Why does the earthquake's "rattle"
arrive before the "shake"?
In an earthquake, a small rattling comes first, and a bigger shaking follows a little later. This order is almost always the same. That's because two kinds of waves set off underground at the same instant, and only one of them has arrived first. And this gap of a few seconds is exactly what makes earthquake early warning possible.
When an earthquake hits, it rarely starts with a big jolt. First there's a light rattling, a faint clatter. Sometimes it's no more than dishes softly trembling.
A few seconds after you think "was that an earthquake?", a big shaking arrives. This is the main event. Things falling off shelves, losing your footing — that's usually this second wave.
The farther away the earthquake, the longer the gap between the two. If the gap feels unusually long, the source is far away.
But why does shaking from one single earthquake arrive in two separate waves?
One is a push-pull wave, the other a side-to-side wave. Both leave the same spot at the same moment, but they travel at different speeds.
The push-pull wave arrives first and shakes gently; the side-to-side wave arrives later and shakes hard. That's why the order is always "small, then big."
This time lag isn't random — it comes straight from the physical properties of the ground itself. Let's look at it step by step.
Why is the push-pull wave faster?
A solid like the ground can be distorted in two ways.
- Squeezing or stretching — a distortion that changes volume
- Shearing — a distortion that changes shape only, without changing volume
And materials resist being squeezed strongly, but resist being sheared comparatively weakly. Think of an eraser: hard to crush, but easy to twist sideways.
Wave speed depends on how strong the resistance is and how heavy the material is — stronger resistance means a faster wave, more mass means a slower one. It's the same reason a tightly strung cord gives a higher note (a faster wave) than a slack one.
So the push-pull wave, which relies on the strongly-resisted "squeezing," travels fast, while the side-to-side wave, which relies on the weakly-resisted "shearing," travels slow. This order can never flip.
The order they arrive in comes from how hard the ground resists being distorted.
Here's where the story gets surprising. The side-to-side wave cannot travel through a liquid. Liquids have no "restoring force" that holds their shape, so when sheared they don't spring back — and the wave simply cannot propagate.
In the early 20th century, researchers gathering seismograph records from observatories around the world after major earthquakes noticed something odd. In a certain band on the far side of the globe, the side-to-side wave never arrived at all.
There was only one explanation: the Earth's interior contains a liquid layer. A structure thousands of kilometres underground, which no one had ever seen, was revealed purely from records of ground shaking. It's thanks to this observation that we know the Earth's outer core is liquid.
A gap of seconds makes warnings possible
Earthquake early warning systems rely on exactly this time gap.
A station close to the epicentre catches the faster-arriving push-pull wave and sends that information out as an electric signal. Because an electric signal travels far faster than a wave through the ground, the warning can reach people before the big shaking arrives.
But this method has limits that can't be avoided.
- If the epicentre is close, there's no time. For an earthquake directly below you, the two waves arrive almost together. In a quake right under you, the warning may sound only after the shaking starts
- Because the size is estimated from the first few seconds, the estimate can be revised. Some earthquakes start small and grow, so the initial signal alone can't always reveal the eventual size
- The lead time is, at most, tens of seconds. You need to decide in advance what to do in that window
- Protect your head first, and get lowGet under a sturdy table and hold the legs. If there's no table, move somewhere nothing can fall on you, topple onto you, or slide into you. Stay away from windows, tall furniture and hanging light fixtures.
- Don't rush outside. Don't run to the stoveOutside, falling debris is dangerous. Turn off the stove only after the shaking stops. Most modern gas appliances sense shaking and shut off automatically, so forcing your way to one is riskier than leaving it.
- Once the shaking stops, check you're safe before you actProtect your feet with slippers or shoes (there may be broken glass), and secure an exit route. If anyone is hurt, call 119. In places at risk of a tsunami, head for high ground immediately — don't wait for a warning.
The "gap" tells you the distance
Try counting the seconds from when the rattling starts to when the shaking arrives. Multiply that number of seconds by roughly 8, and you get the approximate distance to the source, in kilometres.
5 seconds means about 40 km; 10 seconds means about 80 km. It's a fairly simple relationship, but it's one that was worked out in 19th-century Japan, and it's still used as a rule of thumb today.
Put the other way round: an earthquake with almost no gap has its source very close by. A sudden, hard upward jolt with no warning rattle usually means the source is directly beneath you, or nearby.
