Name Class Date

EARTH EXPLAINED · LESSON GUIDE

The Day Japan Moved

The 2011 Tōhoku earthquake and tsunami, from the plates up

Video 9:20Level Grades 6–12 · Introductory university geologyStandards MS-ESS2-2, MS-ESS3-2, HS-ESS2-1, HS-ESS3-1, HS-PS4-1
Play fromfor school networks that block YouTube

Why did one of the best-prepared countries on Earth still suffer so much on March 11, 2011, when the warning signs had been written in its history and buried in its soil for more than a thousand years?

On March 11, 2011, the Pacific plate slipped as much as 50 meters beneath northern Japan along the Japan Trench, in a magnitude 9.0–9.1 earthquake, the strongest ever measured in Japan. This lesson follows the event from the plates up: the locked subduction fault that stored strain for centuries, the Jōgan tsunami of 869 whose sand still lies beneath the Sendai plain, the first minutes of warnings, a tsunami that ran up about 40 meters and crossed the Pacific, the human and physical aftermath, and the changes Japan made afterward. It ends with the Nankai Trough, the fault Japan is watching next.

Captions (.vtt)Captions (.srt)

Learning objectives After this lesson, students can:

  1. Identify the four plates that meet at Japan and explain how subduction of the Pacific plate at the Japan Trench stores strain on a locked fault and releases it in great earthquakes.
  2. Explain how a sudden uplift of the seafloor generates a tsunami, and why a tsunami travels fast and low in the deep ocean but slows and grows as it reaches the coast.
  3. Use historical records and tsunami sand layers to explain how the 869 Jōgan tsunami revealed the hazard, and why forecasts that expected an earthquake of around magnitude 8 still underestimated it.
  4. Describe how earthquake early warning and tsunami warnings performed on March 11, 2011, including why the first magnitude estimate of 7.9 was too small and how offshore observations corrected the warnings.
  5. Interpret measurements of the event's physical effects: rupture length and slip, run-up height, GPS displacement and subsidence of the coast, and the calculated change in the length of the day.
  6. Evaluate what protected people and what failed, using the Aneyoshi tsunami stone, the Kamaishi school drills, overtopped seawalls and the Fukushima Daiichi nuclear plant.
  7. Explain how Japan changed its warnings and monitoring after 2011, and interpret a long-term probability such as the 30-year forecast for the Nankai Trough.

Before watching

Key vocabulary ★ extension terms for older students

Tectonic plate
A large slab of Earth's rigid outer shell, commonly around 100 km thick, that moves slowly over the hotter, softer mantle below. Japan sits where four plates meet: the Pacific, Philippine Sea, North American and Eurasian plates.
Subduction
The process in which one plate bends and sinks beneath another, back down into the mantle. Off northeastern Japan, the Pacific plate sinks beneath Japan.
Ocean trench
The long, deep valley in the seafloor where one plate bends down beneath another. The Japan Trench, where the Pacific plate begins its descent, is about 8 km deep.
Locked fault
A part of a plate boundary where friction holds the two plates together, so their motion is stored as elastic strain in the rock, like a spring being wound, until the fault suddenly slips in an earthquake.
Tsunami
A series of ocean waves made when a large mass of water is suddenly pushed, here by the seafloor lurching upward. The first wave is often not the largest.
Magnitude
A measure of an earthquake's size. Each whole step releases about 32 times more energy. The 2011 earthquake was magnitude 9.0 by the Japan Meteorological Agency's count and 9.1 by the U.S. Geological Survey's.
Tsunami deposit
A layer of sand carried inland from the beach and seafloor by a tsunami and left on top of soil or marsh mud. Marine microfossils in the sand show that it came from the sea.
Run-up
The height above sea level of the highest point on land that the water reaches. It can be much greater than the wave's height at the shoreline, especially at the heads of narrow, steep-sided bays.
Earthquake early warning
A system that detects the first, faster seismic waves near the source of an earthquake already under way and sends an alert before the slower, stronger shaking arrives farther away. It gives seconds to about a minute of warning, not hours.
Subsidence
The sinking of the land surface. In 2011 the Oshika Peninsula sank by about 1.2 m in minutes, leaving parts of the coast permanently lower and more exposed to the sea.
Tsunami stone
A stone monument on Japan's coast that records a past tsunami and warns later generations, such as the stone at Aneyoshi: do not build your homes below this point.
Length of day
The time Earth takes to turn once, about 86,400 seconds. Moving mass closer to Earth's axis makes the planet spin slightly faster, as a skater does by pulling in her arms.
Megathrust earthquake★
An earthquake on the gently dipping fault between a subducting plate and the plate above it. All earthquakes of magnitude 9 and larger, including the 2011 Tōhoku earthquake, have been megathrust events.
Seismic moment★
The physical size of an earthquake: the rigidity of the rock × the area of the fault that slipped × the average slip. Moment magnitude is calculated from it.
Paleoseismology★
The study of earthquakes and tsunamis that happened before instruments existed, using evidence such as buried sand layers, sudden changes in land level and dated plant material.
Design basis★
The largest hazard, such as an earthquake, flood or tsunami, that a structure is engineered to withstand. At Fukushima Daiichi the tsunami far exceeded it.

Chapter outline Select a time to play from there

0:00The Moment

At 2:46 p.m. on Friday, March 11, 2011, the seafloor about 70 km off northeastern Japan gave way. The ground shook for about six minutes, and in many places along the coast people had roughly half an hour before the sea arrived.

0:37Four Plates

Japan sits where the Pacific, Philippine Sea, North American and Eurasian plates meet. The Pacific plate moves toward Japan at about 8 cm a year and sinks at the Japan Trench; where the plates lock, strain builds until the fault slips and the seafloor lurches upward. This is why Japan has about one in five of the world's strong earthquakes and more than a hundred active volcanoes.

