1906 에콰도르-콜롬비아 지진: 북안데스 위험을 정의한 M8.8 메가스러스트

January 31, 1906: The Northern Nazca Megathrust Breaks

The morning of January 31, 1906 began quietly along the Colombian and Ecuadorian Pacific coast — a remote, sparsely populated coast of tropical lowlands, river deltas, and small fishing communities that had seen little of the economic development transforming South America's interior cities. At approximately 10:36 AM local time, the ground shook with a violence that no living person in the region had experienced.

The earthquake lasted several minutes. Contemporary accounts from the coastal town of Tumaco in Colombia and from Esmeraldas in Ecuador describe total destruction of adobe and timber buildings. The shaking was felt throughout Colombia, Ecuador, and across the border into Peru. In Quito, more than 400 kilometres inland and across a major mountain range, the shaking was strong enough to crack walls and topple chimneys. In Bogotá, hundreds of kilometres further north, residents felt a prolonged rolling motion.

Modern analysis, using the tsunami records that the earthquake generated at stations around the Pacific, places the MagnitudeA single number that quantifies the total energy released by an earthquake. Each whole number increase represents roughly 31.6 times more energy released. of the 1906 Ecuador-Colombia earthquake at approximately 8.8 — making it one of the five largest earthquakes of the 20th century and one of the largest instrumentally recorded events at the time. The seismograph had been invented only two decades earlier, and the networks of the early 1900s were sparse; the few stations that recorded the event provided limited but real instrumental evidence of the earthquake's extraordinary size.

Use Earthquake Energy Calculator to understand the enormous energy release of a M8.8 earthquake and how it compares to other events in the historical catalogue.

Colombia-Ecuador Subduction: 500 km of Coupled Fault

The geological setting of the 1906 earthquake is the Subduction ZoneA region where one tectonic plate dives beneath another into the mantle. Subduction zones produce the world's largest earthquakes (M8.5+) and are associated with deep ocean trenches and volcanic arcs. where the Nazca Plate descends beneath the South American Plate along the Colombia-Ecuador segment of the Convergent BoundaryA plate boundary where two plates move toward each other. Can produce subduction zones (ocean-continent), mountain building (continent-continent), or deep trenches (ocean-ocean).. In this region, the Nazca Plate is moving eastward at approximately six centimetres per year, diving beneath the western edge of South America. The plate interface extends from near the surface at the trench to depths of approximately 60 to 80 kilometres beneath the Andes.

The coupling between the Nazca and South American plates along this segment is high — the plates tend to be locked together rather than sliding smoothly. This locking allows stress to accumulate continuously, building the elastic strain energy that is ultimately released in large Fault RuptureThe breakage of rock along a fault during an earthquake, releasing stored elastic energy as seismic waves. Rupture length can range from meters (small quakes) to 1,000+ km (great earthquakes). events. The 1906 earthquake broke approximately 500 kilometres of this locked interface, from near the Ecuador-Peru border in the south to approximately the latitude of the Colombian port of Buenaventura in the north.

The specific geometry of the Fault RuptureThe breakage of rock along a fault during an earthquake, releasing stored elastic energy as seismic waves. Rupture length can range from meters (small quakes) to 1,000+ km (great earthquakes). in 1906 has been studied by comparing the spatial distribution of TsunamiA series of ocean waves generated by sudden displacement of the seafloor during an underwater earthquake. Tsunamis can travel across entire ocean basins at jet speed (700+ km/h). run-up heights at stations around the Pacific with those from subsequent earthquakes in the same region. The conclusion from these studies is that the 1906 rupture was unusually large and uniform — a continuous break along the full 500-kilometre extent of the segment — and that the seafloor deformation it produced was sufficient to generate a substantial TsunamiA series of ocean waves generated by sudden displacement of the seafloor during an underwater earthquake. Tsunamis can travel across entire ocean basins at jet speed (700+ km/h). across the entire Pacific basin.

The region's geological complexity adds additional hazard factors. The Carnegie Ridge — a submarine volcanic feature associated with the Galapagos hotspot — is currently being subducted beneath the Ecuador segment of the trench. This ridge subduction is thought to influence the coupling pattern and segmentation of the megathrust, potentially creating patches of higher locking and thus greater Seismic MomentA measure of the total energy released by an earthquake, calculated as the product of the fault area, average displacement, and the shear modulus of the rocks. The basis of moment magnitude. potential between larger ruptures. The interaction of normal Subduction ZoneA region where one tectonic plate dives beneath another into the mantle. Subduction zones produce the world's largest earthquakes (M8.5+) and are associated with deep ocean trenches and volcanic arcs. mechanics with the anomalous bathymetry of the Carnegie Ridge makes the Ecuador segment one of the most geologically distinctive sections of the South American megathrust.

The seismological community took decades to fully characterize the 1906 event because of the limitations of the early instrumental record. The earthquake predates the widespread deployment of modern broadband seismographs by more than half a century. Reanalysis of the original seismogram records from European stations — using digital scanning and modern waveform modelling — has progressively refined the magnitude estimate and source parameters. The current consensus of approximately M8.8 is based on the convergence of tsunami modelling, historical intensity reports, and instrumental reanalysis, but small uncertainties remain.

Tsunami Impact on Coastal Communities

Within minutes of the earthquake, a large TsunamiA series of ocean waves generated by sudden displacement of the seafloor during an underwater earthquake. Tsunamis can travel across entire ocean basins at jet speed (700+ km/h). began inundating the Colombian and Ecuadorian coastlines. Contemporary accounts from Tumaco describe waves that overwhelmed the low-lying town, sweeping away entire neighbourhoods and killing hundreds. The towns of Salinas and Esmeraldas in Ecuador experienced similar inundation. Total direct casualties from the earthquake and tsunami are estimated at 500 to 1,500, though the remote character of the affected coast and the limited communications of the era make precise figures impossible to determine.

The low death toll relative to the magnitude of the event reflects the sparseness of the 1906 coastal population. Tumaco, the largest Colombian Pacific city affected, had only a few thousand inhabitants. The river delta and coastal mangrove environments that dominate this coast were inhabited primarily by small fishing communities that in some cases were entirely swept away but whose absolute numbers were limited. The same tsunami striking the same coastline today would encounter a vastly larger population.

The 1906 TsunamiA series of ocean waves generated by sudden displacement of the seafloor during an underwater earthquake. Tsunamis can travel across entire ocean basins at jet speed (700+ km/h). generated significant waves throughout the Pacific. At Hawaii, instruments and observer accounts recorded wave amplitudes of approximately one metre, arriving approximately twelve hours after the earthquake. On the Japanese coast, small but measurable waves arrived some twenty hours after the event. The far-field tsunami data from the 1906 event — though limited by the sparse instrumentation of the era — has provided one of the key constraints in reconstructing the source parameters of the earthquake through inverse modelling.

Tsunami inundation modelling for the Colombia-Ecuador coast consistently identifies the 1906 Fault RuptureThe breakage of rock along a fault during an earthquake, releasing stored elastic energy as seismic waves. Rupture length can range from meters (small quakes) to 1,000+ km (great earthquakes). scenario as the design event for regional tsunami preparedness planning. The near-field character of the source — the EpicenterThe point on the Earth's surface directly above the hypocenter (focus) where an earthquake originates underground. Often reported as the earthquake's location in news reports. is close to the most exposed coastlines — means that communities like Tumaco have warning times measured in minutes rather than hours. This reality drives the "earthquake equals warning" philosophy: coastal residents are trained to begin evacuating to high ground immediately upon feeling strong shaking, without waiting for any official warning.

The 1906 event also highlighted the vulnerability of the Colombia-Ecuador coast to earthquake-triggered Secondary Earthquake HazardsHazards triggered by earthquake shaking rather than the shaking itself — including tsunamis, landslides, liquefaction, fires, dam failures, and chemical releases. Often cause more damage than shaking. beyond the direct tsunami. Coastal subsidence — the permanent downward displacement of the land surface that accompanies the seaward motion of the overriding plate in a megathrust rupture — lowered coastal land elevations by up to one to two metres in some areas. This subsidence left communities more exposed to subsequent high tides and storm surges even after the tsunami itself had receded, and is now recognized as a standard component of megathrust hazard assessment.

Segmented Re-Rupture: 1942, 1958, 1979 Partial Repeats

One of the most scientifically fascinating aspects of the Colombia-Ecuador subduction zone is its behaviour in the decades after the 1906 megathrust rupture. Rather than simply rebuilding stress uniformly along the entire 500-kilometre rupture length, the zone re-ruptured in segments — smaller earthquakes breaking portions of the 1906 area in what seismologists call a "segmented re-rupture" pattern.

Three major events contributed to this sequence. A M7.9 earthquake in 1942 broke the southern portion of the 1906 rupture zone, offshore Ecuador. A M7.7 earthquake in 1958 ruptured the central portion, in the border region between Ecuador and Colombia. And a M8.1 earthquake in 1979 broke the northern portion, offshore Colombia. Together, these three events re-ruptured much of the 1906 area, but individually none approached the size of the original Fault RuptureThe breakage of rock along a fault during an earthquake, releasing stored elastic energy as seismic waves. Rupture length can range from meters (small quakes) to 1,000+ km (great earthquakes)..

This segmented re-rupture behavior provides important insights into the mechanics of Tectonic PlateA massive segment of Earth's lithosphere that moves, floats, and sometimes fractures. There are 7 major and about 8 minor plates, and their interactions cause most earthquakes. interaction and Seismic GapA section of an active fault that has not produced an earthquake for a long time compared to neighboring sections. Seismic gaps may indicate increased probability of a future earthquake. dynamics. It suggests that the 1906 rupture — by releasing stress over a very long fault length in a single event — set up a pattern where different segments of the fault subsequently recovered their locking at different rates, leading to asynchronous re-rupture rather than synchronous re-rupture of the entire segment.

The Earthquake Recurrence IntervalThe average time between major earthquakes on a particular fault. Estimated from paleoseismology and historical records. The Cascadia subduction zone has a recurrence interval of ~500 years. question is complicated by this segmented behavior. Is the relevant recurrence for this zone the interval between 1906-scale M8.8+ events, or the shorter interval between the partial re-ruptures? The answer determines how urgently the zone should be considered at risk for another large earthquake, and what the hazard assessment should assume about the next event's size. If the 1906 full-segment rupture recurs approximately every 200 to 400 years, then the period from 1906 to the present represents roughly the first third to half of the inter-event cycle, suggesting the zone may not be close to another full-segment failure. But if individual segments can produce M7.5-8.1 events on shorter recurrence intervals, the hazard to coastal communities may be more immediate.

The 2016 Pedernales M7.8: Is the Cycle Repeating?

On April 16, 2016 — coincidentally the same day as the M7.3 Kumamoto MainshockThe largest earthquake in a sequence, which defines the overall magnitude of the event. Preceded by foreshocks (sometimes) and followed by aftershocks (always). in Japan — a magnitude 7.8 earthquake struck offshore Ecuador, near the town of Pedernales on the coast of Manabi Province. The earthquake killed 670 people, injured more than 6,000, and caused widespread destruction along the Ecuadorian coast in an area that had seen minimal new construction investment since the 1942 partial re-rupture.

The 2016 Pedernales earthquake broke a portion of the southern Ecuador subduction zone approximately consistent with the southern segment that had ruptured in 1942. This fits the pattern of segmented re-rupture that followed the 1906 earthquake — suggesting that the cycle may be repeating, with partial re-ruptures occurring on segments that had been reset by the 1906 megathrust.

The 2016 earthquake provided an important test of seismic hazard models for the Ecuador coast. The event was well-recorded by regional seismograph networks that did not exist in 1942 or 1906, and the quality of strong-motion recordings, GPS deformation measurements, and tsunami observations provided a comprehensive characterization of the source and its effects. Comparison between the 2016 observations and the historical record of the 1942 event — reconstructed from intensity reports and limited instrumental data — has allowed researchers to assess how well the segmented re-rupture model fits the observations.

The death toll from the 2016 event — 670 people for an M7.8 earthquake in a coastal region — reflects both the improved warning capabilities of modern Ecuador and the persistent vulnerability of the building stock. Ecuador had joined the Pacific Tsunami Warning Centre network and had national-level seismic early warning capabilities that the 1906 coast entirely lacked. Yet the combination of old unreinforced masonry construction in coastal towns and the proximity of the EpicenterThe point on the Earth's surface directly above the hypocenter (focus) where an earthquake originates underground. Often reported as the earthquake's location in news reports. to the coast still proved deadly. The 2016 event served as a reminder that the long recurrence of truly great earthquakes like 1906 does not eliminate the hazard of the more frequent, smaller ruptures on the same fault system.

Megathrust Supercycles in the Northern Andes

The 1906 earthquake has become a key data point in the emerging concept of megathrust "supercycles" — multi-century sequences in which great megathrust earthquakes alternate with periods of partial re-rupture and stress redistribution. The concept, developed partly from studying the historical record of the Sumatra and Japan megathrusts as well as the South American examples, holds that the largest Subduction ZoneA region where one tectonic plate dives beneath another into the mantle. Subduction zones produce the world's largest earthquakes (M8.5+) and are associated with deep ocean trenches and volcanic arcs. earthquakes are not simply periodic events with a fixed Earthquake Recurrence IntervalThe average time between major earthquakes on a particular fault. Estimated from paleoseismology and historical records. The Cascadia subduction zone has a recurrence interval of ~500 years. but part of longer cycles that can span centuries.

In the northern Andes segment, the available historical record — extending back to the 16th century with the arrival of Spanish colonial administration — suggests that very large earthquakes (M8.5+) have occurred in 1906, and possibly in 1687 in the southern segment. The gaps between these events, and the pattern of smaller events in between, are consistent with a supercycle model in which the megathrust periodically charges and discharges, with the charging period punctuated by smaller but still destructive partial re-ruptures.

[[Paleoseismology]] evidence from the Colombia-Ecuador coast adds another dimension. Studies of coastal stratigraphy — examining how the land surface has risen and fallen in response to earthquake cycles — provide a multi-millennium record of megathrust activity. These studies indicate that very large Fault RuptureThe breakage of rock along a fault during an earthquake, releasing stored elastic energy as seismic waves. Rupture length can range from meters (small quakes) to 1,000+ km (great earthquakes). events like 1906 occur on timescales of perhaps 200 to 500 years in this segment, separated by periods of smaller events. The submergence and re-emergence of coastal wetlands and mangrove forests, recorded in organic-rich sediment layers that can be precisely radiocarbon dated, provides a biological clock for tracing the history of large earthquakes on this coast.

Understanding megathrust supercycles has direct implications for the Seismic Risk AssessmentThe process of evaluating earthquake hazard, building vulnerability, and potential losses for a specific area or structure. Combines hazard maps, building inventory, and damage models. of cities like Guayaquil, Quito, Cali, and Medellín — major urban centres that would all be strongly affected by a repeat of the 1906 scenario. The 1906 Ecuador-Colombia earthquake thus remains a living and actively studied seismological document, re-examined with each new scientific advance as our understanding of Subduction ZoneA region where one tectonic plate dives beneath another into the mantle. Subduction zones produce the world's largest earthquakes (M8.5+) and are associated with deep ocean trenches and volcanic arcs. mechanics deepens and as the populations exposed to its potential successor continue to grow rapidly along the Pacific coast.

The Nazca Plate's Seismic Budget

The Colombia-Ecuador subduction zone is part of the larger Nazca Plate system, and understanding the 1906 earthquake requires placing it in the context of the Nazca Plate's overall seismic budget. The Nazca Plate is one of the most seismically productive oceanic plates in the world, responsible not only for the 1906 and 1960 events but for a series of great earthquakes spanning the entire length of the South American subduction zone. The concentration of very large earthquakes along this margin — the highest density of M8+ events of any comparable length of subduction zone in the world — reflects the combination of high convergence rate, young and buoyant plate, and high coupling between the descending and overriding plates.

Within this context, the 1906 Colombia-Ecuador rupture occupied the northern end of the Nazca-South America boundary — a section that transitions northward into the transform boundary of the Panama Fracture Zone and southward into the more extensively studied Peruvian and Chilean segments. The relative lack of study of the Colombia-Ecuador segment compared with Peru and Chile reflected partly the historical accident of where major earthquakes happened to occur during the instrumental era, and partly the logistical difficulties of working in the remote Colombian and Ecuadorian Pacific coast region.

The 2016 Pedernales earthquake changed this situation by motivating a concentrated scientific investment in the region. The international research collaborations it catalysed have substantially improved characterization of the megathrust structure, the distribution of coupling, and the potential for future large ruptures. The 1906 earthquake — previously known primarily through its tsunami and its place in the historical catalogue — is now understood with substantially greater precision as a result of the scientific work motivated by its 2016 partial re-rupture.

The Tumaco Hazard: Building Vulnerability and Population Growth

The Colombian Pacific coast, centred on the port city of Tumaco, presents one of the most challenging seismic risk management problems in South America. Tumaco is built largely on low-lying deltaic islands at the mouth of the Mira and Patía rivers, barely above sea level. Its building stock is predominantly light timber frame construction, which typically performs better than adobe in earthquakes but is extremely vulnerable to TsunamiA series of ocean waves generated by sudden displacement of the seafloor during an underwater earthquake. Tsunamis can travel across entire ocean basins at jet speed (700+ km/h). inundation and to Lateral SpreadingThe horizontal movement of soil blocks toward a free face (cliff or stream bank) during liquefaction. Can cause extensive damage to infrastructure, bridges, and pipelines. in the poorly consolidated deltaic soils. The population of the Tumaco metropolitan area has grown from a few thousand in 1906 to over 200,000 today, with much of this growth in informal settlements in the most hazardous coastal locations.

The combination of extreme tsunami hazard — Tumaco sits in the near field of the Colombia-Ecuador megathrust, with potential wave arrival times of three to five minutes — and inadequate Seismic Hazard MapA map showing the probability of earthquake shaking exceeding specified levels over a given time period. Used by engineers, planners, and insurers to assess earthquake risk.ping and enforcement of setback regulations means that a large proportion of Tumaco's population is exposed to life-threatening tsunami inundation with essentially no time for warning-based evacuation. Studies by Colombian geologists and international partners have repeatedly identified Tumaco as one of the highest tsunami risk cities in the Americas, and have recommended investment in vertical evacuation structures, public education, and ultimately the relocation of the most exposed informal settlements to higher ground. Progress on these recommendations has been slow, constrained by the economic and political challenges that characterize coastal development management in lower-income settings globally.

The 1906 Ecuador-Colombia earthquake thus anchors a hazard assessment challenge that grows more urgent with each passing decade as the cities of the Colombia-Ecuador Pacific coast continue to expand into terrain that would be catastrophically inundated by a future megathrust rupture. The Earthquake Recurrence IntervalThe average time between major earthquakes on a particular fault. Estimated from paleoseismology and historical records. The Cascadia subduction zone has a recurrence interval of ~500 years. of such events, while measured in centuries, does not diminish the urgency: the question is not whether a comparable event will occur, but whether the communities exposed to it will be any better prepared than they were in 1906.

The experience of the 2016 Pedernales earthquake — which killed 670 people despite modern warning systems and better construction standards than existed in 1906 — suggests that the answer to this question is "somewhat, but not enough." Each partial re-rupture along the Colombia-Ecuador megathrust is an opportunity to test and improve the preparedness systems that would be needed for the eventual full-segment rupture, and the improvements driven by each event represent genuine progress. But the gap between available Seismic Risk AssessmentThe process of evaluating earthquake hazard, building vulnerability, and potential losses for a specific area or structure. Combines hazard maps, building inventory, and damage models. knowledge and the actual state of buildings, warning systems, and public awareness in the most exposed coastal communities remains substantial, and will require sustained political will and financial investment to close before the next major event arrives.

International Scientific Collaboration After 2016

The 2016 Pedernales M7.8 earthquake catalysed a significant expansion of international scientific collaboration on the Colombia-Ecuador subduction zone. Research teams from Ecuador, Colombia, France, Germany, the United States, and Japan deployed temporary seismograph arrays, GPS networks, and ocean-bottom seismographs in the months following the earthquake, generating a dataset of unprecedented quality for characterizing the structure and locking pattern of this segment of the megathrust.

The temporary networks recorded thousands of AftershockA smaller earthquake that follows the mainshock in the same fault region. Aftershock sequences can last weeks to years, with the largest aftershock typically 1.0-1.2 magnitudes below the mainshock.s over several months, allowing three-dimensional mapping of the fault geometry at depths that were previously unknown. GPS measurements of post-seismic deformation — the slow creep that continues for years after a large earthquake as the crust adjusts to the new stress state — provided constraints on the distribution of slip during the mainshock and on the rheological properties of the lower crust. Offshore survey campaigns mapped the seafloor topography and sediment structure in the source zone, identifying submarine landslides triggered by the 2016 earthquake and characterizing the bathymetric features that influence tsunami propagation toward the coast.

This collaborative scientific effort has substantially improved the Seismic Risk AssessmentThe process of evaluating earthquake hazard, building vulnerability, and potential losses for a specific area or structure. Combines hazard maps, building inventory, and damage models. of the Colombia-Ecuador coast and has produced a new generation of Seismic Hazard MapA map showing the probability of earthquake shaking exceeding specified levels over a given time period. Used by engineers, planners, and insurers to assess earthquake risk.s and tsunami inundation models that are being used to update building codes, land use regulations, and emergency response plans in both countries. The 1906 earthquake — the event that defines the upper bound of the regional hazard — thus continues to shape scientific and policy agendas more than a century after it occurred.