Tsunamis occur wherever large volumes of water get suddenly displaced, but the overwhelming majority strike coastlines that border subduction zones, where one tectonic plate dives beneath another. This makes the Pacific Ocean basin, ringed by subduction trenches, the single most tsunami-prone region on Earth. Countries along the so-called Ring of Fire, from Chile and Peru to Japan, Indonesia, and the Philippines, face the highest frequency of events. Yet the Indian Ocean, the Mediterranean, the Caribbean, and even parts of the Atlantic coast carry real and sometimes underestimated risk, driven by a mix of tectonics, volcanic activity, and underwater landslides.
Why Subduction Zones Drive Most Tsunamis
The engine behind the vast majority of destructive tsunamis is the megathrust fault, the boundary where an oceanic plate slides beneath a continental or overriding plate. Stress builds along this boundary for decades or centuries, and when it releases, the seafloor can lurch upward by several meters in seconds, shoving the entire water column above it into motion. The shallowest portions of these faults are the most dangerous for tsunami generation because they sit directly under the ocean floor and produce the largest vertical displacements.
Earthquakes that rupture these shallow megathrust segments can be deceptively moderate in magnitude while still creating enormous waves, a class of events seismologists call “tsunami earthquakes.”1Geophysical Research Letters. Weak Near‐Field Behavior of a Tsunami Earthquake: Toward Real‐Time Identification for Local Warning Because the rupture happens so close to the seafloor, even relatively modest seismic energy translates into outsized ocean displacement. Understanding how and when these shallow fault segments slip remains one of the harder problems in geophysics, partly because making direct observations on the ocean floor is expensive and logistically brutal.2PubMed Central. Rapid shallow megathrust afterslip from the 2021 M8.2 Chignik, Alaska earthquake revealed by seafloor geodesy
The Pacific Ring of Fire
If you drew a horseshoe around the Pacific Ocean, tracing from New Zealand up through the Philippines, Japan, the Aleutian Islands, down the west coast of North and South America, and ending near Antarctica, you would outline the most tsunami-active region on the planet. Nearly every country along this arc has a documented history of destructive waves, and several have been hit repeatedly over relatively short timescales.
Japan sits in a uniquely precarious spot. It borders multiple subduction zones, and geological records from its coastlines show tsunami deposits going back thousands of years. At one excavation site in northeastern Japan, researchers identified eleven distinct tsunami deposit layers formed over the past 4,000 years.3Quaternary Science Reviews. Tsunami deposits and recurrence on a typhoon-prone coast of northern Taiwan from the last millennium The Ryukyu Islands in Japan’s subtropical south carry evidence of a magnitude-9 class megathrust earthquake around 0 B.C. that coincided with the historically recorded Sakishima tsunami.4PubMed Central. Evidence of megathrust earthquakes and seismic supercycles in subtropical Japan from millennia-old coral microatolls The 2011 Tōhoku event, which killed roughly 18,000 people, was only the latest entry in a very long catalog.
Alaska and the Aleutian Islands are another hotspot. Modeling of stratigraphic records along the eastern Aleutians has reconstructed a 700-year sequence of great earthquakes, with one 18th-century event producing tsunami deposits at elevations of 23 meters above mean sea level.5PubMed Central. A 700-year rupture sequence of great eastern Aleutian earthquakes from tsunami modeling of stratigraphic records Some of these events sent damaging waves all the way to Hawaiʻi, while others, depending on rupture geometry, barely registered there despite being enormous locally. That variability matters for warning systems: a huge quake in the Aleutians does not automatically mean a huge wave in Honolulu.
Chile and Peru round out the South American side of the Ring of Fire. Chile’s subduction zone produced the largest instrumentally recorded earthquake in history, the 1960 magnitude-9.5 event, which sent a transoceanic tsunami that killed people as far away as Japan and the Philippines. Taiwan, the Philippines, and New Zealand all sit on active plate boundaries capable of generating locally devastating tsunamis as well. Geological work on the northern coast of Taiwan has identified four distinct tsunami event layers from the last millennium, with a recurrence interval estimated between about 83 and 436 years.3Quaternary Science Reviews. Tsunami deposits and recurrence on a typhoon-prone coast of northern Taiwan from the last millennium
The Indian Ocean and the Sunda Megathrust
Before December 26, 2004, many people outside the geosciences had never heard the word “megathrust.” The Indian Ocean tsunami that day, generated by a sudden slip along roughly 1,600 kilometers of the Sunda megathrust between Aceh and Myanmar, killed over 230,000 people across fourteen countries.6Journal of Earthquake and Tsunami. THE SUNDA MEGATHRUST — PAST, PRESENT AND FUTURE Three months later, another great earthquake ruptured 350 kilometers of the same fault system farther south, beneath Simeulue and Nias islands, producing a smaller but still destructive tsunami.
