Tsunami Physics, Explained

After reading this you will understand why a tsunami crosses an ocean at the speed of a jetliner, how to compute its arrival time and how much it grows as it slows near shore, and how to read the maps and gauge readings the simulator produces.

A tsunami is not a surface wave like the ones surfers ride. It is a disturbance of the entire water column, from the sea floor to the surface, driven by a sudden vertical shift of the ocean bottom during an earthquake. That single fact explains almost everything strange about tsunamis: their enormous speed, their long wavelength, their small height in deep water, and their violent growth at the coast.

What the simulator does and one example

The tool solves the linear shallow-water equations on a rotating sphere over the real ocean floor. When you set off an earthquake, it lifts the sea floor over a fault patch, which pushes up a mound of water. That mound collapses and spreads as a ring of waves that the model then carries across the Pacific, Atlantic or Indian ocean, reflecting off coastlines and reading off heights at more than thirty tide gauges.

Take the 2011 Tōhoku event off Japan. The rupture happened about 130 km offshore at roughly 38°N. The first wave reached the Japanese coast in under 30 minutes, hit Hawaii near Hilo about 7 to 8 hours later, and reached the coast of Chile after roughly 22 hours, having crossed the entire Pacific. The simulator reproduces these arrival times because they follow from one number: the local water depth.

When this model helps and when it does not

Use the simulator to build intuition. It shows you why deep basins act as fast lanes, why mid-ocean ridges bend and focus wave energy into beams, and why a coast that faces the source directly gets hit harder than one shielded behind an island arc. It is a teaching model, and a fairly faithful one for the open-ocean part of the problem.

Do not use this tool to make any real decision. It reports heights offshore and never simulates a wave running up a beach, so it cannot tell you how deep the flooding will be anywhere. If you feel strong shaking near a coast, do not wait for a model or an alert: move to high ground immediately.

The model also stops at 72°S and 72°N and treats the grid edges as absorbing, so waves that would wrap far around Antarctica or through narrow straits below the grid resolution are not tracked well.

The wave speed formula

A wave is called "shallow-water" when its wavelength is much larger than the water depth. A tsunami has a wavelength of 100 to 300 km, and the deep ocean is about 4 km, so the ratio is roughly 50 to 1. In that limit the wave speed depends only on depth:

c = \sqrt{g \cdot h}

Here c is the wave speed in metres per second, g \approx 9.81 is the acceleration of gravity in metres per second squared, and h is the water depth in metres. Nothing about the earthquake or the wave height enters this formula. Depth alone sets the speed.

Put in numbers. Over an abyssal plain at h = 4000 m, the speed is \sqrt{9.81 \times 4000} \approx 198 m/s, which is about 713 km/h, most of the way to the roughly 850 km/h cruising speed of a jet airliner. Over a continental shelf at h = 100 m, the speed drops to \sqrt{9.81 \times 100} \approx 31 m/s, about 113 km/h. In water 10 m deep it is only 9.9 m/s, roughly 36 km/h.

Speed climbs with the square root of depth. Doubling the depth multiplies the speed by about 1.41, not by 2.

Why the wave grows as it slows

As the wave moves from deep water onto the shelf, it slows down. The energy has to go somewhere, so the wave piles up and grows taller. For a wave train crossing gently changing depth, the height H follows Green's law:

H_2 = H_1 \left( \frac{h_1}{h_2} \right)^{1/4}

Here H_1 and H_2 are the wave heights at depths h_1 and h_2. The exponent is one quarter, so the growth is gentle at first and steep only when the depth becomes very small.

Suppose a wave is 0.5 m tall in the deep ocean at 4000 m. When it reaches the 100 m shelf, its height becomes 0.5 \times (4000/100)^{1/4} = 0.5 \times 40^{0.25} \approx 0.5 \times 2.515 = 1.26 m. At 10 m depth it would reach 0.5 \times (4000/10)^{1/4} = 0.5 \times 400^{0.25} \approx 0.5 \times 4.472 = 2.24 m. The final run-up on land is larger still and depends on the shape of the beach, which is exactly the part this model does not compute.

A tsunami barely 0.5 m high in the open sea passes ships unnoticed, because its 200 km wavelength lifts the whole vessel over several minutes. The same wave becomes destructive at the coast because Green's law and the shoaling bathymetry concentrate its energy.

A worked arrival time from the Tōhoku demo

Estimating the crossing time to Hilo

Load the Tōhoku 2011 demo, which sets off the rupture near 38°N, 143°E. Hilo, Hawaii sits about 6200 km away across deep Pacific water. Estimate the travel time by hand, then compare with the gauge.

  1. Take a representative deep-ocean depth of h = 4500 m along the path.
  2. Compute speed: c = \sqrt{9.81 \times 4500} \approx 210 m/s, which is 756 km/h.
  3. Divide distance by speed: 6200 / 756 \approx 8.2 hours.
  4. The real first wave reached Hilo about 7.7 hours after the quake. The estimate is close because most of the path is deep and near-uniform.

The estimate runs about half an hour slow because the real path is deeper than the 4500 m assumed: much of the northwest Pacific basin is 5000 to 6000 m deep, where the wave moves at 800 to 870 km/h. The tool's Hilo gauge records the first wave at 7.7 hours.

The general rule is simple. The arrival time is the sum, along the path, of small path steps divided by the local speed:

T = \int \frac{ds}{\sqrt{g \cdot h(s)}}

Here ds is a step along the wave's path and h(s) is the depth at that point. The simulator does this integral for you by stepping the whole wave field forward in small time steps.

Reading the three map modes

The live view shows the sea surface height as the wave spreads. Blue troughs and red crests alternate. Watch how the ring of waves distorts: it moves fast over deep basins and lags over ridges, so the wavefront bulges and pinches.

The maximum-height map records, for every cell, the largest amplitude that ever passed through during the run. This is where the beams show up. Mid-ocean ridges and undersea plateaus act like lenses that bend the wave and focus its energy along preferred directions, so you see bright streaks reaching thousands of kilometres. A coast that happens to sit at the end of a beam gets a far larger wave than a coast at the same distance off to the side.

The arrival-time bands colour the ocean by the hour the first wave arrives. Each band is one hour wide. Where the bands are packed close together, the wave is slow, meaning shallow water. Where they are spread far apart, the wave is racing across a deep basin.

Rough first-wave arrival times from a Tōhoku-type source
GaugeApprox. distance (km)Model time (h)
Hilo, Hawaii62007.7
Adak, Alaska38005.0
San Diego, USA850010.5
Valparaíso, Chile1730022.4

Try it: depth, speed and shoaling

Over 4000 m of water a tsunami travels at 713 km/h; at 100 m it slows to 113 km/h and, per Green's law, a 0.5 m deep-water wave grows to about 1.26 m. Halving the depth from 4000 m to 2000 m cuts speed only to 504 km/h, because speed follows the square root of depth.

Common mistakes when reading a tsunami model

The first mistake is expecting a single wave. A tsunami is a train of several waves arriving over hours, and the first is often not the largest. Reflections off distant coasts and focusing by ridges can make the third or fourth crest the worst. The maximum-height map, not the first-arrival time, tells you where the danger is.

The second mistake is reading offshore height as flood depth. The gauge heights here are the wave amplitude in open water near the coast. Actual run-up on land is commonly 2 to 4 times larger and depends entirely on local beach shape, which this quarter-degree grid cannot resolve.

The third mistake is trusting magnitude alone. A magnitude 9 earthquake on a steep dip-slip fault that lifts the sea floor makes a large tsunami. A magnitude 9 strike-slip fault that mostly slides sideways lifts very little water and makes a small one. In the simulator, the fault orientation you drag matters as much as the magnitude you set.

Related simulators

If you like watching a field equation evolve on a grid, the River Erosion tool carves landscapes by a related flow rule, and the Plate Tectonics Toy builds the trenches and ridges that make the tsunamis in the first place. For other spreading-process models, see the Forest Fire Model and the SIR Epidemic Simulator, where a disturbance propagates through a medium much as a wavefront does here.

Frequently asked questions

Why does a tsunami travel so fast in the open ocean?

Because its speed is \sqrt{g \cdot h} and the deep ocean is about 4000 m. That gives roughly 200 m/s, or 713 km/h. Ordinary wind waves are much slower because their speed depends on wavelength and they never feel the full ocean depth.

How can the wave be only half a metre tall yet still dangerous?

In deep water the energy is spread through a 4 km column and stretched over a 200 km wavelength, so the surface barely rises. As the wave reaches shallow water it slows and, by Green's law, grows: a 0.5 m open-ocean wave becomes about 1.26 m over the 100 m shelf and much higher again as it runs up the beach.

Does the earthquake magnitude decide the wave height?

It sets the scale of the fault, but the geometry matters just as much. A fault that thrusts the sea floor upward displaces a lot of water; one that slides sideways displaces very little. Two magnitude 9 events can produce very different tsunamis.

Why do some far coasts get hit harder than nearer ones?

Mid-ocean ridges refract the wave and focus its energy into narrow beams. A coast at the end of a beam can see a larger wave than a coast that is closer but off the beam axis. The maximum-height map shows these beams directly.

Can I use these arrival times for planning?

No. This is a teaching model at a coarse resolution, and the footnotes on the tool say so plainly. Use it to understand the physics, and follow official tsunami warnings for anything real.