2:10Warnings from the Past

The Jōgan tsunami of 869 flooded the Sendai plain and drowned about 1,000 people, and from about 1990 geologists found its sand, with older layers beneath it, far inland; yet official forecasts expected an earthquake of around magnitude 8, not 9. On the Sanriku coast, tsunamis in 1896 and 1933 killed about 22,000 and 3,000 people, and the village of Aneyoshi carved a warning into stone.

3:362:46 PM

The rupture began about 70 km off the Oshika Peninsula at about 30 km depth, tore about 400 km along the trench and slipped up to about 50 m. Early warning gave Tokyo about a minute; the first tsunami warning, issued three minutes after the earthquake, was based on a magnitude of 7.9 and forecast 3 to 6 m, and was raised to more than 10 m once sensors at sea detected the real wave.

4:53The Wave

Why a tsunami crosses the deep ocean at jet speed and piles up near shore. The water ran up about 40 m near Miyako, overtopped seawalls, swept across Sendai Airport and flooded more than 500 km², while Aneyoshi, built above its stone, was spared. At Fukushima Daiichi a wave about 14 m high flooded a site designed for about 6 m, and three reactors melted down; the tsunami reached Hawaii in about 7 hours, California in about 10 and Chile in about a day.

6:37Aftermath

More than 18,000 people died or were never found, most by drowning, and about 122,000 buildings were destroyed; the World Bank estimated the damage at up to $235 billion. GPS showed the Oshika Peninsula moved 5.3 m and sank 1.2 m, and a NASA calculation suggests the day became about 1.8 millionths of a second shorter. In Kamaishi, years of drills meant that almost all of the city's nearly 3,000 schoolchildren survived.

7:49What Japan Changed

Since 2013, when an earthquake is too large to measure quickly, the first tsunami warning says the tsunami will be 'huge' instead of guessing a number. A network of 150 cabled seafloor stations now watches the Japan and Kuril trenches, and more than 400 km of new seawalls, some taller than 12 m, line the coast.

8:24Still Moving

Japan still feels about 2,000 earthquakes a year. In 2025 the government estimated the 30-year chance of a great Nankai Trough earthquake at 60–90% or more by one method and 20–50% by another. The summary: three things to remember, and the lesson of the Aneyoshi stone.

Common misconceptions For the teacher

MisconceptionA tsunami is a single giant wall of water that is just as tall out at sea as it is at the coast.

What the evidence showsIn the deep ocean a tsunami may be barely a meter high, though it moves at about 800 km/h. It grows only as it slows in shallow water, and it arrives as a series of waves. In 2011 the first small readings at the coast led some people to believe the tsunami would be minor, and the Japan Meteorological Agency later changed how it announces early observations for that reason.

MisconceptionThe first tsunami warning tells you how big the wave will be, so if the forecast is lower than the local seawall, you are safe.

What the evidence showsThe first forecast in 2011, 3 to 6 m, was based on a magnitude estimate of 7.9 that was far too small, and Japan's weather agency concluded that it may have delayed some evacuations. The safe response to strong, long shaking at the coast is to go to high ground at once, whatever the first number says.

MisconceptionScientists had no way of knowing that such a large tsunami could strike this coast.

What the evidence showsThe 869 Jōgan tsunami was described in Japan's historical records, and from about 1990 geologists found its sand, and older layers, kilometers inland on the Sendai plain. The evidence existed; official forecasts still expected an earthquake of around magnitude 8, because the instrumental record was short and the geological work was still being turned into hazard assessments.

MisconceptionEarthquake early warning predicts earthquakes before they happen.

What the evidence showsEarly warning detects an earthquake that has already started and races the slower, more damaging waves. It can give tens of seconds to about a minute far from the source, as in Tokyo in 2011, but places closest to the rupture may feel strong shaking before any alert arrives.

MisconceptionThe 2011 earthquake knocked Earth off its axis and noticeably changed the length of the day.

What the evidence showsBy one NASA calculation the day became shorter by about 1.8 millionths of a second, and the axis around which Earth's mass is balanced shifted by about 17 cm; Earth's axis in space did not move. The change was calculated, not felt, and it is far smaller than the natural variations in day length caused by winds, ocean currents and Earth's core.

Discussion questions

  1. Geologists had found the Jōgan tsunami's sand on the Sendai plain more than twenty years before 2011, yet official forecasts expected an earthquake of around magnitude 8. Why might it take a long time for new geological evidence to change official plans? How should a society act on evidence that is strong but not yet certain?
  2. In Aneyoshi, a stone carved after 1933 kept every home above the reach of the 2011 tsunami. Elsewhere, seawalls built for earlier tsunamis were overtopped. Compare the two kinds of protection: what are the strengths and weaknesses of a remembered rule and of an engineered wall?
  3. Kamaishi's schoolchildren had practiced one rule for years: after a big earthquake, run for high ground immediately and don't wait for anyone. Why might a simple rule like this save more lives than waiting for an official warning? Can you think of situations where it would be hard to follow?
  4. The first tsunami warning in 2011 gave a number that turned out to be far too low. Since 2013, when the size is uncertain, the first warning simply says the tsunami will be 'huge'. Is a warning without a number better or worse than a precise but possibly wrong forecast? Explain your reasoning.
  5. At Fukushima Daiichi, a wave about 14 m high struck a site designed for about 6 m. How should engineers and governments decide how large a hazard a nuclear plant, hospital or school must be built to withstand, when the largest events happen only every few hundred or thousand years?
  6. For the Nankai Trough, Japan's government gave two 30-year probabilities: 60–90% or more by one method and 20–50% by another. What does each number mean for a family, a city and a national government? Which would you emphasize in public messages, and why?

Quiz Choose an answer; it checks itself

1.What happens where the Pacific plate meets northern Japan at the Japan Trench?

2.Why does a subduction fault produce sudden great earthquakes instead of slipping smoothly all the time?

3.What evidence showed that giant tsunamis had struck the Sendai plain long before 2011?

4.Three minutes after the earthquake, the first tsunami warning forecast waves of 3 to 6 meters. Why was this forecast far too low?

5.In the deep ocean a tsunami may be barely a meter high, yet it travels at about 800 km/h. What happens as it reaches shallow water near the coast?

6.What led to the disaster at the Fukushima Daiichi nuclear plant?

7.GPS showed that the Oshika Peninsula moved about 5.3 meters toward the sea and sank about 1.2 meters. What does this tell us?

8.Almost all of the nearly 3,000 schoolchildren in Kamaishi survived the tsunami. What made the biggest difference?

9.Since 2013, if an earthquake is too large to measure quickly, what does the first tsunami warning from Japan's weather agency do?

10.In 2025 Japan's government estimated the 30-year chance of a great Nankai Trough earthquake at 60–90% or more by one method and 20–50% by another. Which interpretation is best?

Written response

1. Explain, step by step, how the 2011 earthquake produced a tsunami that reached the coast of Japan in about half an hour and crossed the entire Pacific. Use the terms subduction, locked fault, strain, seafloor uplift and shallow water.

What a strong answer includes

The Pacific plate converges with Japan at about 8 cm a year and subducts at the Japan Trench. Where the plates touch, the fault is locked, so the edge of Japan is dragged down and squeezed and elastic strain builds. In 2011 the strain overcame friction: the fault slipped up to about 50 m over a zone about 400 km long, the squeezed edge sprang back, and the seafloor lurched upward, lifting the whole column of water above it. Gravity pulled that raised water back down, sending waves outward in every direction. In the deep ocean they traveled at about 800 km/h while barely a meter high; in shallow water near the coast they slowed and grew, reaching nearby shores in roughly half an hour and running up about 40 m in places. Across the open Pacific they reached Hawaii in about 7 hours and Chile in about a day.

2. Using the Jōgan tsunami of 869 as your example, explain how geology can reveal a hazard that the instrumental record misses, and why the 2011 earthquake was still larger than official forecasts expected.

What a strong answer includes

Seismographs have recorded earthquakes in Japan for only about a century, much shorter than the time between the largest events. Japan's historical records described the 869 tsunami flooding the Sendai plain and drowning about 1,000 people. From about 1990, geologists found its sand, carried in from the sea, kilometers inland, with older sand layers beneath it, showing that giant tsunamis had struck repeatedly, centuries apart. Yet official forecasts were built largely on the better-known recent earthquakes and expected an event of around magnitude 8. The geological work was still estimating how large the Jōgan earthquake had been, and its full lesson, that a much larger section of the fault could rupture at once, had not yet been built into hazard plans when the magnitude 9 struck.

3. Compare three kinds of protection described in the episode: the Aneyoshi stone, the Kamaishi school drills, and seawalls together with the Fukushima Daiichi plant. Which protected people, which failed, and why?

What a strong answer includes

The Aneyoshi stone worked because its rule did not depend on predicting the size of the next wave: every home was built above it, and the 2011 wave stopped about 90 m downhill from it. The Kamaishi drills worked because a simple, practiced rule, run for high ground immediately and don't wait, did not depend on warnings or estimates; about 99.8% of the city's students survived. Seawalls built for earlier tsunamis, and the Fukushima Daiichi site designed for a wave of about 6 m, failed because they were engineered to a fixed design height that the 2011 tsunami, about 14 m at the plant, exceeded. The lesson is that defenses sized to past events can be overtopped, while rules that move people out of reach work whatever the size of the wave; the best plans use both.

Going deeper University and advanced high school

How fast does a tsunami cross the Pacific, and why does it grow?

A tsunami's wavelength (hundreds of kilometers) is far longer than the ocean is deep, so it behaves as a shallow-water wave whose speed depends only on the depth d: c = √(g·d), with g = 9.81 m/s². As it moves into shallower water, conservation of energy flux makes its height grow roughly in proportion to 1 ÷ ⁴√d, the inverse fourth root of the depth (Green's law), before it breaks or floods the land.

ProblemCalculate the speed for a typical depth of 5,000 m, in m/s and km/h. The great-circle distance from the 2011 epicenter to Honolulu is about 5,960 km: predict the travel time and compare it with the roughly seven hours in the episode. Then estimate how much a wave 1 m high over 4,000 m of water would grow by the time it reaches water 10 m deep, and its speed there. Why can the run-up on land, about 40 m in 2011, be far greater still?

Worked answer

c = √(9.81 × 5,000) ≈ 221 m/s ≈ 800 km/h, the speed in the episode. 5,960 km ÷ 800 km/h ≈ 7.5 hours, close to the observed arrival of about seven hours (the real wave follows the actual seafloor, partly deeper than 5,000 m). Growth: ⁴√(4,000 ÷ 10) = ⁴√400 ≈ 4.5, so the 1 m wave becomes about 4.5 m high, moving at √(9.81 × 10) ≈ 10 m/s, or about 36 km/h, faster than most people can run. Run-up is the height the water climbs on land, not the wave's height offshore: on the Sanriku coast, narrow V-shaped bays funnel the flow and the moving water carries momentum uphill, so run-up there reached about 40 m.

Magnitude, energy and slip

Radiated energy rises by a factor of about 101.5 ≈ 32 for each whole step of magnitude. Moment magnitude is defined from the seismic moment M₀ (in newton-meters) by M₀ = 101.5·Mw + 9.1, and M₀ = μ·A·D, where μ is the rigidity of the rock, A the area that slipped and D the average slip. The USGS models the 2011 rupture as about 400 km long and 150 km wide, with peak slip of 50 to 60 m.

ProblemHow many times more energy did the final magnitude, 9.1 (USGS), represent than the first estimate of 7.9? How much does the difference between JMA's 9.0 and USGS's 9.1 matter? Using μ = 4 × 1010 Pa, calculate the average slip for Mw 9.0 and Mw 9.1 and compare it with the peak slip.

Worked answer

101.5 × 1.2 = 101.8 ≈ 63 times more energy than the first estimate, which is why the first tsunami forecast was so low. 101.5 × 0.1 = 100.15 ≈ 1.4: the two agencies' values differ by about 40% in energy, small next to the first error. For Mw 9.0, M₀ = 1022.6 ≈ 4.0 × 1022 N·m, so D = 4.0 × 1022 ÷ (4 × 1010 × 4 × 105 m × 1.5 × 105 m) ≈ 17 m; for Mw 9.1, M₀ ≈ 5.6 × 1022 N·m and D ≈ 23 m. The average slip of roughly 17 to 23 m is less than half the peak of 50 to 60 m, which was concentrated near the trench, where it lifted the seafloor most and drove the largest tsunami.

How much warning can early warning give?

P waves travel through the crust at roughly 6 to 7 km/s and S waves, which carry most of the strong shaking, at roughly 3.5 to 4 km/s. An early warning system detects the P waves at stations near the source, estimates the earthquake and sends an alert, which then travels at the speed of electronic communication. The 2011 rupture began about 30 km deep; Tokyo is about 376 km from the epicenter and the tip of the Oshika Peninsula about 76 km.

ProblemSuppose the alert goes out 30 seconds after the rupture begins. Estimate how long Tokyo had before the first S waves arrived. Then estimate the radius of the 'blind zone', the area where S waves arrive before the alert, and decide whether the Oshika Peninsula was inside it. Why might the strongest shaking in a great earthquake arrive later than this simple calculation suggests?

Worked answer

S-wave travel time to Tokyo ≈ 376 km ÷ 3.5 to 4 km/s ≈ 94 to 107 s, so the warning time is about 64 to 77 s, about a minute, as in the episode. In 30 s the S waves travel about 105 to 120 km from the hypocenter; at 30 km depth that reaches the surface out to √(105² − 30²) ≈ 100 km to √(120² − 30²) ≈ 116 km from the epicenter. The Oshika Peninsula, about 82 km from the hypocenter, felt S waves after only about 20 to 23 s, before such an alert. In a great earthquake the rupture itself keeps growing for minutes and sends out strong waves from parts of the fault that break later and lie closer to distant cities, so real warning times depend on the whole rupture, not only on its starting point.

The strain budget and the recurrence of Jōgan-type tsunamis

The Pacific plate converges with northern Japan at about 83 mm a year (USGS). Minoura and colleagues (2001) found three tsunami sand layers on the Sendai plain and estimated that the largest tsunamis struck about every 1,000 years; Sawai and colleagues (2012), using radiocarbon ages from many sites, suggested intervals of 500 to 800 years. A simple way to turn a recurrence interval T into a probability is to assume events occur at random: P = 1 − exp(−t/T) for a window of t years.

ProblemHow many years of plate convergence does a slip of 50 m represent? An average slip of about 20 m? How much convergence accumulated between 869 and 2011? Then calculate the probability of a Jōgan-type tsunami within 30 years and within 50 years for T = 1,000 years and T = 500 years. What do these numbers suggest about designing coastal plants or towns for a 'once in a thousand years' event?

Worked answer

50 m ÷ 0.083 m/yr ≈ 600 years; 20 m ÷ 0.083 m/yr ≈ 240 years. From 869 to 2011 (1,142 years) about 95 m of convergence accumulated, more than the slip in 2011, so not all convergence is stored as strain on the locked patch: some is released by slow, quiet slip and by the many smaller earthquakes on this plate boundary. Ozawa and colleagues (2011) note the other side of the budget: before 2011, GPS showed strain accumulating off Tōhoku much faster than earlier earthquakes on the plate boundary had released it on average, which raised the question of how that strain would eventually be released. Probabilities: for T = 1,000 years, 1 − exp(−0.03) ≈ 3% in 30 years and 1 − exp(−0.05) ≈ 5% in 50 years; for T = 500 years, about 6% and 10%. A few percent over the life of a power plant or a town is far from negligible when the consequences are catastrophic, which is why rare events need to be part of design, not left out of it.

Did the earthquake really change the length of the day?

With no outside torque, Earth's angular momentum L = I·ω is conserved, where I is the moment of inertia and ω the rate of rotation. If an earthquake moves mass slightly closer to the axis, I decreases, ω increases, and the day T = 2π/ω becomes shorter: ΔT/T = ΔI/I. NASA's Richard Gross calculated that the 2011 earthquake shortened the day by about 1.8 microseconds. Over the long term, tidal friction from the Moon lengthens the day; measurements from ancient eclipses show an average increase of about 1.8 milliseconds per century (Stephenson and colleagues, 2016).

ProblemCalculate the fractional change in Earth's moment of inertia implied by a 1.8 µs shorter day. How many days of the long-term lengthening trend would cancel it? If the change were permanent, by how much would a clock based on Earth's rotation gain on an atomic clock in one year?

Worked answer

ΔI/I = ΔT/T = −1.8 × 10−6 s ÷ 86,400 s ≈ −2.1 × 10−11. The trend of 1.8 ms per century is 1.8 × 10−3 s ÷ 36,525 days ≈ 4.9 × 10−8 s per day per day, so it would cancel the change in about 1.8 × 10−6 ÷ 4.9 × 10−8 ≈ 37 days. Over a year, 365 × 1.8 µs ≈ 0.66 ms. The effect is real but tiny, much smaller than the natural fluctuations in day length caused by winds, ocean currents and Earth's core, which is why it was calculated rather than measured.

Standards alignment

Next Generation Science Standards

MS-ESS2-2

Construct an explanation based on evidence for how geoscience processes have changed Earth's surface at varying time and spatial scales.

In this episode: Slow plate convergence of about 8 cm a year against minutes of slip, uplift of the seafloor and subsidence of the coast.

MS-ESS3-2

Analyze and interpret data on natural hazards to forecast future catastrophic events and inform the development of technologies to mitigate their effects.

In this episode: Tsunami sand layers and recurrence, early warning, tsunami warnings, seafloor cables, seawalls and the Nankai Trough forecast.

HS-ESS2-1

Develop a model to illustrate how Earth's internal and surface processes operate at different spatial and temporal scales to form continental and ocean-floor features.

In this episode: The subduction-zone model: trench, locked fault, strain build-up and elastic rebound of the overriding plate.

HS-ESS3-1

Construct an explanation based on evidence for how the availability of natural resources, occurrence of natural hazards, and changes in climate have influenced human activity.

In this episode: The Aneyoshi stone, Kamaishi's drills, coastal rebuilding and new seawalls, and the siting and design of Fukushima Daiichi.

HS-PS4-1

Use mathematical representations to support a claim regarding relationships among the frequency, wavelength, and speed of waves traveling in various media.

In this episode: Tsunami speed c = √(g·d) in water of different depths, and why the wave slows, shortens and grows as it reaches the coast.

Advanced Placement

University courses

Physical geology

Plate boundaries and microplates around Japan; subduction zones and trenches; the earthquake cycle and elastic rebound

Natural hazards

Megathrust earthquakes; tsunami generation, propagation, shoaling and run-up; paleoseismology and tsunami deposits; recurrence and long-term probability; warning systems and evacuation

Geophysics

Seismic moment and moment magnitude; magnitude saturation; earthquake early warning; GPS geodesy and coseismic displacement; earthquakes and Earth's rotation

Disaster risk management

Design basis and rare events in engineering; risk communication and warning wording; community drills and memory of past disasters

Transcript The narration, word for word

0:00The Moment

On March 11, 2011, at 2:46 on a Friday afternoon, the seafloor 70 kilometers off northeastern Japan suddenly gave way.

0:13The ground shook for about six minutes, and in many places along the coast, people had roughly half an hour before the sea arrived.

0:22This is the story of the strongest earthquake ever measured in Japan. We'll look at why it happened, why its warnings were already 1,000 years old, and what Japan changed afterward.

0:37Four Plates

To understand that day, we have to start underground, because Japan sits where four of Earth's tectonic plates meet.

0:45Those four are the Pacific plate, the Philippine Sea plate, the North American plate and the Eurasian plate. Each one is a slab of rock around 100 kilometers thick, drifting slowly across the planet.

1:00None of them ever stops moving. The Pacific plate is creeping toward Japan at about 8 centimeters a year, roughly twice as fast as your fingernails grow.

1:11Off the northeast coast, the Pacific plate meets northern Japan at the Japan Trench, a gash in the seafloor about 8 kilometers deep.

1:22At the trench, the Pacific plate bends and dives beneath Japan, sinking slowly into the hot mantle below.

1:32It does not slide smoothly, though. Where the two plates touch, they lock together, so year after year the edge of Japan is dragged down and squeezed, like a spring being wound tighter and tighter.

1:45Eventually, the strain overcomes the friction. In a matter of seconds the fault slips, the squeezed edge of the plate springs back, and the seafloor lurches upward, lifting the entire ocean above it.

1:57That process explains why Japan experiences about one in five of the world's strong earthquakes, and why the country has more than 100 active volcanoes.

2:10Warnings from the Past

Japan has written records of earthquakes going back more than 1,000 years, and one of them describes a night in the year 869.

2:20That night, a great earthquake struck near what is now the city of Sendai, and the sea rushed far inland across the coastal plain. About 1,000 people drowned.

2:31Starting around 1990, geologists digging on that same plain found the tsunami's calling card: a layer of ocean sand, kilometers from the coast, with older sand layers beneath it, left by even earlier giant waves.

2:47The evidence showed that giant tsunamis had struck this coast again and again, centuries apart. Official forecasts did expect a large earthquake here, but they expected one of around magnitude 8, not nine.

3:02Farther north, on the rugged Sanriku coast, the memories were more recent. In 1896, a tsunami ran up as high as 38 meters there, and killed about 22,000 people.

3:16In 1933, another tsunami killed about 3,000 more. Afterward, the people of a small village called Aneyoshi carved a warning into stone.

3:27The inscription reads: Do not build your homes below this point.

3:362:46 PM

On March 11, 2011, at 2:46 in the afternoon, the fault finally let go. The rupture began about 70 kilometers off the Oshika Peninsula, roughly 30 kilometers below the seafloor.

3:50From there, it tore along the trench for roughly 400 kilometers, and in places the two plates slipped as much as 50 meters past each other.

4:01Japan's earthquake early warning system worked as designed. Its sensors detected the first, faster waves, and alerts went out before the strongest shaking arrived.

4:12Tokyo, about 370 kilometers away, had about a minute of warning.