The countries most affected by Indian Ocean tsunamis cluster around the Sunda megathrust: Indonesia, Sri Lanka, India (especially the Tamil Nadu and Andaman coasts), Thailand, Myanmar, and the Maldives. Indonesia is by far the most exposed. It straddles the fault itself and has thousands of populated coastal communities within minutes of potential wave arrival. Researchers studying the section of the megathrust south of Java have identified seismic gaps, long quiet stretches of the fault where stress has been accumulating, that could host future great earthquakes. The worst-case modeling for the Java segment projects an earthquake as large as magnitude 9.1 and tsunami heights exceeding 20 meters near small islands south of Java’s westernmost province.7PubMed Central. Implications for megathrust earthquakes and tsunamis from seismic gaps south of Java Indonesia Meanwhile, the stretch offshore of West Sumatra and Bengkulu provinces remains a concern: researchers have warned that another great earthquake and tsunami there is likely within the coming decades.6Journal of Earthquake and Tsunami. THE SUNDA MEGATHRUST — PAST, PRESENT AND FUTURE
The Indian Ocean had no coordinated tsunami warning system before 2004. One now exists, but the warning time for communities nearest the fault can be agonizingly short, sometimes only fifteen to twenty minutes. For Indonesia in particular, local tsunamis remain the primary threat because the waves arrive before most centralized alerts can reach coastal populations.
The Mediterranean and the Caribbean
People tend to associate tsunamis exclusively with the Pacific and Indian Oceans, but the Mediterranean has a documented tsunami history stretching back thousands of years. The 365 CE Crete earthquake generated a wave powerful enough to cause massive seafloor sediment disturbances along the Italian and African continental margins as it traveled westward.8PubMed Central. Recognizing megatsunamis in Mediterranean deep sea sediments based on the massive deposits of the 365 CE Crete event Today, the eastern Mediterranean carries the higher tsunami hazard, driven by the Hellenic subduction zone south of Crete and the Cyprus Arc, while the western Mediterranean’s risk comes more from local crustal faults.9PubMed Central. Tsunamigenic potential of crustal faults and subduction zones in the Mediterranean Countries most at risk include Greece, Turkey, Italy (particularly Sicily and the Calabrian coast), Egypt, and Libya. The 1908 Messina earthquake and tsunami killed tens of thousands along the strait between Sicily and the Italian mainland.
In the Atlantic and Caribbean, the risk profile is different but real. The Puerto Rico Trench, the deepest point in the Atlantic, sits north of the Caribbean island arc. Seafloor mapping there has revealed that the tsunami hazard to the northern coasts of Puerto Rico and the U.S. Virgin Islands from submarine slope failures is high.10Eos Transactions American Geophysical Union. New seafloor map of the Puerto Rico trench helps assess earthquake and tsunami hazards A 1918 earthquake off northwestern Puerto Rico produced a local tsunami that killed over 100 people. The broader Caribbean has experienced at least a dozen tsunami events since European colonization, and the combination of tectonic activity and underwater volcanic slopes keeps the hazard level elevated for islands like Hispaniola, Martinique, and the Lesser Antilles.
Portugal and Morocco face an Atlantic tsunami threat from the Azores-Gibraltar fault zone, the same system that produced the catastrophic 1755 Lisbon earthquake and tsunami. That event killed tens of thousands and was felt across the entire eastern Atlantic basin.
Tsunamis Without Earthquakes
Subduction zone quakes get most of the attention, and rightly so, but tsunamis can also be triggered by volcanic eruptions, submarine landslides, and even atmospheric disturbances. These non-seismic sources complicate risk assessment because they do not always show up on the same hazard maps.
The January 2022 eruption of the Hunga Tonga-Hunga Haʻapai volcano produced one of the largest volcanic tsunamis ever recorded. The eruption generated waves through multiple mechanisms: explosions, volcanic mass flows, and sudden caldera subsidence. Researchers found that the largest tsunami was caused not by the initial explosion, which was globally detected, but by a caldera-subsidence event that occurred about an hour later.11PubMed Central. Delayed submarine caldera subsidence creates extreme tsunami hazard This made timing unpredictable. The event also generated atmospheric pressure waves that traveled faster than the ocean waves themselves, producing secondary tsunami-like disturbances on coastlines thousands of kilometers from Tonga, including in the Caribbean and the Mediterranean.12PubMed Central. Global Tonga tsunami explained by a fast-moving atmospheric source
Submarine landslides are another underappreciated source. Along the eastern margin of the Sea of Japan, researchers have identified rotational submarine landslides in shallow continental-shelf waters that could generate initial wave amplitudes exceeding 10 meters.13Journal of Geophysical Research: Oceans. Tsunami Potential and Scaling Characteristics of Rotational Submarine Landslides in the Eastern Margin of the Sea of Japan Because the locations and sizes of future landslides are difficult to predict, they represent a stubborn source of uncertainty in hazard planning. Landslide-generated tsunamis are typically more localized than earthquake-generated ones, but they can be extremely intense near the source. Some of history’s tallest documented wave run-ups, including the 1958 Lituya Bay event in Alaska, were caused by landslides.
Then there are meteotsunamis, tsunami-like waves driven by atmospheric pressure disturbances rather than geological events. In June 2013, a series of mesoscale convective weather systems moving from inland to offshore along the U.S. East Coast generated shallow water waves that were amplified by resonance between the water column and the atmospheric forcing.14PubMed Central. Mesoscale convective system surface pressure anomalies responsible for meteotsunamis along the U.S. East Coast on June 13th, 2013 Meteotsunamis are usually smaller than their seismic counterparts, but they can catch coastal communities off guard because they are not preceded by any earthquake. They have been recorded in the Mediterranean (particularly the Adriatic and the Balearic Islands), the Great Lakes, and parts of East Asia.
How Sea-Level Rise Changes the Map
Tsunami risk is not static. Rising sea levels are poised to amplify both the frequency and intensity of tsunami-induced flooding, even without any change in the underlying seismic hazard. A study modeling Macau’s coastline found that a conservative sea-level rise of half a meter, potentially reached by around 2060, would roughly double the frequency of tsunami-induced flooding events. A one-meter rise, plausible by 2100, could increase flooding frequency by a factor of roughly 1.5 to nearly 5, depending on the scenario.15PubMed Central. A modest 0.5-m rise in sea level will double the tsunami hazard in Macau
The mechanism is straightforward. Higher baseline water levels mean that a tsunami of any given size starts from a higher platform. Waves that would previously have been absorbed by beaches, seawalls, or gently sloping coastal land can now overtop defenses and push further inland. For low-lying island nations like the Maldives, Tuvalu, and parts of Indonesia and the Philippines, this is a compounding threat. These places already face existential risks from rising seas and storm surges; adding an amplified tsunami hazard on top of that worsens an already difficult adaptation picture.
This interaction between sea-level rise and tsunami hazard is often missing from coastal planning. Many hazard maps are drawn using current sea levels and then treated as fixed. Updated probabilistic assessments that incorporate future sea-level scenarios are being developed for certain coastlines, including southeastern China and Taiwan.16Journal of Geophysical Research: Solid Earth. Probabilistic Tsunami Hazard Assessment (PTHA) for Southeast Coast of Chinese Mainland and Taiwan Island But many of the world’s most vulnerable communities still lack these forward-looking models.
Why Some Countries Suffer More Than Others
Geophysical exposure tells you where tsunamis can happen. It does not tell you where they do the most damage. The human toll of a tsunami depends enormously on socioeconomic factors: income levels, housing quality, infrastructure resilience, and warning-system reach. The 2004 Indian Ocean tsunami illustrated this starkly. Across the affected countries, wealthier communities with sturdier buildings and higher ground nearby suffered far fewer casualties per capita than poor fishing villages built right at the waterline.
In Galle, Sri Lanka, many residents work in fishing and tourism and live in poorly constructed or informal houses that sit within what local authorities designate as a coastal buffer zone. Low income limits both the ability to build resilient structures and the financial capacity to recover after a disaster, creating a cycle where each event deepens economic vulnerability.17IntechOpen. Socio-Economic Vulnerability in Tsunami Devastations: A Critical Review Based on Galle City, Sri Lanka Critical infrastructure damage can trigger a long-term economic downturn because the very industries people depend on, fishing boats, tourism facilities, and coastal roads, are the first things destroyed.
