Global Historical Megatsunamis Catalog (GHMCat)

Abstract

The Global Historical Megatsunamis Catalog (GHMCat) is presented for the first time, including events with the largest waves recorded in historical times. An objective criterion is established to identify megatsunamis based on the maximum wave height (runup) of all recorded events. A threshold value of 35 m for maximum wave height is proposed based on the analysis of the statistical distribution of the maximum wave heights documented. The catalog was compiled through a systematic review and verification of tsunami events from the two existing Global Historical Tsunami Databases (GHTDs). A list of 40 megatsunamis from 1674 to the present is presented, including descriptions of their maximum wave heights, causes and sources according to the available and verified information, along with the main bibliographical references that support the data gathered in the catalog. The majority of megatsunamis have originated from large landslides, predominantly subaerial, with fewer caused by submarine landslides or associated with volcanic explosions. The geographical distribution of source locations shows that megatsunamis most frequently occur in bays and fjords in glaciated areas and in inland bodies of water, such as lakes and rivers. Notably, certain regions of Alaska and Norway experienced an unusual frequency of megatsunamis, particularly in the early 20th century. The information provided by the GHMCat allows for a comprehensive historical overview of megatsunamis, establishing relationships between their causes, wave heights, and geographic distribution over the past 350 years. This may contribute to advancing the study of the causes and origins of megatsunamis and aid in their prevention in high-risk regions.

1. Introduction

The main objective of this study is to present a comprehensive global catalog of historical megatsunamis, compiling and describing all events documented to date through written records from eyewitnesses or direct observations and measurements.

Megatsunamis, or giant waves, are extreme events that impact cliff shores in any geographical or climatic region, contingent upon the presence of a body of water and the occurrence of a geological triggering process with sufficient energy to cause a sudden and large vertical displacement of water, generating waves of great height.

Despite lacking a strict or universally agreed definition, the term megatsunami has been widely used in scientific publications in recent decades, primarily associated with prehistoric oceanic tsunamis linked to large bolide impacts and giant volcanic-island-flank collapses [1,2,3,4,5]. Additionally, it has been arbitrarily applied to extensive and destructive trans-oceanic tsunamis such as those in Indonesia in 2004 and Japan in 2011, as well as earlier events in the Aleutians in 1946, Chile in 1960, and Alaska in 1964 [6].

Megatsunamis are local events originating from nearby sources, with their effects confined to small areas, as the wave height rapidly decreases with distance from the source. Unlike conventional tsunamis triggered by offshore earthquakes, megatsunamis can reach very large heights upon impacting coastlines, with observed or measured maximum values of several hundred meters on the shores of enclosed marine or inland bodies of water.

The number of known megatsunamis documented in historical records or inferred from geological evidence is very limited. The oldest recorded event is the one triggered by the massive explosive eruption on the island of Thera (Santorini) in Greece around 1600 B.C. Numerous megatsunamis must have occurred in historical times beyond the limited records available, but they were neither observed nor reported. Over Earth’s history, such processes must have been relatively frequent, but due to the short historical record and the disappearance of their effects from the Earth’s shores, we have access to only a very limited number of records.

Most historical megatsunamis have been generated by large subaerial landslides, with some cases involving submarine landslides and a few cases related to violent volcanic eruptions. Subaerial landslide-generated tsunamis cause the largest waves due to the kinetic energy of huge rock masses falling from considerable heights and violently entering confined bodies of water. Landslides are sometimes triggered by large earthquakes and often occur in bays or lakes. Megatsunamis caused by subaerial landslides have been documented since the 18th century in Norway and Japan.

To date, only a few studies have compiled lists of “large tsunamis” for specific areas or time intervals [6]. Partial lists of landslide-generated “giant waves” or “large waves” have also been published [7,8,9,10,11].

To compile a catalog of historical megatsunamis, and in the absence of a universally accepted scientific definition of a megatsunami, it is essential to define the term based on available data from documented events.

A methodology based on objective data and criteria is developed, utilizing measurable and universally applicable tsunami parameters, such as wave height, that can be consistently applied across all documented events, regardless of their origin. In this context, a megatsunami is defined as a giant wave with a height of 35 m or more, as proposed in this study and detailed in Section 5, where the selection of this threshold is discussed.

Information from the two existing Global Historical Tsunami Databases (GHTDs) [12,13] was used as primary sources, supplemented by numerous publications of various scopes and contents, ranging from regional and local catalogs to publications or reports describing or investigating individual events. Each documented historical event was meticulously reviewed and investigated, with references to original sources whenever possible. As a result, some information in the databases and catalogs regarding significant tsunamis required corrections due to errors in maximum wave height data and their causes.

Among the problems associated with historical information are those derived from the inaccuracy and incompleteness of the data—especially concerning old or geographically remote events—and errors in the interpretation of old descriptions and conversion of measurement units for wave height. Although only a few records from the 19th century have been properly described, the number of recorded events significantly increased during the 20th century; in just the first 20 years of the 21st century, as many cases have been recorded as in the entire 20th century. This increase in recorded large tsunamis correlates with the advancement of scientific knowledge and industrial development from the second half of the 19th century, and especially in the 20th century.

From 1674 to 2024, a total of 40 megatsunamis (i.e., those with wave heights ≥35 m) have been included in the Global Historical Megatsunami Catalog (GHMCat), following a thorough investigation of the available documentation; more than 300 publications from 1888 to 2024 have been consulted. For some of these events, extreme heights in excess of 100 m have been reported.

In addition to the cataloged events, a previously undocumented megatsunami has been identified through the review of old documents by Alaskan explorers from the late 18th century: a megatsunami prior to 1786 in Lituya Bay [14]. However, it has not been included in the GHMCat as its year of occurrence is unknown, and we can only establish its maximum-limiting age.

The GHMCat compiles all historical megatsunamis recorded so far and provides, for the first time, a complete list of events along with information on their causes and most significant effects, damages, or casualties, when available. Data on the maximum recorded wave heights, their causes and effects, as well as their frequency and geographic distribution, are crucial for understanding the incidence of this phenomenon, its generation, and potential risks and harmful effects. This information can contribute to prevention and mitigation efforts.

2. Background, Definitions, and Data Availability

2.1. Previous Definitions

Although the term megatsunami lacks a precise definition, it is generally used to describe events based on maximum wave height, which serves as a practical parameter for distinguishing it from “conventional” or seismic tsunami waves. Megatsunamis are typically defined as waves reaching or exceeding a certain height; however, there is no specific wave height or threshold above which the term can be applied, nor is there consensus on whether wave height should be measured at the shore or at the source of the tsunami.

The term megatsunami emerged in scientific literature around 1990 [15] linked to giant waves (“mega-tsunami”) caused by a massive asteroid impact in the Gulf of Mexico millions of years ago, at the Cretaceous–Paleogene boundary. Earlier studies by Bourgeois et al. [16] referred to wave heights of 50 to 100 m for this event, using the term “very large tsunami”. Another early reference by Paskoff [17] suggests a possible “mega-tsunami” to explain large boulders embedded in shelly beach deposits on a 40 m elevated marine terrace on the Chilean coast, likely caused by an exceptionally large earthquake during the Plio–Quaternary. The term “giant wave” was previously used to explain the origin of elevated marine coarse deposits on Lanai, Hawaii [18,19], where wave heights reaching between 190 and 375 m are inferred to have occurred, likely triggered by a massive submarine landslide. The same term “giant wave” was also used [7] to describe waves generated by a large rockslide in Lituya Bay, Alaska, in 1958.

Few publications provide specific threshold values for classifying waves as megatsunamis. Some authors have subjectively suggested a minimum wave height of 40 or 50 m a.s.l., or even 100 m, while others propose heights from several tens to hundreds of meters. Alexander and Neall [20] indicate that megatsunamis are above 40 m, and up to hundreds of meters, contrasting with “normal” or earthquake-generated tsunamis, which typically reach up to about 10 m. According to Krehl [21], “mega-tsunamis” are defined in the literature as waves over 100 m, even over 300 m, but this lacks specific bibliographical references.

Goff et al. [22] compile and discuss these and other references, and propose a definition of a “mega-tsunami” based on the initial wave height/amplitude at source exceeding 100 m/50 m, respectively. This definition would only include extreme events from large bolide impacts, extremely violent volcanic activity, or giant landslides, and “perhaps an extremely large and as yet historically undocumented earthquake generated event” [22].

Naranjo et al. [23] define a megatsunami as a tsunami with an initial wave amplitude or height of several tens or hundreds of meters, significantly surpassing that of a “normal” tsunami; most megatsunamis result from major impacts, such as large-scale landslides, devastating volcanic eruptions, or meteorite strikes, whereas “normal” tsunamis typically result from tectonic activity causing abrupt seabed displacements.

As reflected in the previous definitions, wave height is the primary parameter for defining a megatsunami, but the cause of the event is also considered to play a determining role.

2.2. Wave Height and Runup Measurements

Some of the previous definitions focus on wave height at the source, while others consider wave heights on the shore. Data on tsunami wave heights at the source or near the source can only be obtained instrumentally through deep oceanic sensors, only available for the last 50 years [24] and exclusively for tsunamis originating in the open ocean. The earliest reliable data were obtained in the early 1980s. Most of the current monitoring stations have been deployed in recent years, particularly following the devastating Indonesian tsunami of 2004. In the same way, tide gauge measurements of wave height along coastlines are unavailable for non-seismic tsunamis, with such data not existing for events prior to the mid-20th century, except in rare instances [24,25,26]. The first tide gauges began to function in the Pacific Ocean in the mid-19th century, with the network becoming sufficiently dense only by the mid-20th century [27].

The only typically available data for most events recorded in the GHTDs—particularly for significant wave heights exceeding 10 m—are the maximum height reached by the waves above a reference water level (mean sea, lake, or others) at any location along the affected coast, from any type of measurement. The term commonly used to refer to this parameter is maximum water height Hmax, runup height, or simply runup [28]. The maximum horizontal distance on land (inundation or horizontal runup) is not considered in the definition of megatsunami. Both values depend largely on the elevation and topography of the coast.

In most GHTD entries, the Hmax or runup value for significant tsunamis comes from witness accounts or, for the last three decades, post-tsunami survey measurements. Much less frequently, these values come from tide gauge measurements; only in about ten cases does the measurement correspond to depth tide gauges, all from 2006 onwards. The American GHTD [12] includes the following data:

• Tide gauge measurements: 63 cases (40%) among the events with Hmax ≥10 m, all from 1945 onwards; 11 cases (24%) with Hmax ≥30 m, from 1946 onwards, except for the 1883 Krakatoa tsunami.
• Depth gauge measurements: 11 cases (7%) among the events with Hmax ≥10 m, and only 1 case for Hmax ≥30 m, the 2011 Japan tsunami.

These data reflect the scarcity of instrumental measurements and the absence of such measurements for older tsunamis (prior to the mid-20th century).

According to UNESCO [29], and other sources [30,31], runup is the difference between the maximum ground elevation wetted on a sloping shoreline and the sea level, i.e., the maximum elevation of a point inundated by the waves; in the simplest case, runup can be measured at the maximum tsunami penetration inland where the topography is flat. However, with large tsunamis and megatsunamis, waves can exceed the height corresponding to this maximum inundation limit. For example, when waves hit rocky cliffs, they splash and reach elevations higher than those at the maximum inundation distance. Local topography plays a crucial role in determining tsunami height and the enhancement of runup. As observed during the 1960 Chilean tsunami, runup measured as the maximum height at the landward limit significantly underestimated the heights reached closer to the shore [32].

Currently, accurate measurements of the largest tsunami waves are obtained through post-tsunami field surveys, examining the marks left by the waves on the coastline. Maximum height or runup is inferred from indicators such as the vertical limit of vegetation removed by the wave, marks on wooded slopes, damage to trees or debris deposited on hillsides or in trees.

It is important to note that the maximum height reached by the water is always specific to a single point, representing a discrete measurement that may not be representative of average or even maximum wave heights along the coast. This data point often corresponds to the peak of surges or splash events caused by violent water movements or seiches within enclosed bodies of water. For instance, the 1958 Lituya Bay megatsunami recorded a maximum runup of 525 m where the water surge impacted a steep slope and splashed to very high elevations, establishing this as the runup value for that event. However, researchers sometimes differentiate between this local runup measurement and other metrics such as the maximum wave height on the bay’s slopes (~150 m) or the maximum wave amplitude (~50 m) [30].

Measurements taken on cliffs or steep coastlines can yield local runup values that significantly differ from the maximum wave heights observed in surrounding areas. For instance, during the 2004 Indonesian tsunami, a maximum runup of 50–51 m was measured west of Banda Aceh in northern Sumatra [33,34]. This measurement corresponds to the height of vegetation washed away by waves on cliff walls along a small isthmus. While this figure represents the highest ever measured for a tsunami attributed to an earthquake (specifically for a non-landslide tsunami), its highly localized nature prevents it from being considered representative of the tsunami; in fact, some researchers [33,35] clearly distinguish between the largest wave heights, approximately 35 m in the area most affected, and the maximum point runup height, influenced by specific coastal topography.

2.3. Megatsunamis and Large Tsunamis

“Megatsunamis” and “large tsunamis” are terms used with some frequency to refer to large-scale seismic waves in the ocean. However, “megatsunami” typically implies waves of exceptionally large dimensions, usually caused by catastrophic events such as subaerial or submarine landslides or volcanic eruptions. Terms such as “large tsunami”, “oceanic tsunami”, or “regional tsunami” are more general, and can refer to any significant or major seismic tsunami in terms of geographical extent and damages occurring in the ocean, without necessarily connoting exceptionally large waves. On the other hand, “large tsunami” is often used to describe tsunamis with significant wave heights, effectively making it synonymous with “megatsunami” in this context.

Tsunamis generated by earthquakes have a wavelength far greater than that of non-seismic origin; the latter, unlike those of tectonic origin, can reach an initial wave amplitude of tens or hundreds of meters, and their impacts and effects are much more localized compared to “normal” seismic tsunamis, which usually dissipate their energy over a broader area as they travel across oceans reaching coastlines even thousands of kilometers away from their earthquake source.