4:19Three minutes after the earthquake, Japan's weather agency issued its highest level of tsunami warning. The first estimate of the earthquake's size, however, was far too small, at magnitude 7.9, so the first wave forecasts were also too low, at 3 to 6 meters.

4:36Then sensors out at sea detected the real wave, and the warnings were raised to more than 10 meters. The final magnitude was 9 by Japan's count, and 9.1 by the U.S. Geological Survey.

4:53The Wave

In the deep ocean, a tsunami moves about as fast as a jet, around 800 kilometers an hour, even though it may be barely a meter high.

5:03As the water grows shallower near the coast, the wave slows down, and because the water behind it keeps arriving, it has nowhere to go but up.

5:11This tsunami spread out in every direction, across the entire Pacific.

5:18Near Miyako, the water ran up a hillside to 40 meters above sea level, and seawalls built for earlier tsunamis were simply overtopped.

5:28Ships were carried far inland. This ferry in Miyako was left stranded among the houses.

5:35On the Sendai plain, the same ground that had flooded in 869, the sea swept across Sendai Airport and continued for kilometers inland.

5:46In all, more than 500 square kilometers of Japan went underwater, and whole fishing towns along the coast were destroyed.

5:55In Aneyoshi, the wave stopped about 90 meters downhill from the old stone. Every home in the village had been built above it, and every one was spared.

6:06At the Fukushima Daiichi nuclear plant, a wave about 14 meters high flooded a site designed for about six. The plant lost power, three reactors melted down, and it became the worst nuclear accident since Chernobyl.

6:22Meanwhile, the tsunami kept traveling. It reached Hawaii about seven hours later and California about ten, wrecking boats in their harbors, and about a day after the earthquake it reached Chile.

6:37Aftermath

More than 18,000 people died or were never found, most of them by drowning, and more than 120,000 buildings were destroyed.

6:47The World Bank estimated the damage at up to $235 billion, which at the time made it the costliest natural disaster ever recorded.

6:59The earthquake even reshaped the land itself. The Oshika Peninsula moved more than 5 meters toward the sea, and sank by more than a meter.

7:10It also shifted so much of Earth's mass that, by one NASA calculation, the planet now spins very slightly faster. Each day became shorter by about 1.8 millionths of a second, far too little for anyone to notice.

7:25There are also stories of survival. In the city of Kamaishi, schoolchildren had practiced one simple rule for years: after a big earthquake, run for high ground immediately, and don't wait for anyone.

7:40Almost all of the city's nearly 3,000 students survived.

7:49What Japan Changed

Japan rebuilt its tsunami warnings. Now, when an earthquake is too large to measure quickly, the first warning no longer guesses at a number. It simply tells people that the tsunami will be huge.

8:00Along the Japan and Kuril trenches, 150 seafloor stations now listen for earthquakes and tsunamis, connected to land by about 5,500 kilometers of cable.

8:14More than 400 kilometers of new seawalls now line the coast, and some of them are taller than 12 meters.

8:24Still Moving

The plates have not stopped moving. In a typical year, Japan feels about 2,000 earthquakes, or five to six every day. Most are too small to notice, but each one is a sign of the same plates, still moving.

8:37Scientists are also watching another fault, the Nankai Trough, off southwestern Japan. In 2025, Japan's government estimated the chance of a great earthquake there within 30 years at 60 to 90% or more by one method, and at 20 to 50% by another.

8:56So, three things to remember. First, the 2011 earthquake came from the Pacific plate slipping as much as 50 meters beneath northern Japan, at the Japan Trench, which made it a magnitude 9, the strongest ever measured in Japan. Second, the tsunami did most of the harm: it ran up to 40 meters high, and more than 18,000 people died or were never found. And third, the warnings were already there, in the flood of 869, and in stones like the one at Aneyoshi.

9:06The stone at Aneyoshi still stands, and its message is the lesson of this whole story. The Earth keeps its own calendar, and it remembers what we forget.