Post-2004 studies in Malaysia found that household recovery depended heavily on pre-existing income. Families in the highest income bracket recovered faster, while those headed by unemployed individuals or those who lost their fishing boats had significantly reduced odds of regaining their previous economic status.18PubMed. Socioeconomic consequences of the 2004 tsunami: policy implications for natural disaster management In practical terms, a tsunami of identical size hitting two coastlines at the same angle can produce wildly different outcomes depending on how much money and institutional capacity the affected population has.
This matters for the “most at risk” question. A country like Japan, which faces extremely high seismic exposure, has invested billions in seawalls, vertical evacuation towers, building codes, and a sophisticated warning network. By contrast, countries like Myanmar, Bangladesh, or Mozambique have comparable or even lower seismic exposure but far less capacity to absorb a strike. The effective risk, combining probability of impact with likely consequences, puts many developing nations higher on the danger list than raw geophysics alone would suggest.
Probabilistic Hazard Assessment and Its Limits
Governments and international bodies increasingly rely on probabilistic tsunami hazard assessment, or PTHA, to guide coastal planning. These models combine estimated earthquake source parameters, wave propagation simulations, and local coastal topography to estimate how likely it is that a given stretch of coast will experience flooding of a particular depth within a given time window. Modern frameworks can incorporate multiple earthquake sources, both near and distant, and account for variables like tidal state at the time of arrival.19Mathematics of Planet Earth. R4–Event-Tree PTHA of Coastal Inundation with Tidal Uncertainty Treatment
The challenge is that these models are only as good as our knowledge of what can happen. Submarine landslides, volcanic collapses, and unusual fault geometries can all produce tsunamis that fall outside the scenarios the models consider. The 2011 Tōhoku earthquake exceeded what many Japanese hazard assessments had considered a worst-case scenario. The 2022 Hunga eruption generated atmospheric waves that reached coastlines no conventional tsunami model would have flagged. These surprises keep researchers honest about the limits of forecasting.
For individual communities, the practical question often is not “what is the probability?” but “what do we do in the next ten minutes?” Near-source tsunamis, the kind that hit within fifteen to thirty minutes of the triggering event, leave almost no time for formal warning dissemination. In these cases, the most effective mitigation is public education: if you feel strong shaking near the coast, move to high ground immediately without waiting for an official alert. Countries like Japan, Chile, and Indonesia have invested heavily in this kind of awareness. Whether those lessons have truly reached the most vulnerable coastal populations remains an open question.
Regions That Fly Under the Radar
Some coastlines with genuine tsunami exposure receive relatively little attention. Southeastern China, for instance, faces potential waves from the Manila Trench and the Ryukyu Trench, yet tsunami preparedness in the region has historically lagged behind that of nearby Japan and Taiwan.16Journal of Geophysical Research: Solid Earth. Probabilistic Tsunami Hazard Assessment (PTHA) for Southeast Coast of Chinese Mainland and Taiwan Island The eastern coast of Africa, from Mozambique to Tanzania and Kenya, is within range of Indian Ocean events and has limited early-warning infrastructure. Parts of the east coast of the United States and the Gulf Coast face meteotsunami risk and, to a lesser extent, far-field tsunami risk from Canary Islands volcanic flank collapse scenarios that have received sporadic but dramatic media coverage.
The Sea of Japan presents another under-discussed zone. Active submarine faults and thick sediment layers along its eastern margin create conditions for landslide-generated tsunamis that could strike Japanese and Korean coastlines with little warning.13Journal of Geophysical Research: Oceans. Tsunami Potential and Scaling Characteristics of Rotational Submarine Landslides in the Eastern Margin of the Sea of Japan Because these events are triggered by slope failures rather than great earthquakes, they may not produce the kind of prolonged shaking that serves as a natural warning.
Even inland freshwater bodies are not entirely immune. Earthquake-triggered seiches in large lakes have produced wave surges in the past, and a major landslide into a reservoir can generate a localized wave. These are not oceanic tsunamis by strict definition, but the physical process and the danger to shoreline communities are analogous. Awareness of these possibilities tends to be even lower than awareness of oceanic tsunami risk, simply because people do not think of lakes as tsunami-capable environments.