Something you can check at home
- Stretch out a long spring (a coiled toy spring) on the floor, holding one end
- Push the end you're holding quickly back and forth, and you'll see a compressed section travel down the spring. This is the push-pull wave
- Give the same spring (or a rope) a single flick left-right, and you'll see a sideways bulge travel along it. This is the side-to-side wave
- Try tightening or loosening the spring to compare which type travels faster
- Next, flick the side of a tray of water left and right. You'll see that the sideways distortion doesn't travel through it at all
The same spring travels completely differently depending on how you shake it. And in water, the sideways distortion doesn't get through at all — this is exactly the fact that "the side-to-side wave can't pass through a liquid," the same clue that revealed the Earth's interior is liquid. A kitchen tray and the Earth's core are connected by the same piece of physics.
Summary
The rattle arrives first because the ground strongly resists being squeezed, but only weakly resists being sheared. The wave that relies on the stronger resistance travels faster and arrives first. This gap of a few seconds is what makes earthquake early warning possible — and the very same property once revealed the structure of the Earth's interior.
That first rattle is a warning.
Protect your head before you start counting.
After the big shaking stops, the ground itself can weaken in reclaimed land through a process called "liquefaction." We explain how that works in our article on liquefaction. And we cover why small tremors can continue for days after a big earthquake in "Why do aftershocks keep coming for days after a big earthquake?"
Want to know more? ― Terms, formulas, and the textbook connectionWe mark clearly which level each part belongs to, from middle-school science to active research
- MSCovered in middle-school science
- HSCovered in high-school "Basic Physics" / "Basic Earth Science"
- HS+Covered in high-school "Physics" / "Earth Science," or treated as advanced/sidebar content in textbooks
- UnivNot taught in high school — university-level specialist content (seismology, elasticity theory)
- ResearchNot yet settled even at university level — a question researchers are actively investigating
MSTerms: earthquake vocabulary
- P wave: the "push-pull wave" in the main text. Stands for Primary. A longitudinal wave that vibrates along its direction of travel, causing the initial, small tremor.
- S wave: the "side-to-side wave" in the main text. Stands for Secondary. A transverse wave that vibrates at right angles to its direction of travel, causing the main, large shaking.
- Initial tremor duration: the time from when the P wave arrives to when the S wave arrives. Also called the P-S time.
- Seismic intensity vs. magnitude: intensity is how strongly a particular location shakes, and varies from place to place. Magnitude is the size of the earthquake itself — one earthquake has one magnitude.
- Hypocentre and epicentre: the hypocentre is the underground point where the rock first ruptures; the epicentre is the point on the surface directly above it.
HSCheck it with a formula: counting the small shaking gives you the distance
When an earthquake hits, a small rattling comes first, then the big shaking follows. Just by counting the seconds in between, you can work out the distance to the source.
Distance = Speed × Time
| Fast wave (P wave) | about 7 km/s |
| Slow wave (S wave) | about 4 km/s |
| Time gap | duration of the small shaking [seconds] |
Two waves of different speeds leave the same spot at the same instant. The farther they travel, the more their arrival times diverge — just like a fast runner and a slow runner in a marathon pulling apart the longer the race goes.
In other words, the "gap in seconds" is itself a ruler for distance.
Let the distance to the source be D, and the duration of the small shaking be t. The two arrival times are D ÷ 7 and D ÷ 4. Their difference is t. Rearranging gives:
D = t × (7 × 4) ÷ (7 − 4)
| Calculate the numerator | 7 × 4 = 28 |
| Calculate the denominator | 7 − 4 = 3 |
| Divide | 28 ÷ 3 ≈ 9.3 |
| In short | D ≈ t × 9.3 (km) |
The familiar rule of "seconds × 8" is this 9.3 adjusted to match the real structure underground (wave speed changes with depth). Remembering roughly "seconds × 8–9 km" is good enough.
| Small shaking lasted 3 seconds | 3 × 8 = 24 km |
| Small shaking lasted 10 seconds | 10 × 8 = 80 km |
| Small shaking lasted 30 seconds | 30 × 8 = 240 km |
Here's the important part. A long gap does not mean "far away, so safe." If you can still feel shaking from far away, that means the earthquake itself is large.
Conversely, if hard shaking arrives with barely time to count, the source is close. Either way, count only after you've protected yourself.
