oki event. Sedimentary Geology, 282, 1–13.Find this resource:

Goto, K., Miyagi, K., Kawamata, H., & Imamura, F. (2010). Discrimination of boulders deposited by tsunamis and storm waves at Ishigaki Island, Japan. Marine Geology, 269, 34–45.Find this resource:

Hall, J. (1815). IV. On the revolutions of the Earth’s surface. Transactions of the Royal Society of Edinburgh, 7, 139–167.Find this resource:

Hasegawa, S. (2016). The relationship between tsunami susceptibility and Pā locations. Unpublished report for University of Canterbury, School of Anthropology, Christchurch, New Zealand, and Brown University, Providence, RI, Department of Anthropology.Find this resource:

Heaton, T. H., & Snavely, P. D., Jr. (1985). Possible tsunami along the northwestern coast of the United States inferred from Indian traditions. Bulletin of the Seismological Society of America, 75, 1455–1460.Find this resource:

Intergovernmental Oceanographic Commission (IOC). (2013). Tsunami glossary (Revised ed.). Paris, France: UNESCO. IOC Technical Series, 85 (English). (IOC/2008/TS/85rev).Find this resource:

Jackson, K. L., Eberli, G. P., Amelung, F., McFadden, M. A., Moore, A. L., Rankey, E. C., et al. (2014). Holocene Indian ocean tsunami history in Sri Lanka. Geology, 42, 859–862.Find this resource:

Jaffe, B. E., Morton, R. A., Kortekaas, S., Dawson, A. G., Smith, D. E., Gelfenbaum, G., et al. (2008). Reply to discussion of articles in “Sedimentary features of tsunami deposits.” Sedimentary Geology, 211, 95–97.Find this resource:

Jiji Press. (2012). Saitama to teach about Miyagi’s tsunami “angel.” Japan Times, January 30, 2.Find this resource:

Judd, J. W. (1873). The secondary rocks of Scotland. Quarterly Journal of the Geological Society, 29, 97–195.Find this resource:

Kastens, K. A., & Cita, M. B. (1981). Tsunami-induced sediment transport in the abyssal Mediterranean Sea. Geological Society of America Bulletin, 92, 845–857.Find this resource:

King, D., & Goff, J. (2010). Benefitting from differences in knowledge, practice and belief: Māori oral traditions and natural hazards science. Natural Hazards and Earth System Sciences, 10, 1927–1940.Find this resource:

Kortekaas, S., & Dawson, A. G. (2007). Distinguishing tsunami and storm deposits: An example from Martinhal, SW Portugal. Sedimentary Geology, 200, 208–221.Find this resource:

Macdonald G. A., Shepard, F. P., & Cox, D. C. (1947). The tsunami of April 1, 1946, in the Hawaiian Islands. Pacific Science, 1, 21–37.Find this resource:

Matsumoto, H. (2007). 仙台平野北部に認められる‎ 2,000 yr BP 頃の砂の薄層と周辺の地形‎ (A thin sand layer dating to 2,000 years BP found in the northern part of the Sendai Plain). Quarterly Journal of Geography, 59, 169.Find this resource:

McFadgen, B. G. (2007). Hostile shores: Catastrophic events in prehistoric New Zealand and their impact on Māori coastal communities. Auckland, New Zealand: Auckland University Press.Find this resource:

McFadgen, B. G., & Goff, J. (2005). An earth systems approach to understanding the tectonic and cultural landscapes of linked marine embayments: Avon-Heathcote Estuary (Ihutai) and Lake Ellesmere (Waihora), New Zealand. Journal of Quaternary Science, 20, 227–237.Find this resource:

McMillan, A. D., & Hutchinson, I. (2002). When the mountain dwarfs danced: Aboriginal traditions of Paleoseismic events along the Cascadia subduction zone of western North America. Ethnohistory, 49, 41–68.Find this resource:

Moore, J. G., & Moore, G. W. (1984). Deposit from a giant wave on the island of Lanai, Hawaii. Science, 226(4680), 1312–1315.Find this resource:

Morton, R. A., Gelfenbaum, G., & Jaffe, B. E. (2007). Physical criteria for distinguishing sandy tsunami and storm deposits using modern examples. Sedimentary Geology, 200, 184–207.Find this resource:

Moseley, M. E., Wagner, D., & Richardson, J. B., III. (1991). Chapter 10. Space shuttle imagery of recent catastrophic change along the arid Andean coast. In L. L. Johnson (Ed.), Paleoshorelines and prehistory: an investigation of method. Boca Raton, FL: CRC Press.Find this resource:

Okamura, K., Fujisawa, A., Kondo, Y., Fujimoto, Y., Uozu, T., Ogawa, Y., et al. (2013). The Great East Japan Earthquake and cultural heritage: Towards an archaeology of disaster. Antiquity, 87, 258–269.Find this resource:

Peters, R., Jaffe, B., & Gelfenbaum, G. (2007). Distribution and sedimentary characteristics of tsunami deposits along the Cascadia margin of western North America. Sedimentary Geology, 200, 372–386.Find this resource:

Pinegina, T. K., & Bourgeois, J. (2001). Historical and paleo-tsunami deposits on Kamchatka, Russia: Long-term chronologies and long-distance correlations. Natural Hazards and Earth System Sciences, 1, 177–185.Find this resource:

Platania, G. (1909). Il maremoto dello Stretto di Messina del 28 Dicembre 1908 (The Straits of Messina tsunami of 28 December 1908). Bollettino della Società Sismologica d’Italia, 13, 369–458.Find this resource:

Saino, H. (2015). Tsunami disasters of Yayoi period and Heian period in the Sendai Plain. Proceedings of the Symposium on Traces and Experiences of Past Tsunami Disasters in the Pacific Rim, and the Succession of Knowledge. UN World Conference on Disaster Risk Reduction, Sendai, Japan, March 14.Find this resource:

Sakata, M. (2011). Possibilities of reality, variety of versions: The historical consciousness of Ainu folktales. Oral Tradition, 26, 175–190.Find this resource:

Scidmore, E. R. (1896). The recent earthquake wave on the coast of Japan. National Geographic Magazine, 7(September), 285–289.Find this resource:

Scopus. (2016). Elsevier.

Shanmugam, G. (2012). Process-sedimentological challenges in distinguishing paleo-tsunami deposits. Natural Hazards, 63, 5–30.Find this resource:

Smith, S. P. (1910). History and traditions of the Māoris of the west coast North Island of New Zealand prior to 1840. Memoirs of the Polynesian Society, I, 175–185.Find this resource:

Tappin, D. R., Watts, P., & Grilli, S. T. (2008). The Papua New Guinea tsunami of 17 July 1998: Anatomy of a catastrophic event. Natural Hazards and Earth System Sciences, 8, 243–266.Find this resource:

Terry, J. P., & Goff, J. (2013). One hundred and thirty years since Darwin: Reshaping the theory of atoll formation. The Holocene, 23, 613–617.Find this resource:

Wartman, J., Dunham, L., Tiwari, B., & Pradel, D. (2013). Landslides in eastern Honshu induced by the 2011 Tōhoku earthquake. Bulletin of Seismological Society of America, 103, 1503–1521.Find this resource:

Wegman, C. E. (1939). Zwei Bilder für das Arbeitszimmer eines Geologen [Two pictures for the study of a geologist] Geologische Rundschau, 30, 390–391.Find this resource:

Weiss, R. (2008). Sediment grains moved by passing tsunami waves: Tsunami deposits in deep water. Marine Geology, 250, 251–257.Find this resource:

Wells, A., & Goff, J. (2006). Coastal dune ridge systems as chronological markers of paleoseismic activity—a 650-year record from southwest New Zealand. The Holocene, 16, 543–550.Find this resource:

Wells, A., & Goff, J. (2007). Coastal dunes in Westland, New Zealand, provide a record of paleoseismic activity on the Alpine fault. Geology, 35, 731–734.Find this resource:

Wood, B. L. (1967). Geology of the Cook Islands. New Zealand Journal of Geology and Geophysics, 10, 1429–1445.Find this resource:

Yawsangratt, S., Szczuciński, W., Chaimanee, N., Chatprasert, S., Majewski, W., & Lorenc, S. (2012). Evidence of probable paleotsunami deposits on Kho Khao Island, Phang Nga Province, Thailand. Natural Hazards, 63, 151–163.Find this resource: