arrow_back
In Depth
In Depth · 7 min read

Plate Tectonics, Subduction Zones and Megathrust Earthquakes

Where one plate sinks beneath another, the boundary can lock for centuries and then fail in a magnitude 9 earthquake. Here is how subduction works, what the great megathrust earthquakes showed, and how scientists dated Cascadia’s last rupture to a winter night in 1700.

How and Why Tectonic Plates Move

Earth’s outer shell is broken into rigid plates that move relative to one another. The U.S. Geological Survey (USGS) describes their speed as roughly the rate at which fingernails grow. Each plate moves at its own speed and in its own direction, and GPS now measures crustal motion to within a fraction of a millimeter per year (USGS n.d.a).

The forces behind that motion were weighed in a classic study by Forsyth and Uyeda (1975). They balanced the torques acting on individual plates to compare the forces that drive them with the forces that resist them. Their analysis pointed to the sinking of dense oceanic slabs at deep-sea trenches as the main driving force, an effect now usually called slab pull, balanced mostly by viscous resistance in the mantle (Forsyth and Uyeda 1975). A second, smaller force, ridge push, acts where new seafloor forms at mid-ocean ridges.

Slab pull places subduction zones at the center of plate tectonics. The places where old oceanic plates sink back into the mantle are also the places that store and release the greatest amounts of strain, which is why they host the largest earthquakes on record.

The Three Types of Plate Boundaries

Plates interact at three main kinds of boundary (Kious and Tilling 1996). At divergent boundaries, plates move apart and new crust forms from magma rising out of the mantle. The mid-ocean ridges are the main example.

At transform boundaries, plates slide horizontally past one another and crust is neither created nor destroyed (Kious and Tilling 1996). The San Andreas Fault in California is the best known example.

At convergent boundaries, plates move toward each other and crust is destroyed as one plate descends beneath another (Kious and Tilling 1996). Where oceanic lithosphere meets another plate, the denser oceanic plate sinks, and the boundary becomes a subduction zone. Kious and Tilling (1996) also note that some plate boundaries are broad zones of deformation rather than sharp lines, a reminder that the three types are an idealization of a more complicated surface.

What Is a Subduction Zone?

A subduction zone is a convergent boundary where one plate dives beneath another into the mantle. At the surface it is marked by a deep ocean trench, and where an oceanic plate descends beneath a continent, chains of volcanoes such as the Andes and the Cascades rise inland (Kious and Tilling 1996).

The contact between the two plates is a vast, gently dipping fault called a megathrust. At shallow depths, less than about 25 kilometers, the interface can become stuck, or locked, and stress builds along it (USGS n.d.b). In the hundreds of years between great earthquakes, the continued squeezing causes the upper plate to bulge and rise just above and inboard of the locked region, over thousands of square kilometers (USGS n.d.b).

When the accumulated stress exceeds the strength of the fault, the locked section breaks free and releases its stored energy as seismic waves (USGS n.d.b). The upper plate springs back. Areas that had been bulging upward can drop suddenly, while the seafloor above the rupture is thrust upward and displaces the overlying ocean. That sudden vertical motion of the seafloor is what generates the great subduction zone tsunamis.

Why Megathrust Earthquakes Are the Largest

USGS describes the shallow megathrust interface of subduction zones as host to the largest earthquakes and as the only type of fault capable of magnitude 9 or larger ruptures (USGS n.d.c). The reason lies in how magnitude is measured. Moment magnitude is derived from the seismic moment, which is the product of the rigidity of the rock, the area of the fault that slipped and the average amount of slip.

A megathrust offers an enormous fault area. Because the interface dips at a shallow angle, a rupture can extend a long way down the dip of the fault as well as for many hundreds of kilometers along the trench. In a study of subduction interface earthquakes of magnitude 7.1 to 9.5, Allen and Hayes (2017) found that rupture width levels off at about 200 kilometers for the largest events. Beyond that point, earthquakes grow mainly by lengthening along the trench and by larger slip, and the authors concluded that the size of great earthquakes appears to be controlled more strongly by average slip across the strongest patches of the fault (Allen and Hayes 2017).

No other fault type combines such length, width and slip. The two largest earthquakes recorded with modern instruments both struck subduction zones: the magnitude 9.5 Chile earthquake of 1960 and the magnitude 9.2 Alaska earthquake of 1964 (USGS n.d.h).

Chile 1960, Sumatra 2004 and Tōhoku 2011

The Chile earthquake of 22 May 1960 is the largest ever recorded, with a moment magnitude of 9.5 (USGS n.d.d). Early reports gave a surface-wave magnitude of 8.6. The moment magnitude, a better measure of the energy released, was later estimated at 9.5 (USGS n.d.d). It ruptured the boundary where the Nazca Plate descends beneath South America. Cifuentes (1989) estimated the rupture length from aftershocks and crustal deformation at 920 ± 100 kilometers, ending near the point where the Chile Ridge meets the Peru–Chile Trench. USGS records about 1,655 deaths in southern Chile and some 2 million people left homeless. The tsunami crossed the Pacific, killing 61 people in Hawaii, mostly in Hilo, and 138 in Japan (USGS n.d.d).

The Sumatra–Andaman earthquake of 26 December 2004 had a magnitude of 9.1 (USGS n.d.f). Its aftershocks outlined a rupture about 1,300 kilometers long, from northwestern Sumatra to the Andaman Islands (Lay et al. 2005). The Indian Ocean tsunami that followed left 227,898 people dead or missing and presumed dead, and about 1.7 million displaced (USGS n.d.g).

The Tōhoku-Oki earthquake of 11 March 2011 struck the Japan Trench, where the Pacific Plate descends beneath northern Japan. USGS rates it at magnitude 9.1 and reports that the fault moved as much as 50 to 60 meters over an area about 400 kilometers long and 150 kilometers wide (USGS n.d.e). Seafloor surveys made less than two weeks after the earthquake indicated that the seafloor near the trench had moved about 50 meters east-southeast and 7 to 10 meters upward, evidence that the rupture reached the trench axis (Fujiwara et al. 2011). Japan’s National Police Agency had confirmed at least 15,782 deaths six months after the disaster (Nippon.com n.d.).

The Cascadia Subduction Zone and the 1700 Earthquake

The Cascadia Subduction Zone runs offshore from northern Vancouver Island to Cape Mendocino in northern California, where the Juan de Fuca Plate and its neighbors descend beneath North America (PNSN n.d.). The Juan de Fuca Plate moves toward North America at about 4 centimeters per year (PNSN n.d.).

The last great Cascadia earthquake occurred on 26 January 1700, with an estimated magnitude between 8.7 and 9.2 (Wirth et al. 2025). Its date and size were reconstructed from a tsunami recorded in Japan. Samurai, merchants and villagers wrote of waves in January 1700 that arrived without any local earthquake, which later researchers called an orphan tsunami (Atwater et al. 2005). Satake et al. (1996) ruled out South America, Alaska and Kamchatka as sources, because historical and geological records showed no large earthquake there in 1700, and placed the earthquake in Cascadia at about 21:00 local time on 26 January 1700. From the tsunami heights they estimated a magnitude of about 9, which would mean a rupture of the entire subduction zone (Satake et al. 1996).

A later analysis compared flooding reported at seven sites in Japan, with heights of roughly 1 to 5 meters, against tsunamis modeled from six possible Cascadia sources. The most likely earthquake had a moment magnitude of 9.0, with 19 meters of slip on an offshore zone about 1,100 kilometers long (Satake et al. 2003). The range of plausible sources corresponds to magnitudes 8.7 to 9.2. Shorter ruptures could match the Japanese tsunami only with up to 40 meters of offshore slip, and the changes in land level they would cause conflict with the geological evidence along the Cascadia coast (Satake et al. 2003).

Evidence for Cascadia Earthquakes: Ghost Forests and Buried Marshes

The geological evidence came first from the coast. Atwater (1987) found that tidal mud had buried coastal lowlands in westernmost Washington six times in about 7,000 years, each burial probably caused by sudden subsidence of roughly 0.5 to 2 meters. Sand sheets on at least three of the buried surfaces pointed to tsunamis arriving with the same events (Atwater 1987). This is the pattern expected when the upper plate drops at the moment a megathrust fails.

The same subsidence created ghost forests. Along the Copalis River in Washington, a stand of dead western red cedars rises from a brackish tidal marsh. The trees were killed when the land dropped during the most recent Cascadia earthquake and salt water flooded their roots (USGS n.d.i). Yamaguchi et al. (1997) used tree rings to date the deaths. Trees killed by the earthquake had completed growth in 1699 but did not begin growth in 1700, which placed the subsidence between the 1699 and 1700 growing seasons and matched the date from the Japanese records (Yamaguchi et al. 1997).