Sources

  1. U.S. Geological Survey. M 9.1 – 2011 Great Tohoku Earthquake, Japan. Earthquake Hazards Program event page, with tectonic summary. earthquake.usgs.gov/earthquakes/eventpage/official20110311054624120_30/executiveUsed for: Magnitude 9.1, location and depth, plate motion of 83 mm a year, rupture size and slip
  2. Japan Meteorological Agency. (2013). Lessons learned from the tsunami disaster caused by the 2011 Great East Japan Earthquake and improvements in JMA's tsunami warning system. www.data.jma.go.jp/eqev/data/en/tsunami/LessonsLearned_Improvements_brochure.pdfUsed for: The first warning at 14:49 (M7.9, 3 to 6 m), the upgrades to more than 10 m, the final magnitude 9.0 and the 2013 'huge' warnings
  3. National Police Agency of Japan. (2021). Police countermeasures and damage situation associated with 2011 Tohoku district – off the Pacific Ocean Earthquake (March 10, 2021). www.npa.go.jp/news/other/earthquake2011/pdf/higaijokyo_e.pdfUsed for: Dead, missing and buildings destroyed
  4. Ozawa, S., Nishimura, T., Suito, H., Kobayashi, T., Tobita, M., & Imakiire, T. (2011). Coseismic and postseismic slip of the 2011 magnitude-9 Tohoku-Oki earthquake. Nature, 475, 373–376. doi.org/10.1038/nature10227Used for: GPS (GEONET) measurements of how far the land moved, and the slip on the fault about 400 km along the trench
  5. The 2011 Tohoku Earthquake Tsunami Joint Survey Group. 現地調査結果 (Field survey results). In Japanese. coastal.jp/ttjt/index.php?%E7%8F%BE%E5%9C%B0%E8%AA%BF%E6%9F%BB%E7%B5%90%E6%9E%9CUsed for: Tsunami heights and run-up of about 40 m
  6. Minoura, K., Imamura, F., Sugawara, D., Kono, Y., & Iwashita, T. (2001). The 869 Jōgan tsunami deposit and recurrence interval of large-scale tsunami on the Pacific coast of northeast Japan. Journal of Natural Disaster Science, 23(2), 83–88. www.jsnds.org/jnds/23_2_3.pdfUsed for: The Jōgan tsunami, its sand more than 4 km inland, about 1,000 deaths and a recurrence of about 1,000 years
  7. Sawai, Y., Namegaya, Y., Okamura, Y., Satake, K., & Shishikura, M. (2012). Challenges of anticipating the 2011 Tohoku earthquake and tsunami using coastal geology. Geophysical Research Letters, 39, L21309. doi.org/10.1029/2012GL053692Used for: Jōgan deposits along the coast, their recurrence and why the evidence did not anticipate a magnitude 9
  8. Fackler, M. (2011, April 20). Tsunami warnings, written in stone. The New York Times (reprinted by NBC News). www.nbcnews.com/id/wbna42688590Used for: The Aneyoshi stone and where the 2011 wave stopped
  9. International Atomic Energy Agency. (2015). The Fukushima Daiichi Accident: Report by the Director General. IAEA, Vienna. www-pub.iaea.org/MTCD/Publications/PDF/Pub1710-ReportByTheDG-Web.pdfUsed for: The 14 to 15 m wave, the plant's design assumptions and the loss of power and cooling
  10. NASA Jet Propulsion Laboratory. (2011, March 14). Japan quake may have shortened Earth days, moved axis. News release (Richard Gross), as reprinted by Geology.com. geology.com/nasa/japan-earthquakeUsed for: The 1.8 microsecond change in the length of day and the 17 cm shift of the figure axis
  11. Stephenson, F. R., Morrison, L. V., & Hohenkerk, C. Y. (2016). Measurement of the Earth's rotation: 720 BC to AD 2015. Proceedings of the Royal Society A, 472, 20160404. doi.org/10.1098/rspa.2016.0404Used for: The long-term lengthening of the day, about 1.8 ms per century
  12. Japan Science and Technology Agency. Educational activities for disaster prevention backed by scientific evidence behind a 'miracle' (Toshitaka Katada, Kamaishi). www.jst.go.jp/EN/achievements/research/toshitaka_katada2016.htmlUsed for: Kamaishi: 99.8% of about 3,000 students survived
  13. Cabinet Office, Government of Japan. (2013). 平成25年版 防災白書 (White Paper on Disaster Management 2013), Part 1, Introduction. In Japanese. www.bousai.go.jp/kaigirep/hakusho/h25/honbun/1b_0s_01_00.htmUsed for: About 20% of the world's magnitude 6 and larger earthquakes occur around Japan
  14. Headquarters for Earthquake Research Promotion, Earthquake Research Committee. (2025). 南海トラフの地震活動の長期評価（第二版一部改訂） (Long-term evaluation of the Nankai Trough, second edition, partial revision). In Japanese. jishin.go.jp/evaluation/long_term_evaluation/subduction_fault/summary_nankaiUsed for: The 30-year probabilities of 60–90% or more and 20–50%

Images in the episode: USGS, NOAA and NASA (public domain), with maps from NASA/USGS Landsat via NASA GIBS; animations by EarthWatch 24-7. Full credits: earthwatch247.com/credits.html.

Cite this lesson

EarthWatch 24-7. (2026). The Day Japan Moved: The 2011 Tōhoku earthquake and tsunami, from the plates up [Video lesson]. Earth Explained, episode 1. https://earthwatch247.com/learn/japan-2011/

Answer key

  1. C. This is subduction: the Pacific plate, moving toward Japan at about 8 cm a year, bends down at the Japan Trench and sinks beneath Japan.
  2. A. The plates keep converging, but the locked contact drags and squeezes the edge of Japan like a spring being wound. When the strain overcomes friction, the fault slips and the squeezed edge springs back.
  3. D. Japan's historical records describe the 869 tsunami, and geologists digging from about 1990 found its sand far inland, with older sand layers beneath it from even earlier giant waves.
  4. B. The quick magnitude method underestimates very large earthquakes, so the first estimate was 7.9 instead of 9.0. Once sensors at sea detected the real wave, the warnings were raised to more than 10 meters.
  5. A. A tsunami's speed depends on the water's depth. As the front slows in shallow water, the water behind catches up and has nowhere to go but up.
  6. C. The flood knocked out the plant's power, a station blackout, so the reactors could not be cooled. It became the worst nuclear accident since Chernobyl.
  7. B. The locked fault had dragged the coast landward and downward; when it slipped, the overriding plate rebounded toward the trench, and the coast near the rupture sank.
  8. D. The rule was simple and practiced for years: after a big earthquake, run for high ground at once and don't wait. About 99.8% of the city's students survived.
  9. C. Japan changed its warnings so that an underestimated first number can no longer make people think they are safe. Numbers are added once the size of the earthquake is known.
  10. A. A probability is not a date or a certainty. Both values are high for a 30-year window, and the government asks people and local authorities to prepare with the higher one in mind.