According to Goff et al. [22], the term megatsunami has been used arbitrarily in a variety of settings, not only to describe large waves at the source of the event or on land, but also to refer to geographically extensive tsunamis, such as the 2004 and 2011 ocean scale tsunamis, whether or not they involve giant waves. However, these remarkably large events do not fit into the classification of megatsunamis, but rather into a category of unexpected exceptional events [22]. In this context, Gusiakov [6] mentions “catastrophic trans-oceanic mega-tsunamis affecting an entire oceanic basin…”, listing seven events within the last 77 years, resulting from large subduction earthquakes, with maximum runups or wave heights between 23 m and 67 m; one of these notable events was caused by the largest earthquake in recorded history, the 1960 Chile earthquake, with a maximum wave height reported of between 15 and over 20 m, depending on the bibliographical source [36,37].

Both the geographical extent and the damaging impact caused by the waves in coastal areas are irrelevant to the definition of a megatsunami, based solely on the maximum height of the waves on land. Megatsunamis are always local events, originating from a nearby source, and their effects are confined to a small area. Maximum runups are focused along a limited length of coast—typically not exceeding some tens of kilometers at most, in contrast to the hundreds of kilometers that a tectonically-induced tsunami can reach [10]—and decreases rapidly with distance. Near-field runup can be many times higher than that produced by seismic tsunamis, and can reach hundreds of meters in exceptional circumstances.

Literature references including lists of megatsunamis or large waves are quite scarce. Gusiakov [6] provides a compilation of “large tsunamis” collected from the two existing global historical databases (GHTDs) for the period 1900–2019, based on annual maximum runups observed or measured (Figure 1); among these events, 14 exceed or equal 50 m in height, with most of them being landslide-generated. Additionally, short lists of historical tsunamis with runup >50 m have been compiled from the GHTDs for the past century (14 events) and for the period 1737–2004 (13 events), with landslides being the primary cause [38,39]. Miller [7] presents a list of data on localized “giant waves” generated by subaerial landslides, and partial lists have also been published [8,9,10,11].

3. Causes and Origin of Megatsunamis

3.1. Causes of Megatsunamis (Why)

Megatsunamis result from massive large-scale landslides, colossal explosive volcanic eruptions, and large asteroids impacts in the sea or confined bodies of water, whereas seismic tsunamis are caused by tectonic activity violently displacing the sea bed and the water column above. Unlike large seismic tsunamis, megatsunamis are local processes, as wave height rapidly decreases with distance from their origin: “Large breaking waves … are rare and almost unheard of outside of the generating area” [30].

The low probability of large-scale geological processes occurring also implies a low probability of megatsunamis worldwide. Prehistoric volcanic-island-flank collapses, the largest ever to have occurred on Earth (with volumes reaching several hundred km³), have generated megatsunamis up to several hundred meters high, as evidenced by geological deposits left by the waves along coastlines in Hawaii and the Canary Islands [1,2,3]. No volcanic island flank megalandslides or asteroid impacts have occurred in historical times.

The exact cause of megatsunamis with exceptional runup heights associated with large earthquakes has been, and in some cases remains, controversial. While some researchers proposed they are generated by large submarine landslides triggered by the earthquake, others argue that it is the rupture mechanism of the seismic fault itself the cause of the giant waves. These “megathrust earthquakes” occur at convergent plate boundaries, and are referred to as “slow tsunami earthquakes” or “tsunami earthquakes” due to their potential to generate tsunami waves much higher than expected from their surface wave magnitude. However, some events, even though they have been classified as tsunami earthquakes, require the involvement of a major underwater landslide to explain the height and distribution of measured runups [40].

Evidence of landslides can be gathered through marine geological and geophysical surveys, as has been the case for several historical megatsunamis initially attributed to earthquakes [30,41,42,43,44].

The first evidence of a large earthquake-triggered underwater landslide was during the 1929 Grand Banks earthquake (M 7.2), North Atlantic, after numerous undersea telegraph cables were broken as a result of major mass movements downslope. In the more recent 1998 Papua New Guinea earthquake (M 7.1), a submarine landslide was inferred from a variety of evidence [45]. Both of these underwater mass movements generated waves of up to 13–15 m in height on nearby coastlines, disproportionate to the magnitude of the earthquakes, i.e., the earthquakes were too small to generate the recorded tsunami waves. Prior to the 1998 event, it was questionable as to whether a submarine landslide could cause such large tsunamis.

Local megatsunamis generated by earthquake-triggered submarine landslides far exceed the height of the major—much more extensive—tectonic tsunami, and reach the coast within minutes after the earthquake; both superimpose on the shores near the origin of the landslide. As demonstrated by research on the near-field runup heights and their distribution in some of the most significant tsunamis of the 1992–2002 period, the concentration of maximum runups along the affected coast clearly reflects the occurrence of a different generating mechanism besides seismic dislocation [31,40]. This is especially illustrative in the case of the 1998 Papua New Guinea tsunami mentioned above.

3.2. Origin of Megatsunamis (Where)

The term megatsunami is broadly applied to large or giant waves generated by geological phenomena, primarily subaerial landslides and rock avalanches, when impacting oceans, semi-confined and narrow fjords or bays [46,47,48] or still bodies of water such as rivers, lakes [9,47,49,50,51,52], or reservoirs [53,54,55]. According to Gusiakov [6], over 15% of maximum annual runups for the last 120 years’ globally recorded tsunamis occurred in coastal and inland water basins: narrow bays, fjords, lakes, and rivers, with all higher runups (>60 m) falling within this category.

However, in certain scientific literature, the term megatsunami is reserved for exceptionally large waves occurring in oceanic environments, resulting from submarine or subaerial landslides or volcanic explosions. For similar events of giant waves caused by subaerial landslides occurring in confined water bodies, e.g., lakes or reservoirs, alternative terms like impulse wave [56,57,58,59] or seiche [12] are employed, both specific types of waves with characteristics that megatsunami waves do not always meet. Nevertheless, in technical terms, if a giant wave is produced in a confined body of water, lake, or reservoir, due to a geological catastrophic event, it can be defined as and called a megatsunami.

Impulse waves and seiches generated in lakes or other confined bodies of water are unequivocally considered tsunami waves—and explicitly included within tsunamis, or megatsunamis—supported by formal definitions and the scientific literature:

• A tsunami, as defined by NOAA [28], is “a water wave or a series of waves generated by an impulsive vertical displacement of the surface of the ocean or other body of water”, specifying, further, that “locally destructive tsunamis may be generated by landslides into bays or lakes”.
• Subaerial landslides falling into water bodies such as lakes, rivers, reservoirs, and fjords can generate impulse waves or seiches, which are a type of tsunami wave [28,54,59,60].
• The terms landslide–tsunami and impulse wave are interchangeable when describing the waves generated by a mass sliding into a confined water body [58,60,61].
• Impulse waves and seiches are specific types of waves encompassed under the term of megatsunami, since they can be produced by megatsunamis.

4. Methodology

Up to the present, there is no widely accepted scientific definition of the term megatsunami, which is commonly used to refer to both tsunamis with extremely large waves and very extensive and destructive tsunamis. Consequently, the primary aim of this study is to define the term megatsunami based on objective and measurable criteria, specifically focusing on a physical characteristic such as maximum wave height—a parameter that is observable and applicable across all documented cases. This definition will form the basis for developing a global catalog of megatsunamis, which is the ultimate goal of this study. The methodology employed to achieve these two objectives is summarized below.

• Phase 1: Megatsunami Definition Based on Wave Heigh Data

• Analysis of data on maximum wave heights of all historical tsunamis documented in the two existing Global Historical Tsunami Databases (GHTDs).
• Establishment of a wave height threshold for megatsunamis based on the statistical distribution of all recorded maximum wave heights for the historical period.
• Definition of “megatsunami” based on the established wave height threshold.

• Phase 2: Megatsunamis Catalog

• Data sources and literature review

  -

  Review of GHTDs as primary data sources.

  -

  Literature review: comprehensive examination of existing catalogs, reports, studies, scientific papers, and other relevant publications.

  -

  Addressing uncertainties: identification of uncertainties and inconsistencies in the data and interpretations of events.
• Data collection and verification

  -

  Identification of definite megatsunamis meeting the proposed definition in the GHTDs.

  -

  Verification of data accuracy and consistency through credible historical records and/or geological evidence from documentary sources.

  -

  Investigation and verification of other documented megatsunamis not included in the GHTDs.

• Analysis of the relationships between maximum wave height and causes of historical tsunamis.
• Data compilation: Global Historical Megatsunamis Catalog (GHMCat)

  -

  Compilation and description of each verified historical megatsunami, including details on maximum wave height, causes, and other significant data and effects when available.

5. Definition of Megatsunami Based on Maximum Recorded Wave Heights

Based on the data from the American NCEI/WDS database [12], maximum wave heights (Hmax), or runup heights, were analyzed. Hmax is the only objective data documented and available for most significant tsunamis included in historical databases and catalogs, related to the physical characteristics of the events. The Russian TL/ICMMG database [13], which is very similar, was also analyzed for this purpose.

Of the more than 2800 events recorded in the databases, less than half—after excluding those classified as erroneous—include valid information on maximum wave height, on which the analysis focused; among these more than 1200 events, 88% have a maximum wave height of less than 10 m, 97% less than 30 m, and 9% between 10 and 30 m.

Figure 2 and Figure 3 represent the relationship between the number of recorded events and their corresponding Hmax values. Figure 3 illustrates that starting from an Hmax of 35 m, there is a significant decrease in the number of tsunamis, dropping from an average of 11 tsunamis for intervals between 20 and 35 m to an average of fewer than 2 events for Hmax intervals above 35 m. This marked decrease justifies selecting an Hmax value of 35 m as the threshold that distinguishes megatsunamis.

Therefore, a wave height of Hmax ≥35 m is proposed as the threshold to differentiate megatsunamis from regular tsunamis, classifying them as exceptional events. This threshold accounts for only 1.4% of the events recorded in the GHTDs—3% if only cases with documented Hmax are considered—further supporting the selection of 35 m as a significant threshold for defining megatsunamis.

A megatsunami can be defined as an exceptionally large wave that reaches an Hmax of 35 m or more, up to several hundred meters. This threshold is considered highly representative, effectively identifying an exclusive group of fewer than forty events among documented historical tsunamis, thus preserving the exceptional nature associated with the term itself.

6. Data Sources and Literature Review

6.1. Sources of Primary Information

The catalog is primarily based on the compilation of data sourced from the two existing Global Historical Tsunami Databases (GHTDs):

• The NCEI/WDS Global Historical Tsunami Database [12] supported by the National Geophysical Data Center of the National Oceanic and Atmospheric Administration (NOAA), USA.
• The TL/ICMMG Global Historical Tsunami Database [13] supported by the Tsunami Laboratory, Institute of Computational Mathematics and Mathematical Geophysics of Siberian Division of Russian Academy of Sciences, Russia.

The databases collect available and documented information on both the geological processes generating the tsunamis (tsunamigenic events), mostly earthquakes, and on the maximum wave heights observed or measured in all locations where data are available (multiple recorded tsunamis per event). The detail and veracity of the data depends on the source documents and, above all, the date of occurrence. The period of record covers the last 4 centuries. An overview of the development and evolution of the global historical databases since the 1980s and 1990s is provided by Gusiakov [39]; more recently, a detailed comparative study of the content of the two global databases has been published in 2020 [62].

Both databases are generally consistent in terms of data format and content (about 100 events from each database are not recorded in the other), with some differences in the estimates of maximum wave heights, confidence level, or degree of validity assigned.

The NCEI/WDS database contains information on over 2800 tsunami events (of which approximately 1200 include measured or estimated runup or maximum wave height data, 700 of these with heights >1 m) in the Atlantic, Indian, and Pacific Oceans, and the Mediterranean and Caribbean Seas, from 2000 BC to the present (2024). The TL/ICMMG database includes information on 2700 events (of which 1128 include some wave height or runup data, 600 with heights >1 m) in the World Ocean, for the same period of time. Both computerized, easily accessible databases can be kept constantly updated.

Most of the tsunamis have been caused by submarine earthquakes (~75%), especially those prior to 1500; the rest are associated with landslides (~9%), volcanic eruptions (4–5%), and meteorological effects (2–3%). Tsunamis caused by landslides have been recorded since the 17th century, increasing in number by the 20th century, especially from the 1960s onwards.

6.2. Literature Review

The literature review served a dual purpose: firstly, to validate the megatsunami data gathered from the GHTDs databases, focusing on runup and event sources, and secondly, to search for megatsunami events not documented in these global databases.

For these purposes, references cited in the GHTDs have been thoroughly examined, as well as any additional information on a particular megatsunami event, even searching, where feasible, for its earliest references. This involved scrutinizing numerous regional and local catalogs, scientific papers, published case studies, reports, books, and any other reliable source of information—including old catalogs and documents available online containing original data—and personal consultations with expert researchers. More than 300 publications spanning from 1888 to 2024 have been reviewed, with some referencing reports dating back to the 17th and 18th centuries, the earliest in 1675 [63] and ~1750 [64]. Old records frequently suffer from ambiguity and incompleteness, often written well after the event by those who were not eyewitnesses. Furthermore, discrepancies in the dates of older events may arise across bibliographic sources due to adaptations of calendars used in ancient records.

The main information on old events is found in catalogs published for various geographic and temporal scopes. The earliest historical tsunami catalog was compiled by N.H. Heck in 1934 [39], who summarized tsunami data from previous earthquake catalogs. Below are some of the most comprehensive and representative catalogs of historical tsunamis, which provide relevant data and additional insights into some of the megatsunamis referenced in this study:

• Iida, K., Cox, D.C. & Pararas-Carayannis, G. 1967. Preliminary catalog of tsunamis occurring in the Pacific Ocean [65].
• Soloviev, S.L. & Go, Ch.N. 1974. A catalogue of tsunamis on the western shore of the Pacific Ocean (173–1968) [25].
• Soloviev, S.L. & Go, Ch.N. 1975. A catalogue of tsunamis on the eastern shore of the Pacific Ocean (1513–1968) [26].
• Iida, K. 1984. Catalog of tsunamis in Japan and its neighboring countries [66].
• Lander, J.F. 1996. Tsunamis affecting Alaska 1737–1996 [30].
• Harris, R. & Major, J. 2016. Waves of destruction in the East Indies: The Wichmann catalogue of earthquakes and tsunami in the Indonesian region from 1538 to 1877 [67].

6.3. Addressing Uncertainties

The main uncertainties in handling historical information come from temporal (highly non-uniform distribution over time) and spatial biases in their documentation, and the reliability of the data, often compromised by inaccuracy and incompleteness, particularly regarding older or geographically remote events; additionally, uncertainties arise from errors in interpreting old descriptions, which accumulate with each translation. One method of addressing these uncertainties in tsunami catalogs and databases is by assigning a validity index, typically ranging from 0 or −1 (false or error) to 4 (definite or certainty).