Offshore sediments extend the record further back. Turbidites, deposits left by underwater landslides, accumulate in submarine channels along the margin. Goldfinger et al. (2012) traced them across about 500 kilometers of northern Cascadia and concluded that their synchronous timing over 10,000 years is best explained by triggering in great earthquakes. The record suggests clusters of two to five earthquakes separated by unusually long intervals, rather than a regular cycle, and the offshore record correlates well with the shorter record preserved on land (Goldfinger et al. 2012).

USGS now estimates a 10 to 15 percent chance of an approximately magnitude 9 Cascadia earthquake in the next 50 years (Wirth et al. 2025). The estimate rests on the same evidence: written records from Japan, dead trees and buried soils on the Pacific coast, and layered sediment on the seafloor.

Frequently asked questions

What is a subduction zone?
A subduction zone is a convergent plate boundary where one plate, usually oceanic, descends beneath another into the mantle. It is marked by a deep ocean trench, and the contact between the plates is a large, gently dipping fault called a megathrust.
What is a megathrust earthquake?
A megathrust earthquake is a rupture of the locked interface between two plates at a subduction zone. Stress builds for centuries while the shallow part of the interface is stuck, then releases when the fault breaks free. USGS describes these faults as the only ones capable of magnitude 9 or larger earthquakes.
Why do subduction zones produce the largest earthquakes?
Earthquake magnitude depends on the area of fault that slips and the amount of slip. A megathrust dips at a shallow angle and can extend for many hundreds of kilometers along a trench, giving it a far larger fault area than other fault types. The largest events also involve tens of meters of slip.
What was the largest earthquake ever recorded?
The largest earthquake recorded with instruments struck southern Chile on 22 May 1960, with a magnitude of 9.5. Its tsunami killed 61 people in Hawaii and 138 in Japan, in addition to about 1,655 deaths in Chile.
How do we know the Cascadia earthquake happened in 1700?
Japanese records describe a tsunami in January 1700 that arrived with no local earthquake. Tree rings from cedars killed when the Pacific Northwest coast subsided show the trees died between the 1699 and 1700 growing seasons. Together these place the earthquake on 26 January 1700.
How likely is a magnitude 9 Cascadia earthquake?
USGS estimates a 10 to 15 percent chance of an approximately magnitude 9 earthquake on the Cascadia Subduction Zone in the next 50 years. Offshore sediment records suggest such earthquakes come in clusters separated by long intervals rather than on a regular schedule.