Written response

  1. The Pacific plate converges with Japan at about 8 cm a year and subducts at the Japan Trench. Where the plates touch, the fault is locked, so the edge of Japan is dragged down and squeezed and elastic strain builds. In 2011 the strain overcame friction: the fault slipped up to about 50 m over a zone about 400 km long, the squeezed edge sprang back, and the seafloor lurched upward, lifting the whole column of water above it. Gravity pulled that raised water back down, sending waves outward in every direction. In the deep ocean they traveled at about 800 km/h while barely a meter high; in shallow water near the coast they slowed and grew, reaching nearby shores in roughly half an hour and running up about 40 m in places. Across the open Pacific they reached Hawaii in about 7 hours and Chile in about a day.
  2. Seismographs have recorded earthquakes in Japan for only about a century, much shorter than the time between the largest events. Japan's historical records described the 869 tsunami flooding the Sendai plain and drowning about 1,000 people. From about 1990, geologists found its sand, carried in from the sea, kilometers inland, with older sand layers beneath it, showing that giant tsunamis had struck repeatedly, centuries apart. Yet official forecasts were built largely on the better-known recent earthquakes and expected an event of around magnitude 8. The geological work was still estimating how large the Jōgan earthquake had been, and its full lesson, that a much larger section of the fault could rupture at once, had not yet been built into hazard plans when the magnitude 9 struck.
  3. The Aneyoshi stone worked because its rule did not depend on predicting the size of the next wave: every home was built above it, and the 2011 wave stopped about 90 m downhill from it. The Kamaishi drills worked because a simple, practiced rule, run for high ground immediately and don't wait, did not depend on warnings or estimates; about 99.8% of the city's students survived. Seawalls built for earlier tsunamis, and the Fukushima Daiichi site designed for a wave of about 6 m, failed because they were engineered to a fixed design height that the 2011 tsunami, about 14 m at the plant, exceeded. The lesson is that defenses sized to past events can be overtopped, while rules that move people out of reach work whatever the size of the wave; the best plans use both.

Going deeper

  1. How fast does a tsunami cross the Pacific, and why does it grow?: c = √(9.81 × 5,000) ≈ 221 m/s ≈ 800 km/h, the speed in the episode. 5,960 km ÷ 800 km/h ≈ 7.5 hours, close to the observed arrival of about seven hours (the real wave follows the actual seafloor, partly deeper than 5,000 m). Growth: ⁴√(4,000 ÷ 10) = ⁴√400 ≈ 4.5, so the 1 m wave becomes about 4.5 m high, moving at √(9.81 × 10) ≈ 10 m/s, or about 36 km/h, faster than most people can run. Run-up is the height the water climbs on land, not the wave's height offshore: on the Sanriku coast, narrow V-shaped bays funnel the flow and the moving water carries momentum uphill, so run-up there reached about 40 m.
  2. Magnitude, energy and slip: 101.5 × 1.2 = 101.8 ≈ 63 times more energy than the first estimate, which is why the first tsunami forecast was so low. 101.5 × 0.1 = 100.15 ≈ 1.4: the two agencies' values differ by about 40% in energy, small next to the first error. For Mw 9.0, M₀ = 1022.6 ≈ 4.0 × 1022 N·m, so D = 4.0 × 1022 ÷ (4 × 1010 × 4 × 105 m × 1.5 × 105 m) ≈ 17 m; for Mw 9.1, M₀ ≈ 5.6 × 1022 N·m and D ≈ 23 m. The average slip of roughly 17 to 23 m is less than half the peak of 50 to 60 m, which was concentrated near the trench, where it lifted the seafloor most and drove the largest tsunami.
  3. How much warning can early warning give?: S-wave travel time to Tokyo ≈ 376 km ÷ 3.5 to 4 km/s ≈ 94 to 107 s, so the warning time is about 64 to 77 s, about a minute, as in the episode. In 30 s the S waves travel about 105 to 120 km from the hypocenter; at 30 km depth that reaches the surface out to √(105² − 30²) ≈ 100 km to √(120² − 30²) ≈ 116 km from the epicenter. The Oshika Peninsula, about 82 km from the hypocenter, felt S waves after only about 20 to 23 s, before such an alert. In a great earthquake the rupture itself keeps growing for minutes and sends out strong waves from parts of the fault that break later and lie closer to distant cities, so real warning times depend on the whole rupture, not only on its starting point.
  4. The strain budget and the recurrence of Jōgan-type tsunamis: 50 m ÷ 0.083 m/yr ≈ 600 years; 20 m ÷ 0.083 m/yr ≈ 240 years. From 869 to 2011 (1,142 years) about 95 m of convergence accumulated, more than the slip in 2011, so not all convergence is stored as strain on the locked patch: some is released by slow, quiet slip and by the many smaller earthquakes on this plate boundary. Ozawa and colleagues (2011) note the other side of the budget: before 2011, GPS showed strain accumulating off Tōhoku much faster than earlier earthquakes on the plate boundary had released it on average, which raised the question of how that strain would eventually be released. Probabilities: for T = 1,000 years, 1 − exp(−0.03) ≈ 3% in 30 years and 1 − exp(−0.05) ≈ 5% in 50 years; for T = 500 years, about 6% and 10%. A few percent over the life of a power plant or a town is far from negligible when the consequences are catastrophic, which is why rare events need to be part of design, not left out of it.
  5. Did the earthquake really change the length of the day?: ΔI/I = ΔT/T = −1.8 × 10−6 s ÷ 86,400 s ≈ −2.1 × 10−11. The trend of 1.8 ms per century is 1.8 × 10−3 s ÷ 36,525 days ≈ 4.9 × 10−8 s per day per day, so it would cancel the change in about 1.8 × 10−6 ÷ 4.9 × 10−8 ≈ 37 days. Over a year, 365 × 1.8 µs ≈ 0.66 ms. The effect is real but tiny, much smaller than the natural fluctuations in day length caused by winds, ocean currents and Earth's core, which is why it was calculated rather than measured.