For megatsunamis recorded within the last century, data on the maximum wave heights reached on land are reliable and supported by evidence, such as eyewitness accounts or measurements.

The review process led to ruling out some of the initially selected megatsunamis due to a lack of evidence or reliable references supporting the information collected in the GHTDs regarding the maximum wave heights. Since the data could not be verified and considering the uncertainties or significant doubts raised and documented by different authors, some cases were not finally classified as megatsunamis. Additionally, the documentary research led to identifying some events with runups lower or higher than the documented values, as well as other megatsunamis occurring in the last century not included in either GHTDs.

7. Data Collection and Verification

Initially, events were selected from the NCEI/WDS and the TL/ICMMG databases that met the proposed threshold for maximum recorded wave height or runup, Hmax ≥ 35 m. Each event from this list was thoroughly researched, consulting available references, including the earliest sources. The information was subsequently cross-checked and verified from various documentary sources. While some events are extensively documented in papers and reports, others are inadequately recorded, sometimes limited to a single publication. Furthermore, a systematic search was conducted for megatsunamis not included in the GHTDs. Additional contemporary documentary sources were sought, primarily recent scientific papers and technical reports, either published or available online.

7.1. Megatsunamis Included in the GHTDs

The GHTDs [12,13] contain records of 34 and 31 megatsunami events—i.e., with reported maximum water height Hmax ≥35—respectively, for the period 1674–2024, of which 29 coincide in the two databases. All events are classified as “definite”, indicating the highest level of validity of the actual tsunami occurrence, in both databases; however, in a few cases, significant differences exist in the reported maximum height or runup reached by the waves for the same event.

A total of 36 megatsunamis have been compiled. Data on these events are presented in

Table A1. Historical megatsunamis with Hmax ≥35 m recorded in the global databases NCEI/WDS [12] and TL/ICMMG [13] as of June 2024.

Date	Name/Place	Cause *	Runup (m) *
1600 BC	Santorini, Greece	V	90/-
1674	Ambon Island, Indonesia	LEq (6.8)/Eq (8)	100/80
1737	Kamchatka, Russia	Eq (8.5)	15/63
1741	Oshima Island, Japan	V/LV	90/10
1756	Langfjord, Norway	L/-	38/-
1771	Ryukyu Islands, Japan	Eq (7.4)/LEq	85.4
1788	Unga and Sanak Is., Alaska	Eq (8)	88
1792	Kyushu Island, Japan	LV/V	55/57
1853	Lituya Bay, Alaska	L	120
1880	Sitka, Alaska	LEq (6.3)	1.8/60
1883	Krakatoa Island, Indonesia	V	41/35
1896	Sanriku coast, Japan	Eq (8.3)/(8.5)	38.2
1899	Lituya Bay, Alaska	LEq (8.2)	61
1905	Lovatnet Lake, Norway	L	40
1934	Tafjord, Norway	L	62
1936	Lovatnet Lake, Norway	L	74/70
1936	Lituya Bay, Alaska	L	150
1946	Unimak Island, Alaska	LEq (8.6)/Eq (8.6)	42
1958	Lituya Bay, Alaska	LEq (7.8)/L	525
1963	Vaiont Reservoir, Italy	L	235
1964	Port Valdez Bay, Alaska	LEq (9.2)/Eq (9.3)	67
1965	Cabrera Lake, Chile	LV/V	60
1967	Grewingk Glacier Lake, Alaska	L	60
1980	Spirit Lake, EE.UU.	V	250
1985	Yangtze River, China	L	54
2000	Vaigat Strait, Greenland	L	50
2003	Qinggang River, China	L/-	39/-
2004	Sumatra Island, Indonesia	Eq (9.1)	50.9
2007	Aysén Fjord, Chile	LEq (6.2)/L	50/65
2007	Shuibuya Reservoir, China	L/-	50/-
2007	Chehalis Lake, Canada	L	38/37.8
2011	Sanriku coast, Japan	Eq (9.1)	39.7/42
2014	Askja Lake, Island	L/-	60/-
2015	Taan Fjord, Alaska	L	193
2017	Karrat Fjord, Greenland	L	90
2018	Bureya Reservoir, Russia	L	90
2018	Anak Krakatau, Indonesia	LV/V	85

L: Landslide, rockslide, or rock avalanche; LEq: Earthquake-triggered landslide, rockslide or rock avalanche; LV: Volcanic landslide; Eq: Earthquake (magnitude); V: Volcanic. * Where the two global databases differ, data from both are given (NCEI/WDS / TL/ICMMG).

(Appendix A), as recorded in the GHTDs as of 15 June 2024; the most recent megatsunami is from 2018. Six events have reached or exceeded a height of 100 m. The apparent accuracy of some of the older runup figures is due to the precision during unit conversions to the metric system, especially for events dating back in time.

The earliest documented event occurred in 1674 (excluding Santorini: a huge wave ~100 m high, the largest ever documented in Indonesia, generated after a violent earthquake, affecting the north coast of the island of Ambon; the earthquake and tsunami resulted in over 2500 casualties.

The volcanic tsunami originating from the island of Thera or Santorini, Greece, around 1600 B.C., is included in both databases. While only the NCEI/WDS database reports a maximum wave height of 90 m, historical evidence overwhelmingly supports the occurrence of a large tsunami triggered by the island’s explosion and collapse. Wave heights have been estimated at between 40 and 50 m according to geological evidence, although could have exceeded 250 m in height in the near-field [68,69,70,71]. Despite its undeniable significance, this event has not been included in the global catalog of megatsunamis presented here: due to its antiquity and the long period of time elapsed until the next documented case (in 1674), it is considered a particular case.

In the American database, the majority of the megatsunamis, 27 out of 35 events, are attributed to landslide origins (including 6 triggered by earthquakes and 3 by volcanic eruptions); additionally, 3 megatsunamis are attributed to volcanic eruptions and 5 to earthquakes. In the Russian database, 19 events are attributed to landslides (with 3 of them triggered by earthquakes), 5 to volcanic eruptions, and 8 to earthquakes.

7.2. Verification of Megatsunamis Included in the GHTDs

The examination of documentary sources for the events initially listed from the GHTDs (see Table A1, Appendix A) has resulted in:

• The exclusion of three events due to substantial doubts, uncertainties, or discrepancies in the reported data regarding maximum wave height: events of 1737, 1741, and 1880.
• The correction of Hmax values for four events: 1771, 1788 (heights have been lowered); 1756, 1896 (heights have been increased).
• The inclusion of two events, one in Canada in 1946 and another in Alaska in 1905, as megatsunamis, initially documented with Hmax < 35 m.

7.2.1. Events Excluded as Megatsunamis

The following cases have been ruled out: the earthquake-triggered tsunami of 1737 in the Kuril-Kamchatka region, Russia; the tsunami of 1741, caused by a large volcanic landslide on the island of Oshima-Oshima, Japan; and the 1880 tsunami caused by a seismic landslide in Alaska. For these events, there is a substantial disparity between the wave heights reported by the two GHTDs. In each case, only one of the GHTD records wave heights exceeding 35 m, based on oral references from old documents (see Table A1, Appendix A).

The maximum runup height of 63 m attributed to the October 1737 tsunami is considered unreliable. This value is based on observations of driftwood and trees washed over a cliff, not definitively linked to the tsunami [30]. Discrepancies arise from various documentary sources regarding the maximum wave height: 63 m according to written testimonies from an expedition in 1738 [64,72], and a range of 8–10 m, with maximum runup heights of 17–21 m, according to a recent thorough study of field data and written sources on this tsunami [73]. According to Lander [30], the maximum waves in Kamchatka could have reached 30 m, without citing any source. Consequently, given the uncertainties, no definitive conclusions can be drawn for the 1737 earthquake-triggered tsunami.

The October 1880 Alaskan tsunami purportedly reached a maximum wave height of 60 m in Sitka, according to the Russian GHTD, while the American GHTD records a height of 1.8 m, the former without citing the data source. Both include other tsunamis associated with the same earthquake (estimated at a magnitude of 6.3) with runup height not exceeding 2 m. Furthermore, these events are classified as “probable”, rather than definitive. According to [30], information from contemporary newspapers attributes the unusual waves to “very local submarine landslide-generated tsunamis”, with a maximum height of 1.8 m; however, the author also notes: a “huge wave ran into bay”, without further specification. Similarly, other authors describe qualitatively that a “large tidal wave appeared” after the earthquake or a “tidal wave of huge dimensions ran into the bay” [26,74]. These contradictory and inconclusive data, coupled with the fact that the purported 60 m wave occurred in a bay over 200 km away from where the other smaller tsunamis were recorded, suggests that the reported runup height is erroneous.

The August 1741 tsunami was caused by a landslide on the north flank of the small Oshima-Oshima volcanic island in the Sea of Japan, southwest of the island of Hokkaido. Historical records document the occurrence of volcanic activity, but do not record any earthquakes. The landslide, of about 2.4 km³ in volume, one of the largest landslides in history, affected both the subaerial and submarine flanks of the island. The reported maximum wave height of 90 m for the tsunami is not supported by evidence; it appears in the American database, originally from the Catalog of tsunamis in Japan and its neighboring countries [66] based on a single oral reference. According to reliable historical documents, the maximum runup was about 13 m, while data suggesting higher heights remain unverified and less reliable [75]. According to the Russian database and other reviewed catalogs and literature sources, the maximum wave height could reach about 9–10 m off the coast of Hokkaido [25,65,75].

Based on the aforementioned considerations, particularly the inability to verify maximum wave heights due to the absence of reliable historical records or geological evidence, the tsunamis discussed above are not classified as megatsunamis.

After excluding the 1741 tsunami, only two documented historical cases of megatsunamis from volcanic island flank landslide remain: the Mount Mayuyama event in 1792, Japan, and the more recent event on the island of Anak Krakatau, Indonesia, in 2018 (Table A1, Appendix A). Regarding the Krakatoa event in 1883, the ultimate cause of the megatsunami remains undocumented; however, it is very likely that it was caused by a volcanic flank landslide associated with the eruption [76].

The most recent large volcanic island flank landslide, although it did not cause a megatsunami, occurred on Ritter Island in March 1888, with a volume of ~2.4 km³, similar to that of Oshima-Oshima. Both of these events represent the only two large-volume (>1 km³) flank landslides historically recorded on volcanic islands, ranking among the largest historical on Earth. The Ritter landslide generated a tsunami with a maximum runup of 15 m along the nearby coastline [77].

7.2.2. Correction of the Maximum Wave Height Values

The literature review led to corrections in the maximum wave heights, either increasing or decreasing them, in six cases: the events of 1756, 1771, 1788, 1896, 1905, and 1946, the two last events initially being below the proposed megatsunami threshold (see

Table A2. Events recorded in the GHTDs for which the Hmax value has been corrected after the review and verification of the data, resulting in the exclusion of some with Hmax <35 m and documented historical megatsunamis not included in the GHTDs.

Date	Place/Name	Cause		Runup (m)		References
		GHTDs	This Study	GHTDs *	This Study
1737	Kamchatka, Russia	Eq (8.5)	Eq	63	21	[12,73]
1741	Oshima Island, Japan	LV	LV	90	13	[25,65,75]
1756	Langfjord, Norway	L	L	38	>50	[86]
1771	Ryukyu Islands, Japan	Eq (7.4)	SLEq	85.4	35	[78,81,82]
1788	Unga and Sanak Is., Alaska	Eq (8)	SLEq **	88	≥50	[30,84,85]
1880	Sitka, Alaska	LEq (6.3)	SLEq	60	<30	[26]
1896	Sanriku coast, Japan	Eq (8.3)	SLEq **	38.2	55	[88,89]
1905	Disenchantment Bay, Alaska	L	L	33.5	35	[7,30]
1946	Landslide Lake, Canada	LEq (7.3)	LEq (7.3)	30	51	[90]
1936	Lovatnet Lake, Norway	-	L	-	40	[9,49]
1936	Lovatnet Lake, Norway	-	L	-	>74	[9]
2007	Grijalva River, Mexico	-	L	-	>50	[118]
2018	Jinsha River, China	-	L	-	130	[122]
2020	Elliot Lake, Canada	-	L	-	114	[125]

Eq—Earthquake (magnitude); L—Subaerial landslide or rock avalanche; LEq—Earthquake-triggered subaerial landslide (magnitude); SLEq—Earthquake-triggered submarine landslide; LV—Volcanic landslide. * Where the two GHTDs differ, the higher value has been included (see Table A1); ** Probable proposed mechanism.

, Appendix A).

The maximum wave heights of the April 1771 event, known as the Great Meiwa Tsunami, were initially documented to exceed 85 m [12,13]. However, recent studies suggest a more reliable maximum height of around 30–35 m [78,79,80,81,82], discarding earlier overestimated records such as the 85 m reported on the southern coast of Ishigaki Island [66,83]. Soloviev and Go [25] attribute these discrepancies to confusion in measurement unit conversion at the time.

The databases record the occurrence of two tsunamis in 1788 in the Aleutian Islands, just a few days apart, attributed to great earthquakes, with maximum wave heights of 30 and 88 m. These represent the earliest recorded tsunamis in Alaska. In the documentation sources, there are discrepancies in the dates, likely due to the adjustment between the Julian and Gregorian calendars. Similarly, significant variations exist in the recorded maximum wave heights, possibly due to errors in converting measurement units from ancient documents, as many accounts of the 1788 events are based on third- or fourth-hand translations of Russian reports [84].

According to the most reliable documentation [30,84,85], the first tsunami occurred on July 21, coinciding with a major earthquake of magnitude 8 or greater, and the second one 16 days later, on August 6. The second tsunami was the highest, reaching maximum runup heights between 50 m and 90 m, and even over 100 m, on Unga and Sanak Islands, according to various accounts. There is no direct evidence of these maximum heights, as the data come from oral references transcribed and translated from original Russian descriptions. Considering possible conversion errors, it can be assumed that maximum heights reached at least 50 m [30,85]. For the first tsunami, confusing data are given, most of them ≤10 m [13,30,74,85]. The second tsunami has been associated with a strong aftershock of the preceding major earthquake [85]. However, the original texts do not mention that the second tsunami was accompanied by an earthquake, and there are some doubts about the occurrence of a second earthquake [74,84].

