What Decides How Fast Earthquake Waves Travel Through Rock?

1. Earthquake Waves Are Much Faster Than a Train or a Sound

Earthquakes send two main kinds of waves through the ground. In rock, the faster P wave and the slower S wave have a fairly steady relationship: the U.S. Geological Survey (USGS) says S waves generally travel about 60% as fast as P waves [1]. A common rough pair of numbers is about 6 km/s for P waves and about 4 km/s for S waves, although the real values change from place to place [3].

Put those next to things you know. A bullet train running at 300 km/h covers about 83 meters per second. Sound in air at room temperature covers about 340 meters per second. A P wave at 6 km/s is therefore about 70 times faster than that train and about 18 times faster than sound in air (my own arithmetic from the figures above). Even the slower S wave at 4 km/s is nearly 50 times faster than the train.

That is why an earthquake that began 100 km away can reach you in well under a minute. Which raises the question this article follows: what sets the speed, and why isn't it the same everywhere?

2. Stiffer and Lighter Rock Carries Waves Faster

A wave is a disturbance that passes from one bit of material to the next. Two things matter. One is how strongly the material pushes back when it is squeezed or twisted, which is its stiffness. The other is how heavy it is for its size, its density, because heavier material is harder to get moving. Think of a row of balls joined by springs: strong springs pass a push along quickly, while heavy balls make it sluggish. That picture is only an analogy.

Physics gives this a precise form. For P waves the speed is the square root of a stiffness term divided by the density. The stiffness term combines resistance to squeezing (the bulk modulus) and resistance to twisting (the shear modulus). S waves involve only the twisting term [3]. Because the P wave's stiffness term includes the squeezing part as well, a P wave outruns an S wave in the same rock.

In practice, stiffness usually wins over density. Granite, a hard rock, carries P waves at about 5.8 to 6.1 km/s [3]. Loose, crumbly material is much less stiff, so waves crawl through it. This is why speed is such a useful clue: if scientists measure how quickly waves cross a region, they learn something about what the ground is made of, even where nobody can dig.

3. A Sudden Jump in Speed Revealed the Bottom of Earth's Crust

Because rock deeper down is squeezed and different in kind, waves generally travel faster the deeper they go. In 1909 that fact produced a surprise. The Croatian seismologist Andrija Mohorovičić was studying seismograms from earthquakes that began fairly near the surface. He noticed that each recording showed two sets of P waves and S waves. One set had traveled along a direct path near the surface. The other had been bent (refracted) by a layer where waves move faster [2].

At greater distances the second set could arrive first, even though it had taken a longer route. Think of a driver choosing to leave a slow, winding side road to take a longer way along a fast highway: the extra distance pays off once the trip is long enough. The waves' arrival times therefore gave away a hidden layer boundary deep underground.

That boundary is now called the Mohorovičić discontinuity, or "Moho" for short. It separates Earth's crust from the mantle below. Just above it, P waves move at about 6.7 to 7.2 km/s, similar to basalt. Just below it, they move at about 7.6 to 8.6 km/s, closer to what is found in peridotite or dunite, denser mantle rocks. The Moho lies about 5 to 10 km below the ocean floor and about 20 to 90 km below typical continents, with an average of roughly 35 km [2].

The discovery is a good example of how science sometimes works: nobody saw the boundary directly. It was found from the timing of waves alone.

4. Near the Surface, Soft Ground Slows the Waves and Makes the Shaking Stronger

The speed-up with depth stops near the top. The ground we build on is often loose soil or sediment, where waves move far more slowly than in solid rock below. A USGS report on the 1994 Northridge, California, earthquake puts it this way: slower wave velocities imply softer ground and, therefore, larger site amplifications. It adds that softer sediments in a valley usually shake harder than hard-rock sites in nearby mountains [4].

It may seem odd that slower waves can mean stronger shaking. The sources link softer ground to larger amplification, but they do not give a simple reason that fits in one sentence. How much the shaking grows depends on the ground's properties and on how its layers are arranged, not on wave speed alone.

The report gives numbers. Calculations from shear-wave speeds alone predicted a factor of about 2 difference in amplification between sites, while aftershock recordings showed differences as large as 6 [4]. That gap tells us speed is only part of the story: the shape and depth of the layers beneath a site, even those deeper than 100 meters, also matter. Another USGS study measured wave speeds in the top 30 meters at 20 sites in the Los Angeles area. It reported that areas with heavy building damage were associated with high ground-motion amplification [5].

So "how fast do the waves arrive?" and "how hard will the ground shake?" are two different questions. The first depends on the distance from the earthquake and on the speeds of the layers the waves pass through. The second depends a great deal on the slow, soft ground right under your feet. Buildings on soft soil and buildings on rock a block or two away can fare very differently, which the USGS report notes can happen even when the construction is similar [4].

5. What the Speed Lets Us Work Out, and Where the Estimate Breaks Down

If you know a wave's speed, you can estimate its travel time: distance divided by speed. At 100 km from the source, an S wave at 4 km/s takes about 25 seconds, and a P wave at 6 to 7 km/s takes roughly 14 to 17 seconds. These are my own back-of-the-envelope numbers.

They are only rough, because real Earth is layered, with speeds that change with depth. Waves do not travel in straight lines; they curve toward faster layers. The USGS travel-time table, built from a standard Earth model, gives about 17.7 seconds for a P wave to cover 1 degree of arc (about 111 km) and 13.5 seconds more for the S wave to follow [6]. That works out to an average P speed near 6.3 km/s over that distance, a little different from the simple round number. The farther the trip, the more deep, fast rock the waves use, and the more a straight-line estimate drifts.

6. Try It: Estimate the Arrival Time, and Where to Read More

Pick a recent earthquake from the USGS earthquake map (earthquake.usgs.gov) and note its distance from where you live. Divide that distance by 4 km/s to estimate how many seconds the S wave would take to reach you. Add that time to the earthquake's start time to estimate its arrival time. Then compare two quakes at different distances. Treat the answer as a rough guide, since the layers underground will bend the real result away from it.

If you want to see ground differences yourself, look for a hazard map of your area that shows soil type; many geological surveys publish them. For deeper reading, start with the USGS Northridge report [4] and the USGS travel-time table [6]. A question to carry with you: if soft soil makes shaking worse, how might engineers design buildings that stand on it?

Sources

Revision history

October 23, 2026: First published.