References

  1. Allen, Trevor I., and Gavin P. Hayes. 2017. “Alternative Rupture-Scaling Relationships for Subduction Interface and Other Offshore Environments.” Bulletin of the Seismological Society of America 107 (3): 1240–53. doi.org/10.1785/0120160255.
  2. Atwater, Brian F. 1987. “Evidence for Great Holocene Earthquakes along the Outer Coast of Washington State.” Science 236 (4804): 942–44. doi.org/10.1126/science.236.4804.942.
  3. Atwater, Brian F., Satoko Musumi-Rokkaku, Kenji Satake, Yoshinobu Tsuji, Kazue Ueda, and David K. Yamaguchi. 2005. The Orphan Tsunami of 1700: Japanese Clues to a Parent Earthquake in North America. U.S. Geological Survey Professional Paper 1707. doi.org/10.3133/pp1707.
  4. Cifuentes, I. L. 1989. “The 1960 Chilean Earthquakes.” Journal of Geophysical Research: Solid Earth 94 (B1).
  5. Forsyth, D. W., and S. Uyeda. 1975. “On the Relative Importance of the Driving Forces of Plate Motion.” Geophysical Journal of the Royal Astronomical Society 43 (1): 163–200.
  6. Fujiwara, Toshiya, Shuichi Kodaira, Tetsuo No, Yuka Kaiho, Narumi Takahashi, and Yoshiyuki Kaneda. 2011. “The 2011 Tohoku-Oki Earthquake: Displacement Reaching the Trench Axis.” Science 334 (6060): 1240. doi.org/10.1126/science.1211554.
  7. Goldfinger, Chris, et al. 2012. Turbidite Event History—Methods and Implications for Holocene Paleoseismicity of the Cascadia Subduction Zone. U.S. Geological Survey Professional Paper 1661-F. pubs.usgs.gov/publication/pp1661F.
  8. Kious, W. Jacquelyne, and Robert I. Tilling. 1996. This Dynamic Earth: The Story of Plate Tectonics. U.S. Geological Survey general interest publication. pubs.usgs.gov/gip/dynamic/dynamic.html.
  9. Lay, Thorne, Hiroo Kanamori, Charles J. Ammon, Meredith Nettles, Steven N. Ward, Richard C. Aster, Susan L. Beck, et al. 2005. “The Great Sumatra-Andaman Earthquake of 26 December 2004.” Science 308 (5725): 1127–33. doi.org/10.1126/science.1112250.
  10. Nippon.com. n.d. “The Great East Japan Earthquake Disaster: Latest Figures.” www.nippon.com/en/features/h00004.
  11. PNSN (Pacific Northwest Seismic Network). n.d. “Cascadia Subduction Zone.” pnsn.org/education/pnw-earthquakes/sources/cascadia.
  12. Satake, K., K. Shimazaki, Y. Tsuji, and K. Ueda. 1996. “Time and Size of a Giant Earthquake in Cascadia Inferred from Japanese Tsunami Records of January 1700.” Nature 379 (6562): 246–49. doi.org/10.1038/379246a0.
  13. Satake, Kenji, Kelin Wang, and Brian F. Atwater. 2003. “Fault Slip and Seismic Moment of the 1700 Cascadia Earthquake Inferred from Japanese Tsunami Descriptions.” Journal of Geophysical Research 108 (B11). doi.org/10.1029/2003JB002521.
  14. USGS (U.S. Geological Survey). n.d.a. “How Fast Do Tectonic Plates Move?” www.usgs.gov/faqs/how-fast-do-tectonic-plates-move.
  15. USGS. n.d.b. “Introduction to Subduction Zones: Amazing Events in Subduction Zones.” Subduction Zone Science. www.usgs.gov/special-topics/subduction-zone-science/science/introduction-subduction-zones-amazing-events.
  16. USGS. n.d.c. “Earthquakes.” Subduction Zone Science. www.usgs.gov/special-topics/subduction-zone-science/science/earthquakes.
  17. USGS. n.d.d. “Historic Earthquakes: Chile, 1960 May 22, Magnitude 9.5.” Archived web page. webarchive.library.unt.edu/eot2008/20090109060417/http://earthquake.usgs.gov/regional/world/events/1960_05_22.php.
  18. USGS. n.d.e. “M 9.1 – 2011 Great Tohoku Earthquake, Japan.” Earthquake event page. earthquake.usgs.gov/earthquakes/eventpage/official20110311054624120_30.
  19. USGS. n.d.f. “Tsunami Generation from the 2004 M=9.1 Sumatra-Andaman Earthquake.” Pacific Coastal and Marine Science Center. www.usgs.gov/centers/pcmsc/science/tsunami-generation-2004-m91-sumatra-andaman-earthquake.
  20. USGS. n.d.g. “Magnitude 9.1 – Off the West Coast of Northern Sumatra.” Archived web page. webarchive.library.unt.edu/web/20120915214431/http://earthquake.usgs.gov/earthquakes/eqinthenews/2004/us2004slav.
  21. USGS. n.d.h. “1964 Great Alaska Earthquake and Tsunami.” Alaska Science Center. www.usgs.gov/centers/alaska-science-center/science/1964-great-alaska-earthquake-and-tsunami.
  22. USGS. n.d.i. “Dead Red Cedar Forest.” Image. www.usgs.gov/media/images/dead-red-cedar-forest.
  23. Wirth, Erin A., Arthur D. Frankel, Brian L. Sherrod, Alex R. Grant, Audrey Dunham, Ian P. Stone, and Julia Grossman. 2025. Earthquake Probabilities and Hazards in the U.S. Pacific Northwest. U.S. Geological Survey Fact Sheet 2025–3050. pubs.usgs.gov/publication/fs20253050.
  24. Yamaguchi, David K., Brian F. Atwater, Daniel E. Bunker, Boyd E. Benson, and Marion S. Reid. 1997. “Tree-Ring Dating the 1700 Cascadia Earthquake.” Nature 389 (6654): 922–23. doi.org/10.1038/40048.
Next in In DepthFaults and Stress: Why Earthquakes HappenAll In Depth articles
sisma
The Pulse of a Living Planet
Loading earthquakes…