The maximum wave heights reached by the megatsunamis of 1756 and 1896 have been increased (see Table A2, Appendix A). For the 1756 event in Langfjord, Norway, maximum heights of 50 m were reached according to [86,87]. For the 1896 megatsunami, maximum wave heights reached up to 55 m, measured on the basis of traces left on the coast and eyewitness accounts [88,89].

The event that occurred in 1946 on Vancouver Island, Canada, is recorded in the two GTHDs with maximum heights of 9 m and 30 m, respectively. However, according to the reviewed literature, the largest wave reached a runup of 51 m [90]. The 1905 event in Yakutat Bay, Alaska, recorded with a maximum height of 33.5 m, generated waves that reached 35 m, as evidenced by the vegetation swept away on the slopes of the bay [7,30].

7.3. Documented Megatsunamis Not Included in the GHTDs

Five significant events, not recorded in global databases, have been identified as megatsunamis, thus significantly expanding the number of documented historical megatsunamis: two events in Norway in 1936, one in Mexico in 2007, one event in China in 2018, and one in Canada in 2020. These events are listed in Table A2 (Appendix A), with final runup values assigned after the examination of reliable data on the maximum wave height reached in each case.

7.4. A New Event in Lituya Bay Prior to 1786

Lituya Bay, on the northeastern coast of the Gulf of Alaska, was officially discovered by La Pérouse in 1786, during his global expedition. Cenotaph Island, a large wooded elevation at the center of the bay, is named for the 21 members of the La Perouse expedition who drowned when three small boats overturned in rough waters at the bay’s shallow inlet. A memorial was erected in their memory on the island.

The bay is renowned for having experienced the highest wave runup ever documented, reaching 525 m, and for being one of the two locations globally where the greatest number of megatsunamis have been recorded since the mid-19th century: at least four megatsunamis over a span of a little more than a century (1853, 1899, 1936, and 1958) have surged from the head of the bay. Except for the 1899 event, all are among the top ten largest historical megatsunamis worldwide, with heights exceeding 100 m. Each of these megatsunamis devastated the forested slopes, leaving distinct markers of the corresponding runup heights reached.

Another tsunami, with a runup of 20 to 30 m, occurred around 1874, as deduced from the trimline identified in a photograph from 1894, coupled with observations of vegetation growth stages both below and above the tsunami trimlines. The study states [7]: “The lower trimline definitely identified in a photograph of the north shore of Lituya Bay taken in 1894 indicates that at least one giant wave occurred before this date but later than the 1853–54 wave”. These ancient markings left by the waves on the vegetation of the slopes were largely obliterated during the 1936 megatsunami, and nearly entirely during the 1958 event.

The examination of historical documents for this study has led to the inference of another megatsunami occurring prior to 1786 in Lituya Bay [14]: an engraving of the bay created during the summer of that year by the expedition of the Frenchman La Pérouse, its official discoverer, clearly depicts the effects of huge waves on the bay’s slopes, with vegetation completely stripped away up to a certain elevation of several tens of meters a.s.l. (Figure 4) [91]. In the image, another lower trimline is clearly visible on the right-hand slope, indicating a separate, smaller tsunami event; furthermore, compelling evidence of megatsunamis occurring before 1786 is provided by the presence of a large rockslide on the cliff at the bottom of the bay, reminiscent of the earthquake-triggered rockslide of 1958 in the same area.

The details of the drawing undeniably suggest that both processes likely occurred shortly before the expedition’s arrival. Since the recorded history of Lituya Bay begins around 1788 with Russian exploration and settlement, there are no records to verify the occurrence of a megatsunami in earlier years. However, it can no longer be asserted that the oldest known megatsunami in Lituya is that of 1853.

The high frequency of megatsunamis in the bay (with at least five since the mid-18th century), may be partly due to the nearby Fairweather Fault, with very significant seismic activity; the 1958 megatsunami, and likely the 1899 event, were caused by rock and ice avalanches triggered by earthquakes [7,30]. The inferred megatsunami prior to 1786 cannot be linked to a significant earthquake, as the recorded history of Alaska begins around 1784, with the establishment of the first Russian settlements, four years before than a Russian ship entered Lituya Bay to claim the territory. The earliest documented earthquake and associated tsunamis in Alaska occurred in July–August 1788, two years after La Pérouse’s visit. These events were reported by a Russian missionary around 1840, who described a “very strong earthquake” and “terrible floods” [30,85].

8. Causes of Historical Tsunamis

Based on the data from the American NCEI/WDS database [12], relationships between maximum wave heights (Hmax) and tsunami causes were analyzed. The Russian TL/ICMMG database [13], which is very similar, was also analyzed for this purpose.

8.1. Maximum Wave Heights and Causes

Of the more than 2800 total events recorded in the global databases, less than half have information on maximum wave height or runup. After excluding events classified as erroneous or very doubtful, the analysis focused on a subset of about 1200 events, with 75% attributed to earthquakes. Among these 1200 events, 88% have Hmax <10 m, of which 78% are directly attributed to seismic activity. Earthquake-generated tsunamis exceeding 10 m represent 6% (approximately 3% if all recorded events are considered). Figure 5 illustrates the relationship between the total number of recorded events, their causes, and the Hmax values reached, according to the GHTDs.

As shown in Figure 5, starting from Hmax = 10 m, the number of tsunamis decreases sharply. Only 3.5% of cases exceed Hmax of 30 m, and approximately 2.5% exceed Hmax of 40 m, representing ~1.5% and ~1% of the more than 2800 recorded events, respectively. Earthquake-generated tsunamis are predominant up to Hmax = 30 m (75% of events), with only nine cases surpassing this height. From Hmax = 30 m onwards, the majority of tsunamis have been caused by landslides, accounting for 70% of the records, while seismic events represent 17%. From 40 m onward, only 32 cases are recorded, of which 79% are attributed to landslides, and 9% (3 events) to earthquakes, with the remainder attributed to volcanic origin. Between 30 and 40 m of Hmax, there is a substantial change in the causes of tsunamis, shifting from a majority of seismic origin to being almost entirely caused by landslides.

The relationships between the maximum tsunami heights and the physical parameters characteristic of the geological processes that cause them were analyzed for Hmax ≥30 m, as this value has been identified as a significant turning point. For this, the physical parameters that characterize the size or magnitude of the different geological processes causing tsunamis have been used: earthquake magnitude, landslide volume, and volcanic explosivity index. The analysis is included in Appendix B. The results show no significant relationship between these physical parameters and Hmax for each group.

8.2. Data Verification for Tsunamis with Hmax ≥30 m Attributed to Earthquakes

In order to document and verify the causes and their relationship with maximum heights recorded in the databases, tsunamis attributed to earthquakes with Hmax ≥ 30 m were thoroughly investigated. There are only nine cases surpassing this height (see

Table 1. Tsunamis caused by earthquakes with Hmax ≥30 m according to the NCEI/WDS database.

Year		Hmax (m)
1771	Japan	85.4
1788(1)	Alaska	30
1788(2)	Alaska	88
1896	Japan	38.2
1956	Greece	30
1957	Alaska	32
1993	Japan	32
2004	Indonesia	50.9
2011	Japan	39.26

): three events with Hmax >50 m (in the years 1771, 1788(2), and 2004), and the remainder ranging between 30 and 39 m (1788(1), 1896, 1956, 1957, 1993, 2011).

It has been found that, for the majority of these cases, there is either evidence of their origin from submarine landslides triggered by earthquakes, or this origin has been proposed as the only possible explanation for the excessive wave heights, which are not compatible with the earthquake magnitudes. This is the case for all the events except that of 1993 in Japan, as explained below. Additionally, for the cases of 1771, 1788, 1956, and 1957, research by various authors concludes that the maximum wave heights were overestimated (see Section 7.2.2).

In the case of 1771, the occurrence of a triggered-earthquake submarine landslide has been proposed as the cause of the tsunami, based on bathymetric and seismic investigations [78,83]. For the 1788 tsunamis, although there is no evidence, the possibility of massive underwater landslides triggered by the earthquakes has been suggested, also supported by historical descriptions of the events [74,84].

The largest waves of the 1956 tsunami were most probably caused by a series of submarine landslides off the coast of the island of Amorgos, and likely did not exceed a maximum height of 20–25 m [92,93].

It has been ruled out that the 1957 event reached a maximum wave height of 32 m, as all the references consulted indicate maximum measured heights of 12 to 15 m [25,30,84,94]. The only reference to wave heights of up to 22 m comes from an oral communication [30].

For the 2011 and 1896 events, both in Japan and very similar, with maximum runup heights close to 40 m, bathymetric observations and geophysical evidence support the possibility of near-field tsunamis generated by large and rapid submarine landslides as an additional mechanism responsible for the highest near-field waves [41,95,96,97], as happened with the 1946 and 1964 events in Alaska [42,98,99].

The 1993 earthquake in the Sea of Japan registered a magnitude of 7.8, with its epicenter a few kilometers southwest of Hokkaido Island. The earthquake triggered tsunamis and landslides on the slopes of the nearest coasts, killing hundreds of people. No references to a possible submarine landslide origin have been found for the main tsunami on Okushiri Island. According to Yamagishi [100], no submarine landslide occurred. Extensive post-tsunami measurements were made for tsunami runup heights, with a maximum value of ~32 m [101]. This extreme value is the highest recorded in Japan in the 20th century, and is among the highest ever documented for non-landslide tsunamis [102,103]. It was measured at the bottom of a narrow valley, undoubtedly influenced by local effects; the average height in the most affected coastal area, about 10 km long, was about 13 m [101,104]. After our analysis, this tsunami would be the highest recorded for an earthquake-triggered tsunami to date, based on the available information.

The 2004 tsunami is the only one with waves exceeding 40 m, with a maximum runup height of around 50 m, which represents the highest runup measured in history for a seismically generated tsunami. Similar to the 2011 Japan tsunami, it is attributed to one of the largest historical earthquakes ever recorded. The occurrence of extremely high waves has also been linked to the likely occurrence of submarine landslides as part of the tsunami source, in addition to the major thrust fault movement. Evidence of submarine landslides has been observed in the source area [105,106].

8.3. Results

After reviewing, verifying, and analyzing the events collected in the GHTDs [12,13] with available wave height data, the following observations have been made:

• For Hmax values <30 m, 75% of tsunamis originated from earthquakes.
• For Hmax values >32 m, 100% of tsunamis were caused by landslides.

Figure 6 outlines the relationships between the number of events, their causes, and maximum wave heights, following the verification of the data regarding the cause and wave height of tsunamis with Hmax ≥30 m.

Tsunami waves caused by earthquakes have not exceeded 30 m in height, except in a single case (Okushiri Island, 1993, with Hmax ≈32 m), where there is no geological evidence or indication of an associated submarine landslide as an alternative cause. From 32 m onwards, tsunamis have been caused by both subaerial and submarine landslides, often associated with earthquakes or volcanic eruptions, reaching wave heights of up to several hundred meters.

The results obtained have shown that maximum wave height is related to the geological processes generating tsunamis, reflecting the differential hydromechanical mechanisms of seismic and landslides waves, with significant differences in maximum wave heights. Exceptionally high waves are generated by non-seismic processes, such as subaerial and submarine landslides.

Consequently, there is a direct relationship between the causes of tsunamis and the maximum heights they reach, highlighting the exceptional nature of events caused by landslides.

Geomechanical and hydrodynamic conditions in each case primarily determine the potential wave height, influenced also by the morphological characteristics of the coast, such as orientation, slope and the presence of bays or fjords. Shallow water and confined bays or narrow straits can amplify tsunami wave heights.

The highest documented waves have been generated by large-volume subaerial landslides in specific environments such as bays, fjords, and even in rivers, lakes, or reservoirs. Landslides, in general, involve complex mechanisms where natural or induced factors trigger the movement of large volumes of rock on steep slopes, surpassing stability conditions and reaching very high velocities. Upon entering a body of water, these rock masses generate waves of extreme height.

The generating mechanism of tsunamis is also reflected in the extent of the affected areas: regional scale for seismic tsunamis and local for landslide-triggered tsunamis. In cases of submarine landslides triggered by large-magnitude earthquakes, the effects of both types of tsunamis overlap in the near-field zone. Landslide-triggered waves can reach the coast in nearby areas before seismic waves.

9. Results: GHMCat Data Compilation and Presentation

The global catalog of historical megatsunamis GHMCat comprises: (i) verified megatsunamis from the two existing global tsunami databases (GHTDs), and (ii) events not included in the databases, documented after a comprehensive literature research. A total of 40 megatsunamis have been compiled (excluding Santorini, 1600 B.C.), 33 of which are included in the GHTDs, and 7 are new contributions. These are listed chronologically in

Table 2. Historical documented megatsunamis included in the GHMCat *.

Date	Place/Name	Cause	Runup (m)	References
1674	Ambon Island, Indonesia	SLEq	100	[25,67]
1756	Langfjord, Norway	L	>50	[86]
1771	Ryukyu Islands, Japan	SLEq	35	[78,81,82]
1788	Unga and Sanak Is., Alaska	SLEq **	≥50	[30,85]
1792	Kyushu Island, Japan	LV	57	[107]
1853	Lituya Bay, Alaska	L	120	[7]
1883	Krakatoa Island, Indonesia	LV	>40	[76]
1896	Sanriku coast, Japan	SLEq **	55	[88,89]
1899	Lituya Bay, Alaska	LEq	61	[7]
1905	Lovatnet Lake, Norway	L	41	[9,47]
1905	Disenchantment Bay, Alaska	L	35	[7,30]
1934	Tafjord, Norway	L	62	[47]
1936	Lovatnet Lake, Norway	L	74	[9,47]
1936	Lovatnet Lake, Norway	L	40	[9,49]
1936	Lituya Bay, Alaska	L	150	[7]
1936	Lovatnet Lake, Norway	L	>74	[9]
1946	Unimak Island, Alaska	SLEq (8.6)	42	[42,43]
1946	Landslide Lake, Canada	LEq (7.3)	51	[90]
1958	Lituya Bay, Alaska	LEq (7.8)	525	[7,30]
1963	Vaiont Reservoir, Italy	L	235	[58,108]
1964	Port Valdez Bay, Alaska	SLEq (9.2)	67	[109]
1965	Cabrera Lake, Chile	L	60	[110]
1967	Grewingk Lake, Alaska	L	60	[111]
1980	Spirit Lake, USA	LV	260	[12,112]
1985	Yangtze River, China	L	54	[57,113]
2000	Vaigat Strait, Greenland	L	50	[114]
2003	Qinggang River, China	L	39	[113,115]
2004	Sumatra Island, Indonesia	SLEq (9.1)	~50	[33,34]
2007	Aisen Fjord, Chile	LEq (6.2)	65	[116]
2007	Shuibuya Reservoir, China	L	50	[117]
2007	Grijalva River, Mexico	L	>50	[118]
2007	Chehalis Lake, Canada	L	38	[50,119]
2011	Sanriku coast, Japan	SLEq (9.1)	~40	[88,95]
2014	Askja Lake, Iceland	L	>60	[51]
2015	Taan Fjord, Alaska	L	193	[38]
2017	Karrat Fjord, Greenland	L	90	[120,121]
2018	Jinsha River, China	L	130	[122]
2018	Bureya Reservoir, Russia	L	90	[53]
2018	Anak Krakatau, Indonesia	LV	85	[123,124]
2020	Elliot Lake, Canada	L	114	[125]

* The Santorini event, 1600 B.C., is considered a particular case due to its great antiquity and is not included in the catalog (see Section 7.1). The megatsunami in Lituya Bay prior to 1876 is also not included, as its age is unknown (see Section 7.4 for explanation). L: Subaerial landslide, rockslide or rock avalanche; LEq: Earthquake-triggered subaerial landslide; SLEq: Earthquake-triggered submarine landslide; SLEq **: Probable proposed mechanism; LV: Volcanic flank landslide; Eq: Earthquake (magnitude).

. When multiple values for runup height were found in the literature, the largest verified value was included in the table. In addition to the maximum water height (or runup), the causes of megatsunamis have been verified, particularly in the oldest cases for which new documentary and field research has provided more reliable data.

The compilation and description of each event constitute the megatsunami catalog GHMCat, which is included in Section 10. Figure 7 shows the geographic locations of megatsunamis, and Figure 8 shows their distribution over time, along with the respective runup heights.

10. Global Historical Megatsunami Catalog (GHMCat) 1674–2024: Description of Events

• 1674, February 17—Ambon Island, Indonesia
• Runup: 100 m
• Cause: Earthquake-triggered landslide (submarine?)

The Ambon earthquake of 1674 *, one of the most devastating earthquakes in the Moluccan archipelago’s history, and the most violent that the island of Ambon had experienced, was followed by a large tsunami reaching runup heights of up to 100 m in the northern coast of the island, based on the distinctive coastal vegetation trimline it left.

It was the first documented tsunami in Indonesia, with detailed descriptions in a text written by the German botanist and naturalist Rumphius in 1675 [63], and the largest runup height ever recorded in the country. Ancient documents refer to the occurrence of large coastal landslides on the north shore of Ambon as a consequence of the earthquake, generating waves reaching heights of 90–110 m [25,67]. The impact was catastrophic, with part of the coast sliding into the water, carrying whole villages and coastal hills plantations and vegetation, killing more than 2300 people. Witnesses accounts provided detailed descriptions of the events: “the water rose up like a mountain … carrying with it trees, houses, domestic livestock and people” [25] or “a great mountain of sea came on with great rumbling and crashed against the beach”, as documented in Wichmann’s 1918 catalog [67].

The extreme runup height observed suggests that an earthquake as a direct source of the tsunami is unlikely; according to eyewitness accounts, the cause was large landslides on the north coast of Ambon. Some researchers propose that a submarine landslide generated by the earthquake may have been the source [63], which also washed away a large part of the coastal strip.

Ref.: [25,67]

* Although this date is mentioned in most bibliographical references, the event likely occurred in 1675 (Wichmann, 1918; in [67])

• 1756, February 22—Langfjord, Norway
• Runup: >50 m
• Cause: Subaerial rock avalanche

The largest historically recorded rock avalanche in Norway—and northern Europe—took place in 1756 at Tjelle, in Langfjord, involving a volume of 10–15 M m³ that fell from an altitude of 400 m a.s.l. The impact of the huge rock mass on the sea generated waves of up to 50 m in height, with possible maximum runups of 200 m, although there is no evidence for this. A total of 32 people were killed, and hundreds of houses and boats around the fjord were destroyed in areas up to 40 km away. This tsunami is considered one of the most devastating events to occur in Norway in the last 500 years.

Other historical records of megatsunamis associated with rock avalanches in Norway include the 1905 and 1936 events in Lake Lovatnet and the 1934 Tafjord event.

Ref.: [86,87]

• 1771, April 24—Ryukyu Islands, Japan
• Runup: 35 m
• Cause: Earthquake-triggered submarine landslide

The 1771 tsunami, known as the Great Meiwa Tsunami, struck the southern Ryukyu Islands, resulting in 12,000 fatalities. According to historical records, maximum wave heights could reach over 85 m in the southern coast of Ishigaki Island [12,83]. However, recent publications propose a more reliable maximum height around 30–35 m [78,79,80,81,82], dismissing older data from historical records deemed to be overestimations, such as the 85 m height reported just after the disaster [66,83]. Soloviev and Go [25] attribute these discrepancies to confusion in the conversion of units of measurement at the time.

Regarding the cause of the tsunami, it has been attributed to a magnitude 7.4 earthquake or an earthquake-triggered landslide [12,13]. However, there is no historical record of an earthquake coinciding with the tsunami, being considered that a large-scale submarine landslide was the cause of the waves [83]. Subsequently, it has been suggested that the tsunami was accompanied by ground shaking, proposing the occurrence of a large volume landslide based on submarine bathymetric and seismic evidence as the cause of the megatsunami [78,80].

Ref.: [78,81,82]

• 1788, August 6—Unga and Sanak Islands, Alaska
• Runup: ≥50 m
• Cause: Earthquake-triggered submarine landslide (proposed)

The 1788 tsunami, the earliest dated tsunami recorded in Alaska shortly after the area was first settled by Russians, is attributed to a magnitude 8 earthquake [13]. The reported maximum runup of 88 m at Unga Island or Sanak Island is highly questionable and lacks documentation, as it comes from oral testimonies published several decades later.

According to various documentary sources, waves may have reached runup heights between 30 m and 90 m on the south coasts of the islands [30,85]. Most credible estimates, considering measurement corrections, suggest that waves could reached at least 50 m [56,84].

Some of the above references assert that the megatsunami could not solely be attributed to seabed displacement by tectonic rupture. In fact, old reports indicate that the tremors triggered landslides from mountains and shores [30,74]. Despite the lack of geological evidence so far, a source involving a massive underwater landslide triggered by the earthquakes cannot be ruled out [84].

Ref.: [74,84,85]

• 1792, May 21—Kyushu Island, Japan
• Runup: 57 m
• Cause: Subaerial volcanic flank landslide

The 1792 tsunami in the Ariake Sea (Japan) was caused by a flank landslide of Mount Mayuyama (in the Unzen volcanic complex on Kyushu Island), with a displaced volume of about 0.35 km³ (350 M m³). It occurred at the end of a period of intense eruptive and seismic activity in the area, causing collapses and landslides on the mountain sides. The huge rock mass moving towards the sea devastated the city of Shimabara, generating a destructive tsunami which swept around the enclosed Shimabara Bay, causing over 15,100 deaths, more than 10,000 due to the tsunami.

The height of the wave was estimated at 35–55 m at Shimabara. According to [107,126], the maximum wave height was 57 m, as depicted on an ancient map collecting wave height data along the coast, and could have been due to local coastal orographic effects and seafloor topography. The tsunami also caused great destruction and deaths on the Ariake Sea shores facing the landslide, with waves up to 24 m in height [126,127]. The length of the coast affected by the tsunami’s destruction was 75 km, and 17 villages were washed away. Some small islands situated near the coast disappeared, and the slide deposits formed several new small islands in the bay.

It is considered the largest volcanic disaster in Japanese history and one of the greatest landslide disasters of the last millennium, known as the Shimabara catastrophe. It is the second deadliest volcanic tsunami in the world after Krakatoa in 1883 (no data are available for the Santorini tsunami, approximately 1600 years B.C.).

Ref.: [52,107,127,128]

• 1853, November 30—Lituya Bay, Alaska
• Runup: 120 m
• Cause: Subaerial rock/ice avalanche

The first documented megatsunami at Lituya Bay on the northeast coast of the Gulf of Alaska—known for experiencing the largest wave in history in 1958—occurred approximately in 1853 or 1854, triggered by a rockslide. The event does not seem to have been associated with an earthquake. The megatsunami reached runup heights of 120 m as a result of the impact of the huge mass of rocks and ice that fell to the bottom of the narrow bay. According to Indigenous stories, eight canoes filled with people were lost due to a huge flood [129].

This is one of the largest historical megatsunamis in the world, among the eight with runup heights over 100 m. The evidence was discovered by USGS geologist Don Miller during his field investigations in 1952–1953, when he recognized distinct sharp trimlines marking the upper limits of destruction of the forest on the slopes of the bay—with young growth below and old growth above—which he initially attributed to cataclysmic floods or waves of water, without knowing the causes. The trimlines were mapped, and their approximate age determined by counting growth rings on older trees affected by the event, thus inferring that a wave had cleared the forest to a height of 120 m a century earlier [7,30]; the highest trimline was dated to 1936.

Some years later, after the megatsunami of 1958, it was discovered that the giant waves were caused by the impact of huge masses of rock and ice falling violently to the bottom of the narrow bay.

Ref.: [7,30]

• 1883, August 27—Krakatoa Island, Indonesia
• Runup: 41 m
• Cause: Volcanic flank collapse/Caldera collapse

The 1883 tsunami, the largest, most extensive and devastating volcanic tsunami ever recorded, was generated by the colossal eruption of Krakatoa, one of the largest volcanic explosions in history. Several cataclysmic explosions occurred during the eruption, the first one at about 17:00 (GMT) on 26 August. At 10 am on the morning of the 27th, the fourth and largest explosion occurred, with an explosivity index (VEI) of 6. This explosion was followed by massive flank failure and caldera collapse; two thirds of the island of Rakata, or Krakatoa, disappeared, generating the most destructive tsunami waves [130].

The final collapse of a still-standing part of Krakatoa, several hours later, generated additional waves. The local effects of the main tsunami along the Sonda Strait, on the nearest coasts of Java and Sumatra, were devastating: within an hour after the fourth explosion and collapse, waves reached heights of up to 40–41 m, flooding several kilometers inland, and destroyed 295 towns and villages, drowning more than 36,400 people. The island of Rakata disappeared for the most part (70%) after the cataclysm.

Historical documents from the 19th century record maximum wave heights between 35 and 41 m at Merak (Java) [76], which also appear in the catalogs of Heck [131] and Iida et al. [65]. Although there is no evidence, it is highly probable that the violent processes during the eruption and the collapse of the flanks caused waves of much greater height, 100 m or more, on the slopes around the volcano itself and nearby islands.

The ultimate cause of the 27 August megatsunami (flank failure, caldera collapse, underwater explosion) is unknown, as there was no direct observation of the phenomena that caused the giant waves. However, according to the 1888 report of the Krakatoa Committee of the Royal Society of London [76] “the destructive waves in the Strait of Sunda were mainly due to these masses falling into the sea, or to sudden explosions under the sea”.

Ref.: [65,76,130]

• 1896, June 15—Sanriku coast, Japan
• Runup: 55 m
• Cause: Earthquake-triggered submarine landslide (proposed)

The so-called Great Meiji Sanriku Tsunami that accompanied the earthquake of 1896 (estimated to have a magnitude approximately M~8.3), caused waves with runup heights that reached nearly 40 m [12,13]. The tsunami’s maximum heights, based on reports from the original field surveys published soon after the earthquake, range from 38.2 m to 55 m, as measured from traces left on the coast and eyewitness accounts [88,89].

Given the disparity between the magnitude of the earthquake (estimated at M = 7.2) and the height reached by the waves, either a large submarine landslide triggered by the earthquake or a slow tsunami earthquake type has been proposed as a possible source [95]. Due to its similarities with the megatsunami that occurred in the same area in 2011, which is thought to have been caused by a submarine landslide based on geological evidence, this same origin has been proposed for the 1896 tsunami [96,97].

This was the largest tsunami disaster in Japan’s history up to that time, with 26,000 deaths.

Ref.: [88,89,95,97]

• 1899, September 10—Lituya Bay, Alaska
• Runup: 61 m
• Cause: Earthquake-triggered subaerial landslide/rock avalanche (M~8.2)

The 1899 megatsunami was probably caused by a rock and ice avalanche on its steep slopes, triggered by an earthquake. Waves reached a runup height of 61 m [7]. Based on some unconfirmed oral accounts and on marks—trimlines—left in the forests on the slopes, reflected in aerial photographs from the first decades of the 20th century [7], the occurrence of the tsunami was inferred between 1853 and 1936. The date 1899 has been attributed as it coincides with the catastrophic earthquake of magnitude >8 in nearby Yakutat Bay [26,30], which caused a tsunami in Yakutat Bay itself of approximately 10 m high, as described by Tarr and Martin in 1912 [132]—these authors were among the first to use the term tsunami as a synonym for earthquake water waves, discarding the term tidal waves frequently used until then—. The 1899 earthquake caused many large collapses and slides, generating several separated tsunamis in the bays of the area; the largest wave would have been in Lituya Bay, washing away vegetation to a maximum height of 61 m on the coastal slopes at the bottom of the bay.

Ref.: [7,26,30]

• 1905, January 16—Lovatnet Lake, Norway
• Runup: 41 m
• Cause: Subaerial rock avalanche

The sudden entry of a rock and glacial debris avalanche into Lovatnet Lake, at the easternmost end of Nordfjord, caused devastating waves with a maximum runup height of 41 m. The final volume of the rock mass entering the lake, from the eastern side of Mount Ramnefjell, from a height of about 500 m, was estimated at 350,000 m³. The waves spread along the lakeshore, causing 61 deaths, mainly in the town of Loen and other smaller coastal settlements, where the waves reached more than 15 m.

Ref.: [9,47]

• 1905, July 4—Disenchantment Bay, Alaska
• Runup: 35 m
• Cause: Glacier landslide

A hanging glacier, since then known as the Fallen Glacier, slid into Disenchantment Bay, at the end of Yakutat Bay, falling 300 m down a steep slope into the bay. The resulting tsunami swept away vegetation on the slopes of the bay to a height of 33.5 m, and reached 35 m at Haenke island, in front of the glacier (Lander, 1996; Miller, 1960). A renowned geologist, Dr. R.S. Tarr, who was working in Russell Fjord, an inlet of Yakutat Bay, 24 km from the avalanche site, witnessed a series of large waves that reached heights of up to 6 m (Lander, 1996).

Ref.: [7,30]

• 1934, April 7—Tafjord, Norway
• Runup: 62 m
• Cause: Subaerial rock avalanche

The megatsunami that struck Tajford, a town at the end of a long, narrow fjord, was triggered by the impact of a debris and rock avalanche, estimated at 1.5–2 M m³, that fell from the steep wall of the fjord into the sea. The height of the waves generated by the avalanche reached a maximum height of 62 m a.s.l., resulting in the deaths of 40 people. This event is considered one of the most severe natural disasters in Norway.

Ref.: [47]

• 1936, September 13—Lovatnet Lake, Norway
• Runup: 74 m
• Cause: Subaerial rock avalanche

On this occasion, after several minor rockfalls on Mount Ramnefjell, 1 M m³ of rock fell into Lake Lovatnet from a height of about 800 m, causing a giant wave that reached a maximum height of 74 m above the lake, directly in front of the source area of the rock avalanche. The tsunami resulted in 74 deaths, making it the deadliest in Norwegian history. Thirty-one years earlier, in 1905, a similar event had occurred in the same location, and several occurred later in the same year, 1936.

Ref.: [9,47]

• 1936, September 21—Lovatnet Lake, Norway
• Runup: 40 m
• Cause: Subaerial rock avalanche

Approximately 100,000 m³ of rock fell from Mount Ramnefjell, from a height of 800 m above the lake, generating a wave that reach a height of about 40 m.

Ref.: [9,49]

• 1936, October 27—Lituya Bay, Alaska
• Runup: 150 m
• Cause: Subaerial landslide/rock avalanche

The cause of the megatsunami was probably a large subaerial landslide or avalanche of rocks and ice with no apparent relation to any earthquake in the area, as none was reported or recorded in the region around that time. The waves reached a maximum height of 150 on the slopes at the bottom of the bay [7].

Four people observed the event, two in a small hut on Cenotaph Island in the middle of the bay, and two in a small fishing boat. Their eyewitness accounts, along with observations from others who visited the bay a few days after the event, led to the first known published references to the unusual “waves or floods” of water in Lituya Bay, in 1936 [7].

The maximum wave height was deduced from the trim lines left by waves that swept over the wooded vegetation on the bay slopes, detected during field investigations by Miller in 1952–1953. The age of the event was later verified by growth rings on the oldest trees affected by the event [7].

This is one of the largest historical megatsunamis in the world, among the eight with evidence of runup heights exceeding 100 m.

Ref.: [7,27,30]

• 1936, November 11—Lovatnet Lake, Norway
• Runup: >74 m
• Cause: Subaerial rock avalanche

A rockslide from Mount Ramnefjell fell into the lake, generating a wave that reached a heigt of more than 74 m. There were no casualties because the area had been devastated and abandoned due to similar events that had occurred in the preceding weeks. The total estimated volume of rock involved in the slides of 21 September and 11 November 1936, was well over 1 M m³.

Ref.: [9]

• 1946, April 1—Unimak Island, Alaska
• Runup: 42 m
• Cause: Earthquake-triggered submarine landslide (M 8.6)

The tsunami occurred shortly after a submarine earthquake of magnitude 8.6 (the second-largest earthquake in North American history) in the Aleutian Trench, northern Pacific Ocean. Unexpectedly, a huge wave suddenly engulfed and completely destroyed the newly built Scotch Cap lighthouse on Unimak Island, leaving nothing but the foundation of the five-story, 27 m-high concrete structure standing on top of a bluff 10 m a.s.l., killing all five lighthouse keepers.

The highest runup measured at Scotch Cap was 42 m above tide level, based on the height of driftwood and beach materials deposited by the tsunami; runups along the rugged coastline from Scotch Cap to 30 km to the east ranged between 24 m and 42 m [43].

Although scientists initially believe the wave was due to the earthquake rupture, i.e., a tsunami earthquake—that generates waves much larger than expected from the surface-wave magnitude—the arrival time of waves on the coast of Unimak, their large amplitude, distribution and the heights reached in the area near the epicenter (near-field tsunami) are inconsistent with a seismic origin. These observations can only be explained by a local source like a submarine massive landslide. Furthermore, geophysical investigations have revealed the existence of a large 200–300 km³ submarine landslide on the Aleutian shelf, probably triggered by the earthquake, and proposed as the origin of the megatsunami [42,98]. Additionally, a geophysical survey off the coast of Unimak Island identified an accumulation of large boulders that could have originated from a massive submarine landslide, which might have caused the giant waves at Unimak [99].

In addition to the local effects near the earthquake source, the seismic tsunami travelled across the ocean, reaching the Hawaiian archipelago in a few hours, with runup heights exceeding 10–12 m [25,133,134] causing extensive damage and resulting in 170 fatalities. The coasts of Chile and Antarctica were also affected by large waves. This trans-Pacific tsunami was one of the most destructive, leading to the establisment of the first Seismic Sea Wave Warning System in the United States in 1949, which later became the Pacific Tsunami Warning Center (PTWC).

Ref.: [42,43,98,99]

• 1946, June 23—Landslide Lake, Canada
• Runup: 51 m
• Cause: Earthquake-triggered subaerial landslide (M~7.3)

The most recent megatsunami was caused by a rock avalanche triggered by a 7.3 magnitude earthquake on the north face of Mount Colonel Foster, located in the center of Vancouver Island. Approximately 1.5 M m³ of rock and debris fell from an altitude of between 1965 m and 1600 m [90]. About half of the rock mass, approximately 0.7 M m³, fell into a lake at an altitude of 890 m, which has since been called Landslide Lake.

The wave resulting from the tremendous impact reached a runup of 51 m on the opposite shore of the lake, and overtopped the shores by about 29 m, destroying the valley forests up to a distance of 3 km from the lake.

This megatsunami wave is not recorded in global historical databases; however, other tsunamis on the same day and from the same source are recorded in other areas of Vancouver Island, with maximum wave heights of 30 m.

Ref.: [90]

• 1958, July 9—Lituya Bay, Alaska
• Runup: 524 m
• Cause: Earthquake-triggered subaerial rock/ice avalanche (M~7.8)

The highest wave runup in recorded history resulted from a rock avalanche triggered by a magnitude 7.8 earthquake impacting the waters at the head of Lituya Bay. The avalanche violently threw 35–40 M m³ of rocks and ice into the sea from an average height of 600 m [30]. Due to the energy of the impact, the water surged to a height of 524 m a.s.l. on the shore opposite the rockslide, devastating 10 km² of forest.

The waves spread out across the narrow, 11-km-long bay, passing by Cenotaph Island in the center of the bay, clearing trees to a height of about 50 m, and surpassing the entrance to the bay [30]. This is one of the cases where there is no dispute about the maximum height reached by the water, as it is clearly marked on the slopes by the line of trees that were washed away. This was documented by USGS geologist Don Miller immediately after the event [7].

The megatsunami, witnessed by several people on three boats anchored in the bay for the night, gave rise to a series of novel studies, including the rigorous work of Miller, which compiles the information available up to that time and the results of his fieldwork and observations in the bay. The great height reached by the water is attributed to the narrowness of the semi-enclosed bay in the area where the massive landslide occurred.

The 1958 earthquake in southeast Alaska triggered other major subaerial and submarine landslides that caused several tsunamis in the region, resulting in a total of five fatalities, two of them in Lituya Bay.

Ref.: [7,30,46]

• 1963, October 9—Vaiont Reservoir, Italy
• Runup: 250 m
• Cause: Subaerial landslide

The megatsunami at the Vaiont reservoir, in the Alps of northeastern Italy, was caused by a very rapid rockslide on the southern slope of Mount Toc, induced by man-made activities during the filling operations of the reservoir. The catastrophic failure occurred on a curved pre-existing sliding surface in Cretaceous limestones interbedded with clay layers.

The following figures give an idea of the magnitude of the landslide: a rock mass with a volume of 270 M m³, 250 m thick and 1800 m long; the landslide reached a speed between 70 and 100 km/h, and occurred in just over 45 s, filling up a large part of the reservoir, as it remains today.

The violent entry of a huge mass of rock into the dammed water displaced some 50 M m³ of water, triggering a wave that reached a height of up to 250 m above the reservoir level. The wave hit the opposite slope, destroying several small villages and parts of the towns of Erto and Casso, the latter more than 200 m above the reservoir level at the time of the catastrophe, causing 347 deaths.

In addition, the giant wave overtopped the 262 m-high dam (the world’s second highest dam at the time) and swept through the Piave river valley via the Vaiont gorge, destroying the village of Longarone and others downstream, killing 2000 people. As the wave overtopped the dam, the water reached 100 m above its crest.

The Vaiont megatsunami is the second highest wave ever recorded in an enclosed body of water, surpassed only by the 1980 event caused by the St. Helens volcano landslide.

Ref.: [58,108]

• 1964, March 28—Port Valdez Bay, Alaska
• Runup: 67 m
• Cause: Earthquake-triggered submarine landslide (M 9.2)

The 1964 megatsunami in Port Valdez Bay (Prince William Sound, Alaska) was caused by submarine landslides triggered by the Great Alaska Earthquake. This is the largest ever recorded in North America and the second largest in the world, after the 1960 Chile earthquake.

The earthquake generated several submarine and subaerial landslides that caused more than 20 large local tsunamis; the largest of these reached 52 m at Port Valdez, where sand and silt debris deposited by the waves was also found at a height of 67 m high [109]. In some nearby bays, waves exceeded 30 m in height.

Geophysical and bathymetric studies have provided evidence for the cause of these local tsunamis, revealing the presence of large submarine debris deposits in the bay, considered to be the product of major failures and the probable source of the devastating tsunamis on the Port Valdez coastlines [44,135]. The estimated volume of the 1964 submarine debris flows is 1 km³. Several large glide blocks (up to 40 m high and 300 m across) and debris deposits appear on the sea floor in the western part of the bay, where the megatsunami were observed.

Local tsunamis caused the larger waves and most of the more than 100 fatalities, reaching the nearby coast immediately after the earthquake, within a few minutes. They were followed by the main and much more extensive tsunami of tectonic origin, with arrival times between 20 and 45 min later [30]. The far-field tsunami inundated the coast of the Gulf of Alaska with maximum waves of approximately 10 m [109], travelling across the ocean and reaching the entire west coast of the USA, as well as the coasts of Hawaii and Japan. It remains the largest and most destructive tsunami ever observed in Alaska and the west coast of the United States. The Alaskan Tsunami Warning Center was established following the 1964 disaster.

Ref.: [44,109,135]

• 1965, February 19—Cabrera Lake, Chile
• Runup: 60 m
• Cause: Subaerial landslide

The 1965 tsunami occurred in the southern volcanic area of the Andes, caused by a large landslide, possibly induced by heavy rainfall, that included part of the summit of the Yate volcano. The landslide entered Lake Cabrera, generating a wave with a recorded runup height exceeding 50 m, reaching an estimated height of 60 m.

Between 6 and 10 M m³ of rock and ice fell from an altitude of 2000 m, in the form of a large debris avalanche that slid down the slope and violently entered Lake Cabrera, more than 7.5 km away and situated at an altitude of 1500 m a.s.l., causing the megatsunami that killed 27 people [110].

Although classified as a volcanic landslide [12] the rock mass movement that originated the megatsunami was not related to volcanic eruptions or processes.

Ref.: [110]

• 1967, October 14—Grewingk Glacier Lake, Alaska
• Runup: 60 m
• Cause: Subaerial landslide

The megatsunami at Grewingk Lake, located at the toe of the Grewingk glacier in the southern Kenai Mountains, was triggered by a weathered rock and debris avalanche of approximately 84 M m³. The avalanche slid into the lake, producing a large wave more than 60 m high, which stripped forest from the surrounding landscape.

This event is not included in the catalogs and main references on tsunamis in Alaska [30]. The available information comes from fieldwork aimed at characterizing the height and extent of the tsunami inundation [111].

The cause of the landslide is uncertain; the 1964 Alaska earthquake, which triggered many subaerial landslides, could have influenced the stability conditions of the steep valley slopes exposed after the glaciers retreated. In addition, unusually high precipitation was recorded in the month before the landslide, marking the highest September precipitation since records began in 1943.

Ref.: [111]

• 1980, May 18—Spirit Lake, USA
• Runup: 260 m
• Cause: Volcanic flank landslide

In May 1980 the largest geological cataclysm in recorded history on the northeastern Pacific coast occurred: a colossal eruption of the St. Helens volcano, in Washington. The eruption included lateral explosions, pyroclastic flow waves, and the destruction of part of the volcanic edifice, reducing its elevation from 2950 m to 2549 m. It is the deadliest—fifty-seven people were killed—and the most economically destructive volcanic event in the USA history.

The northern flank of the volcano collapsed, producing the largest landslide ever recorded in historical times, with a volume of 2.9 km³. The landslide became a huge avalanche of rock and ice moving at 170 to 200 km per hour. A portion of the mass violently entered Spirit Lake, to the north of the volcano, while most of it flowed westward over mounds and obstacles, advancing more than 20 km. The debris avalanche partly filled the lake, raising its surface by 64 m and plugging its natural drainage under 84 m of rock and soil.

The displaced water generated a local tsunami with a giant wave reaching over 250 m high (260 m and 265 according to sources [112,136], respectively) above the original lake level, sweeping away all vegetation and trees on the surrounding slopes below the sharp trimline left by the wave. This is the second-highest historic megatsunami on record, after the Lituya Bay megatsunami in 1958.

The Spirit Lake event is the only documented case of a megatsunami caused by a flank landslide of a continental volcano. The St. Helens landslide is the first observed and recorded case of a large-scale volcanic flank landslide, marking a definitive milestone in the study of the mechanisms of mega-flank landslides on volcanoes.

Both the Krakatoa eruption event of 1883 and the St. Helens volcano landslide, 100 years apart, are exceptional historical cases that have helped to understand the processes that can cause megatsunamis in volcanic environments.

Ref.: [12,112]

• 1985, June 12—Yangtze River, Three Gorges Region, China
• Runup: 54 m
• Cause: Subaerial landslide

The Xintan landslide, with a volume of 30 M m³, occurred on the left bank of the Yangtze River, destroying the ancient town of Xintan. Fortunately, the residents of the affected area were evacuated before the landslide occurred. About 2.6 to 3 M m³ of the slid rock-soil mass entered the river, generating huge waves with a runup over 54 m high on the opposite shore, affecting 42 km along the river channel, killing 12 people and sinking nearly 80 ships and boats. The water course was blocked for 12 days [57,113].

Other sources provide varying data on the maximum height reached by the water: 49 m [137] and 70–80 m (described as “water jet-flow”) according to witnesses [57].

Ref.: [57,113]

• 2000, November 21—Vaigat Strait, Greenland
• Runup: 50 m
• Cause: Subaerial landslide

The tsunami occurred on the west coast of Greenland and swept along the shores of the narrow Vaigat Strait near Paatuut. It was caused by a large and rapid landslide of 90 M m³ of basaltic rock and debris, of which 30 M m³ entered the sea; the head of the slide reached up to 1400 m a.s.l. The slopes of the area of the strait are prone to instability, with frequent coastal landslides.

The wave reached a height of 50 m near the impact site, and 28 to 30 m on the shores across the strait [114]. There were no fatalities, but the surrounding area and a nearby village were severely damaged. The tsunami reached 250 m inland and 30 m a.s.l., destroying 10 boats.

Ref.: [114]

• 2003, July 14—Qinggan River, Three Gorges Reservoir, China
• Runup: 39 m
• Cause: Subaerial landslide

The Qianjiangping landslide, with a volume of approximately 24 M m³, occurred on the western side of the Qinggang River, a narrow, shallow watercourse tributary of the Yangtze River, during the filling phase of the Three Gorges Reservoir [115].

Part of the landslide volume, about 2.4 M m³, entered and blocked the Qingjiang River, producing waves with a maximum height of about 39 m [117], which propagated through the reservoir reaching 30 km from its source. As a result, 11 fishermen were killed, along with 13 other people on the slope, and hundreds of buildings were destroyed [115,138].

Ref.: [113,115]

• 2004, December 26—Sumatra Island, Indonesia
• Runup: ~50 m
• Cause: Earthquake-triggered submarine landslide (M 9.1)

The 2004 Indonesian tsunami is one of the largest and most extensive in history and the deadliest, with more than 225,000 people killed. The highest waves were on the order of 35 m, with a maximum runup of about 50–51 m—measured over a hill in a peninsula west of Banda Aceh, north of Sumatra—corresponding to the height of the vegetation swept by the waves [33,34,139]. This represents the highest runup measured in history for a seismically generated tsunami, as this was until recently considered the cause of the mega-waves.

The excessive height—even considering the high magnitude of the earthquake (M 9.1), the third largest ever recorded, after the 1960 Chile and 1964 Alaska earthquakes—has been attributed to a slow tsunami earthquake, thus explaining the discrepancy between seismic and tsunami waves. However, the occurrence of the extremely high waves has also been linked to the occurrence of submarine landslides as part of the tsunami source, in addition to the major thrust fault movement. Evidence of submarine landslides has been observed in the source area [105,106].

Ref.: [33,34,35,106]

• 2007, April 21—Aysén Fjord, Chile
• Runup: 65 m
• Cause: Earthquake-triggered subaerial landslide (M 6.2)

The 2007 Aysén fjord earthquake triggered hundreds of landslides on the shores of the fjord, ranging in volume and type from rock avalanches to weathered rock and debris slides. Eyewitnesses took photographs of some of the landslides. The three largest, between 8 and 12 M m³, caused large waves when they violently entered the waters of the narrow fjord. The flooding reached hundreds of meters inland, killing 10 people. The maximum wave runup height of 65 m was estimated from the trimline of the vegetation washed away on the northeastern slope of Mentirosa Island, in front of which one of the largest rockslides occurred [116].

Other previous publications indicate waves runups up to “several tens of meters”, without specifying the exact height, and even just a few meters [13,23]. According to [12] the displaced waters left marks by direct impact of up to 50 m in the northern part of the Isla Mentirosa, based on a document from the Chilean National Mining Service. However, the data could not be verified as the document is not available online.

Ref.: [116]

• 2007, June 15—Shuibuya Reservoir, China
• Runup: 50 m
• Cause: Subaerial landslide

The tsunami occurred at the Shuibuya reservoir in the Qingjian River, located in the Three Gorges Region, following the progressive failure of a slope. A mass of 3 M m³ finally slid into the reservoir, generating a giant wave that reached a maximum height of 50 m on the opposite shore, with disastrous consequences [117]. The subsequent waves travelled downstream to the dam, 21 km away from the landslide, where they reached a runup height of up to 4 m.

The cause of the so-called Dayantang landslide may be related to the combined effects of rainfall and water level fluctuation, which may have altered the rock mass strength conditions in the submerges part of the slopes.

Other landslide-induced tsunamis have occurred previously and subsequently in the area, related to the unfavorable geological structure and hydrogeological conditions of the slopes.

Ref.: [117]

• 2007, November 5—Grijalva River, México
• Runup: 50 m
• Cause: Subaerial landslide

On the night of 4 November, there was a sudden landslide involving approximately 50 M m³ of rocks and soil on the right bank of the Grijalva River, of which about 15 M m³ fell into the river, generating a wave more than 50 m high, according to eyewitnesses [118].

The giant wave swept through the village of Juan de Grijalva, located immediately upstream of the landslide, killing more than 30 people and destroying its 100 houses. The village was literally wiped out by the wave, which then traveled tens of kilometers upstream and returned after hitting a dam, affecting other riverside villages.

The landslide, triggered by heavy rains (with exceptional precipitation in the days prior to the landslide exceeding historical records in the region), formed a dam 800 m long, 300 m wide, and 120 m high. A channel had to be dug to drain the floods upstream, as the river level rose more than 30 m in a few days, flooding hundreds of houses in some villages for several weeks and inundating almost a million hectares. After its destruction, it was determined that the town of Juan de Grijalva would not be rebuilt on the same site, as it was considered a high-risk area.

The Grijalva river tsunami is one of the few documented cases of a megatsunami on a river, with disastrous consequences affecting a nearby riverine population. Due to its magnitude, the Grijalva landslide is likely the most significant to have occurred in Mexico historically.

Ref.: [118]

• 2007, December 4—Chehalis Lake, Canada
• Runup: 38 m
• Cause: Subaerial landslide

A rock mass of 3 M m³ slid from a height of about 550 m on a slope in Chehalis Valley, in the southern Coast Mountains of British Columbia. The mass disintegrated into a debris avalanche, travelling approximately 800 m and reaching the northwestern shore of Chehalis Lake. Approximately, 1.5 M m³ entered the lake, generating a destructive tsunami that reached a height of up to 38 m on the opposite shore. The wave destroyed trees, stripped vegetation up to several tens of meters above the shoreline, and caused severe damage to roads and facilities in the area. The landslide was probably triggered by an intense rain-snow meteorological event.

Ref.: [50,119]

• 2011, March 11—Sanriku coast, Japan
• Runup: ~40 m
• Cause: Earthquake-triggered submarine landslide (M 9.1)

For the 2011 tsunami, one of the largest and most extensive in recent history, the two existing global tsunami databases record maximum runup heights of 55.88 m [12] and 42 m [13], and indicate a magnitude 9.1 earthquake as the cause—the largest earthquake ever recorded, along with the 2004 Indonesian earthquake, after the 1960 Chilean and the 1964 Alaskan earthquakes—. The figure of 55.88 m could not be verified, as it is not included in the reference documents cited in the global NCEI/WDS database, and appears to be erroneous, as the database itself classifies it as “doubtful runup” [12].

The data reported from post-tsunami field survey and all other reliable references, give maximum runup heights not exceeding 40 m on coasts near the epicenter, estuaries or narrow valleys that amplified the wave heights [139,140]. The data are similar to the historical record for a seismic tsunami in the same area: The Great Meiji Sanriku tsunami of 1896 [88,89].

The discrepancy between the magnitude of the earthquake and the tsunami runups along the Sanriku coast—much larger than expected—cannot be explained by a single earthquake tsunami mechanism and, as in the case of Indonesia in 2004, the large waves were attributed to a tsunami earthquake. Alternatively, to explain the wave heights and their directional characteristics, frequency and arrival times—not typical of seismic tsunamis—some authors have proposed a submarine landslide as the most likely additional tsunami source beside the major thrust fault movement [41,95,97]. Recently, bathymetric observations and geophysical evidence support the possibility of near-field tsunamis generated by large and rapid submarine landslides as an additional mechanism responsible for the highest waves [96]. Thus, the tsunami would have a dual origin: seismic (far-field) and submarine landslide, as occurred with the 1946 and 1964 events in Alaska [42,98,99].

Ref.: [89,95,96]

• 2014, July 21—Askja Lake, Iceland
• Runup: 80 m
• Cause: Subaerial rockslide

The megatsunami was triggered by a large rockslide on the inner wall of the Askja volcano caldera, in the northern volcanic zone of Iceland. The rockslide entered the lake at the bottom of the depression, which had a water depth of 200 m, causing a large tsunami that inundated the shore with waves reaching up to 60–80 m high in some places [51]. Snow patches around the caldera were covered by sediment deposited by the tsunami, marking the height reached by the waves.

The rockslide is one of the largest known since the settlement of Iceland, with a volume estimated from field surveys at about 20 M m³. It was registered as tremor at a seismic station near Askja.

Ref.: [51]

• 2015, October 17—Taan Fjord, Alaska
• Runup: 193 m
• Cause: Subaerial landslide

The megatsunami in Taan Fjord, Icy Bay, is the largest in Alaska since the 1958 Lituya event. It was caused by the rapid entry of a 180 million tons (~65 M m³) rock avalanche from the fjord wall into the sea, generating a violent wave that swept away about 20 km² of forest and land along the shores of the bay [38].

The maximum documented wave runup was 193 m, the fourth highest historically. Although there were no eyewitnesses, automatic seismic systems detected the waves within hours, identifying the source area and direction of the landslide. Satellite images later identified both the landslide and the tsunami impact areas, which were months later visited by scientists for field investigations to document the landslide and tsunami. Prior to the 2015 rockslide, field surveys and satellite measurements over several decades [141] had documented the unstable slope, which ultimately failed after a period of heavy rains.

Ref.: [38]

• 2017, June 17—Karrat Fjord, Greenland
• Runup: 90 m
• Cause: Subaerial landslide

A sudden landslide with a volume of approximately 45 M m³ [120], on the steep slopes of Karrat Fjord, located on the west coast of Greenland, triggered the largest tsunami ever recorded in Greenland. The wave exceeded 90 m in height near the landslide and reached almost 50 m on the opposite shore of the fjord.

The tsunami caused severe damage to the nearest village, located about 20 km away on an island in the fjord, destroying houses and killing four people.

The impact of the rockslide, which occurred from a height of about 1000 m, generated a seismic signal comparable to that of a magnitude 4.1 earthquake.

Ref.: [120,121]

• 2018, October 10—Jinsha River, Tibet, China
• Runup: 130–140 m
• Cause: Subaerial landslide

More than 15 M m³ of rock from the right bank slid into the Jinsha River, in Baiyu County, with a vertical drop of approximately 800 m. The landslide generated enormous waves on the opposite bank, destroying vegetation and scattering debris across the slope. The waves quickly dissipated both upstream and downstream. Based on the maximum elevation of the landslide-generated wave on the left bank (approximately 3040–3050 m above water level) and the initial average elevation of the river surface (2910 m), the wave reached a height of at least 130 m [122].

The landslide mass completely blocked the channel, forming an 80 m-high dam. Fortunately, there were no casualties, and the landslide dam naturally discharged after a few hours.

Ref.: [122]

• 2018, December 11—Bureya Reservoir, Russia
• Runup: 90 m
• Cause: Subaerial landslide

The megatsunami was caused by a massive landslide on the slope of the hydropower reservoir, which blocked the Bureya river and generated waves up to 90 m high on the opposite slopes, 2.8 km from the source. The estimated volume of the landslide is 25 M m³, and the depth of the reservoir up to 70 m.

The tsunami extended up to 10 km upstream and downstream of the impact zone. The waves penetrated inland up to 3.75 km, eventually reaching a height of 78 m above the initial reservoir level, leaving piles of uprooted and washed-out tree trunks several meters thick.

The waves washed away all vegetation and ground cover in the affected area. Subsequent field surveys revealed that the actual height of the waves was 5 to 10 m (sometimes up to 15 m) higher than the cut-off line of the swept vegetation marked on aerial photos and satellite images of the affected area [53].

Given the absence of seismicity in the area, it is likely that the landslide was caused or influenced by the formation of ice cover in the preceding days and the freezing of water in pre-existing tension cracks, or possibly affected by reservoir operations.

Ref.: [53]

• 2018, December 22—Anak Krakatau Island, Indonesia
• Runup: 85 m
• Cause: Volcanic flank landslide

This volcanic megatsunami is the most recent of any origin recorded in the two existing global databases [12,13] as of June 2024. It was caused by a flank landslide at the Anak Krakatau volcano in the Sunda Strait, generating waves with runups exceeding 80 m on the slopes of nearby small islands. Along with the Krakatoa megatsunami of 1883, these are the only two historical cases of megatsunamis associated with explosive volcanic island eruptions, occurring more than 135 years apart.

The small island of Anak Krakatau was built on the giant submarine caldera formed after the 1883 eruption. In 1927, after a period of 43 years of relative quiescence, new explosive eruptions began forming the new island, which reached a height of 335 m by 2018 as a result of continuous eruptive processes. Following the 2018 eruption, its elevation dropped to 110 m a.s.l.

The landslide of the southern flank of the island occurred after an eruptive period, involving a relatively small volume of less than 0.2 km³. The collapse swept most of the island, including its summit, into the sea, triggering the tsunami. On the rugged coasts of the closest small islands to the volcano, fortunately uninhabited, tsunami runup heights exceeded 85 m on the northern coast of Rakata Island, 83 m on the southern coast of Sertung [123], and 82.5 m on the west coast of Panjang island [124].

The tsunami reached the coastlines of Java and Sumatra—located 40–50 km away—in less than one hour, with waves up to 13.5 m high and maximum inundation distances exceeding 300 m. The waves caused 437 fatalities, marking the highest death toll from a volcanic tsunami since those triggered by the catastrophic eruption of Krakatoa in 1883 and the flank landslide of Ritter Island—the largest island volcano landslide in historical records—in Papua New Guinea in 1888, where maximum waves reached 15 m, claiming hundreds of lives.

The flank failure was likely influenced by over-steepening of the flank on the edge of the 1883 caldera, combined with the weakening of the deeper materials of Anak Krakatau due to alteration, and did not necessarily require an extraordinary eruptive event to trigger the landslide and subsequent megatsunami [142].

Ref.: [123,142]

• 2020, November 28—Elliot Lake, Canada
• Runup: 114 m
• Cause: Subaerial landslide

The most recent documented megatsunami (as of June 2024) occurred in the southern Coast Mountains of western Canada, triggered by a large rockslide. Approximately 18 M m³ of rock descended nearly 1000 m from the steep wall of a glacial valley, plunging into a 0.6 km² lake known as Elliot Lake, formed in the 20th century due to glacial retreat. The impact produced a runup exceeding 100 m high [125]. Water surged over the lake, travelling more than 10 km before depositing the debris at the lake’s outlet. The lake was partially filled by the deposits, forming several small islands of hummocky debris.

The waves left distinctive marks on the slopes of both sides of the lake, with a runup trimline reaching about 114 m above the lake level on the west side, and about 73 m on the east side. This event marked the largest tsunami ever recorded in the Canadian Mountain chain.

Ref.: [125]

11. Discussion

The GHMCat compiles and describes the megatsunami events that occurred in historical times, providing verified and cross-checked information about each of them. For this purpose, an objective criterion has been proposed to define and classify megatsunamis based on the analysis of maximum wave heights for all recorded historical events included in the GHTDs. The maximum wave height (Hmax) is the only available parameter that is measurable and universally applicable across all documented cases. A threshold of 35 m has been selected to classify megatsunamis based on a statistical analysis of maximum wave heights, with special attention given to the most significant tsunamis, regardless of their causes.

While the Hmax ≥35 m threshold appears robust based on the current dataset, it is important to acknowledge that this criterion may evolve as more data becomes available and our understanding of high tsunami dynamics advances. Therefore, further research and data collection are encouraged to validate and potentially adjust this threshold in the future.

11.1. Maximum Wave Height, Data Availability, and Measurement

The maximum wave height or runup is a physical parameter available for most documented historical events above a certain height, although it corresponds to different types of measurements and observations, from sometimes inaccurate or even doubtful witness account—in the cases of old tsunamis—to accurate post-tsunami measurements in recent cases. Instrumental measurements by tide or ocean gauges are scarce and available only for the last decades (with few exceptions), and therefore, these data cannot be used as a universal criterion for historical tsunamis of any source (Section 2.2).

Although a criterion based on wave height is the most widely used and has been proposed by various authors for the definition of a megatsunami, both the specific height threshold and the location where it should be measured (onshore or at the tsunami source) have not been consensually agreed upon so far [22]. When dealing with ancient documented historical tsunamis dating back to the 17th century, the available data are very limited; and generally, the maximum wave height reached on land, or runup, is the only parameter that can be used for comparison and classification.

Some definitions of megatsunami focus on wave height at the source, while others consider wave heights onshore (Section 2.1). However, the term “tsunami” intrinsically implies a wave that impacts the coast, and thus, the height is necessarily influenced by the nearshore bathymetry and the onshore topography and elevation (e.g., cliff coasts versus flat coasts). The proposal by Goff et al. [22], based on the initial wave height at the source, does not account for the influence of coastal topography. Additionally, two principal objections arise: (i) measurements of tsunami source wave heights are generally not available for the vast majority of large historical tsunamis, and (ii) only exceptionally extreme geological events would generate megatsunamis according to their proposed wave heights.

Wave height data at or near the source can only be obtained through deep oceanic sensors, which have been available for the last 50 years [24] and exclusively for open ocean tsunamis. Reliable data began to be collected in the early 1980s, with most current monitoring stations deployed after the 2004 Indonesian tsunami (Section 2.2). Therefore, a criterion based on wave heights at the source is largely theoretical and cannot be applied, even for the largest historical tsunamis.

Wave heights along coastlines from tide-gauge measurements are unavailable for most non-seismic tsunamis, with such data not existing for events prior to the mid-20th century, except in rare instances [24,25,26,30]. The first tide gauges began operating in the mid-19th century in the Pacific Ocean, with the network becoming sufficiently dense only by the mid-20th century. Additionally, tide gauge measurements may not always accurately reflect coastal conditions or provide a representative height of a tsunami, as evidenced by the 2004 tsunami in Indonesia [27].

Consequently, instrumental wave height measurements derived from deep ocean devices or coastal tide gauges cannot be used to define or classify historical megatsunamis occurring more than a few decades ago, especially for non-seismic local tsunamis triggered by subaerial landslides. In the majority of cases recorded in the GHTDs, data on the maximum height come from witness observations or post-tsunami survey measurements, with tide gauge measurements available only in a few cases.

11.2. Causes of Historical Megatsunamis

The analysis of the relationship between maximum wave heights and the causes of tsunamis—primarily earthquakes, 75%, followed by landslides, 14%—has shown that tsunamis of seismic origin rarely generate waves with a runup height exceeding 20 m, accounting for 2% of the events (less than 1% when considering all tsunamis recorded in the GHTDs). Above 30 m, the vast majority of tsunamis have been caused by subaerial or submarine landslides.

The highest recorded runup for a tsunami attributed to an earthquake, measuring 32 m, occurred in 1993 at Okushiri Island, Japan [100]. Other previous and subsequent tsunamis with Hmax ≥30 m, attributed to earthquakes in the GHTDs, have shown evidence of associated submarine landslides, which are proposed as the origin of these higher waves, thereby explaining inconsistencies with seismic magnitudes (Section 8.2). Evidence supported by research has demonstrated that the largest waves of tsunamis and megatsunamis such as those occurring in 1929 (Newfoundland, Canada), 1998 (Papua New Guinea), 1946, and 1964 (Alaska), initially attributed to earthquakes, were actually caused by submarine landslides [41,42,43,44,45].

Based on the analysis results, a value of 32 m for maximum wave height on land marks a significant change in the cause of the tsunamis (i.e., in their geological triggering mechanism) (see Section 8.3). From 32 m onwards, tsunamis have been generated by both subaerial and submarine landslides, often associated with earthquakes or, less frequently, volcanic eruptions. Large landslides accompanied by megatsunamis have occurred in both historical periods (as documented in the GHMCat) and prehistoric times: geological evidence from the Canary Islands and Hawaii has documented waves several hundred meters high, generated by the impact of large volumes of rock masses into the sea, hundreds of thousands of years ago [1,2,3].

Since the limit value for the height associated with seismic tsunamis (32 m) is below the threshold that defines megatsunamis (35 m), this indicates that no megatsunami has been generated by an earthquake.

Submarine landslides are recognized as a significant mechanism for generating megatsunamis associated with large underwater earthquakes. While tectonic dislocation explains the generation of far-field tsunamis, earthquake-triggered landslides produce near-field waves of exceptional height [43]. The interaction of these two types of tsunamis along coastlines near the source underscores their complex dynamics, with near-field waves typically arriving first.

The existence of submarine landslides has been substantiated in some cases through bathymetric surveys and geophysical investigations, as seen in the events mentioned above. For ancient events lacking direct evidence, submarine landslides are hypothesized as plausible mechanisms for megatsunami generation [41,84]. The volume of submarine landslides, the distance from the source and coastal topography determine megatsunami characteristics and the resulting wave height.

On the shore, the distribution of runup heights for nearfield tsunamis clearly reflects the occurrence of a different generating mechanism beyond seismic dislocation [31,40]. Analysis of runup data from post-tsunami field surveys and numerical simulations of various nearfield tsunamis shows that a local concentration of exceptional runups is inconsistent with a seismic cause, necessitating consideration of an alternative source, such as a landslide [40]. This is especially illustrative in the cases of the 1946 Unimak Island tsunami and the 1998 Papua New Guinea tsunami.

The main and most effective method for determining the direct cause of megatsunami waves associated with large earthquakes is through underwater investigations to identify co-seismic landslides. Although not direct evidence, the analysis of the spatial distribution of the highest waves (maximum runups) along the affected coast, through detailed post-tsunami field surveys, can suggest the probable existence of submarine landslides.

Regarding the 32 m limit for seismic tsunamis, future investigations might confirm the occurrence of a submarine landslide related to the 1993 earthquake. Additionally, it could be possible that a large future earthquake generates waves exceeding this threshold that are clearly not associated with a submarine landslide, in which case the proposed value would need to be reconsidered.

To advance the understanding of earthquake-triggered submarine landslides as a megatsunamis-generating mechanism, it is crucial to conduct submarine research and develop models based on validated data. This knowledge is essential for assessing megatsunami risks and implementing effective prevention measures in vulnerable coastal regions worldwide.

11.3. GHMCat Temporal and Spatial Scope

The historical record of megatsunamis begins in the 17th century, excluding the Santorini event. Until the mid-19th century, the data are scarce, incomplete and uncertain, with only three recorded megatsunamis during this period. From 1850 onwards, there is a notable increase in recorded events, with 16 documented throughout the 20th century. This number gradually rises over time, and in the 21st century, there has been a surge, with 15 documented megatsunamis since 2000.

As might be expected, significant spatial and temporal biases are present, reflected in the frequency and distribution of events, conditioned primarily by the presence of human settlements and the availability of written historical records. Over the past 90 years, the record of megatsunamis has increased, with several of the most significant occurring in glaciated areas, where half of the events with wave heights exceeding 70 m have taken place (Table 2). Particularly noteworthy is the fact that the bays and fjords of Alaska and Norway have experienced more than a third of all recorded megatsunamis worldwide over the past 125 years, primarily during the first half of the 20th century.

Besides the inherent spatial biases in the global historical record, regional clusters of megatsunamis may reflect the high number of landslide investigations in these areas and the availability of documentation [143]. On the other hand, the role of geological and climatic conditions cannot be dismissed, as they influence the frequency and occurrence of megatsunamis in these regions and other glaciated regions prone to large rockslide-triggered events. In contrast, Japan, with a historical record of tsunamis spanning many hundreds of years, has documented only three megatsunamis. Similarly, only one such event has been recorded in Russia—on its far east coast—in almost 300 years.

The cases of Lituya Bay (Alaska) and Lake Lovatnet (Norway) are renowned for their history of megatsunamis, with each location experiencing at least three events within a little over a century. Additionally, these areas have also experienced numerous tsunamis with smaller yet still significant wave heights, consistently originating from rockslides or rock avalanches [7,9]. Similarly, in the regions of Vaigat Strait and Karrat Fjord (central West Greenland), recent megatsunamis occurred in 2000 and 2017, respectively, with several other landslide-generated tsunamis occurring in recent decades [144]. These facts suggest the possibility that megatsunamis could occur again in the same or nearby locations.

11.4. Future Trends

In glaciated areas, the combined effects of steep slopes, fractured rock masses, glacial retreat, high precipitation, and freeze–thaw cycles play a crucial role in the occurrence of large rockslides and rock avalanches. Global warming, which leads to rising temperatures, significantly impacts the colder high latitudes of both hemispheres, particularly in frozen regions, where rapid deglaciation results in the gradual retreat of glaciers, significantly influencing slope instability processes. Climate warming is likely driving an increase in the frequency of large landslides [141].

Following glacier retreat, steep slopes are exposed to erosion and weathering processes, as well as changes in the in situ state of stress of the rock mass. These changes in the stress regime modify the equilibrium conditions of slopes, potentially leading to large rockslides and avalanches that, when suddenly falling into confined deep-water bodies such as fjords and bays, may trigger massive megatsunami waves.

The close link between rising temperatures and the increasing occurrence of large waves in confined waters in glacial coastal areas vulnerable to slope instability seems evident. It is important to direct attention towards preventing and mitigating potential damaging effects in these regions prone to large landslides and associated megatsunamis.

12. Conclusions

• The Global Historical Megatsunami Catalog (GHMCat) compiles the events with the largest waves recorded in historical times. It provides a comprehensive list of 40 verified megatsunamis, detailing their maximum wave heights, causes and primary bibliographic sources. It also describes the main characteristics, attributes, and consequences or damages of each megatsunami. Additionally, a previously unrecorded megatsunami that occurred before 1786 on the coast of Alaska has been documented.
• A definition of megatsunami is proposed based on the objective criterion of maximum height reached by the waves, or runup, with a proposed threshold value of 35 m, derived from the analysis of all historical tsunamis, particularly those with a maximum wave height (Hmax) ≥30 m. The 35 m threshold effectively distinguishes an exclusive group of 40 events, that represent ~1.5% of documented historical tsunamis.
• No tsunami waves caused by earthquakes have been recorded over 32 m. In contrast, all tsunamis exceeding this value have been generated by subaerial or submarine landslides.
• Large subaerial landslides or rock avalanches, occasionally triggered by high-magnitude earthquakes or large explosive volcanic eruptions, account for 80% of megatsunamis, while 20% of the events have been caused by large submarine landslides triggered by very high magnitude earthquakes.
• Megatsunamis generated by subaerial landslides or rock avalanches in confined bodies of water yield the highest recorded runups, reaching up to several hundred meters. The highest is the 1958 Lituya Bay megatsunami, with a runup of 525 m, more than double that of the second highest, Spirit Lake, 1980.
• Submarine landslides triggered by great earthquakes represent a critical mechanism for generating near-field megatsunamis. This dual earthquake-landslide mechanism helps explain the exceptional tsunami wave heights independently of earthquake magnitudes.
• Historical megatsunamis have been documented in America (~40%), Asia (~32%), and Europe (~22%). Alaska and Norway’s bays and fjords have the highest frequency, accounting for 40% of global recorded megatsunamis in the last 350 years. Other affected areas include the coasts of Indonesia, Japan, Canada, and China, with the latter experiencing four megatsunamis in rivers or reservoirs.
• Megatsunamis have occurred in glaciated regions’ bays, fjords and lakes (45%), open sea coasts (25%), mountain lakes (12%), rivers (10%), and reservoirs (8%). Notably, human activity has influenced landslides in certain instances, such as reservoirs in China and Italy.
• The possibility of more frequent megatsunamis in glaciated regions due to global warming-induced retreat warrants consideration. In contrast, the likelihood of megatsunamis associated with large explosive eruptions or volcanic island flank failures is very low, as is the occurrence of local, near-field megatsunamis generated by large-magnitude earthquake-triggered submarine landslides.
• The information provided by the GHMCat allows for a comprehensive historical overview of megatsunamis, establishing relationships between their causes, wave heights, and geographic distribution over the past 350 years. This may contribute to advancing the knowledge and understanding of the causes and origins of megatsunamis, and aid prevention efforts in high-risk regions.