Early warning works by observing the faster P wave first, then alerting people before the slower S wave arrives. The available lead time comes from the same calculation as in ②.
| Distance to source | assume 100 km |
| Time for P wave to arrive | 100 ÷ 7 ≈ 14 s |
| Time for S wave to arrive | 100 ÷ 4 = 25 s |
| Difference | 25 − 14 = 11 s |
11 seconds. And since it takes a few more seconds to observe and issue the alert, the time you actually have to act is even shorter.
And the closer the source, the smaller this gap gets. At 20 km, it's 20 ÷ 7 ≈ 3 s versus 20 ÷ 4 = 5 s — a difference of just 5 − 3 = 2 seconds. For a quake directly beneath you, the warning cannot beat the shaking. This isn't a technology problem — it's a hard limit set by wave speed itself.
So waiting to think after the alert arrives is too late. Securing furniture in advance, keeping nothing that could fall near where you sleep — preparations made beforehand are what reliably work.
Magnitude is a scale where each step of 1 means about 32 times the energy. It grows by multiplication, not addition.
| Difference between M7 and M8 | 32× |
| Difference between M7 and M9 | 32 × 32 = 1024× |
"M7 and M9 differ by only 2" is misleading — it's about 1000 times the energy. Because the numbers look small, the scale of the gap is easy to underestimate. Knowing what the scale actually means changes how you read the same report.
HSWhere does "seconds × 8" come from?
Let the distance to the source be d, and the speeds of the P and S waves be Vp and Vs. The difference in arrival time is:
t = d/Vs − d/Vp
Solving for d gives d = t × Vp·Vs/(Vp − Vs). Let's plug in numbers.
| P-wave speed | about 8 km/s |
| S-wave speed | about 4 km/s |
| Coefficient Vp·Vs/(Vp−Vs) | 8 × 4 ÷ (8 − 4) = 8 |
| If P-S time is 10 seconds | d ≈ 8 × 10 = 80 km |
※ Actual speeds vary with underground structure, so the coefficient is generally taken to be somewhere between 6 and 8. Use it as a rough guide.
This relationship is known as the Omori formula, named after Fusakichi Omori and colleagues, who studied it in Meiji-era Japan. Working out the distance to the source with nothing but a stopwatch — it's still taught today as a foundation of earth science.
HS+What determines wave speed
The speed of a wave through an elastic body roughly takes the form v = √(elastic modulus / density). The stiffer the material, the faster; the heavier, the slower.
Here, the two wave types draw on different kinds of "stiffness."
- P wave: Vp = √((K + 4μ/3) / ρ) — driven by both K, resistance to volume change, and μ, resistance to shape change
- S wave: Vs = √(μ / ρ) — driven only by μ, resistance to shape change
The expression for the P wave is always larger, so Vp > Vs is guaranteed mathematically. There's no way, in principle, for the order to flip. For typical rock, the ratio is roughly 1.7.
And in a liquid, μ = 0, so Vs = 0. The S wave simply doesn't propagate. Everything said in the main text follows from this one line.
UnivSeeing through the Earth with shake records
Seismic waves bend and reflect at boundaries where underground material changes — the same way light bends through glass. Using this property, the arrival times of waves at stations worldwide can be used to work backward to the speed structure inside the Earth.
One of the earliest major discoveries was the S-wave "shadow." In the 1910s, researchers found a region — roughly 103 to 143 degrees from the epicentre, measured as an angle — where S waves never arrived at all, confirming the existence of the liquid outer core. This was followed by the discovery of the (solid) inner core in 1936 by Inge Lehmann.
Today, seismic tomography combines data from many earthquakes and stations to work out three-dimensional speed structure. It's used to trace the path of subducted plates and estimate mantle flow — the same basic idea as a medical CT scan.
ResearchWhat's still unknown
- Earthquake prediction still doesn't work. No scientifically established method exists for pinpointing "when, where, and how large" before an event happens. This has been studied for decades in Japan and elsewhere, but the current consensus is that short-term prediction hasn't become practical. Given that prediction isn't possible, effort goes instead into preparation beforehand and information right after the fact (early warning).
- Whether an earthquake's start determines its final size is unknown. Big and small earthquakes often start with very similar-looking waveforms. There are two competing views — that whether a small start grows into a big quake is essentially random, or that the final size is already determined at the moment it starts — and the question remains unresolved. This bears directly on the accuracy of early warning. If the second view is correct, it would become possible to estimate the eventual size more accurately from the very first signals.
- The meaning of "slow slip" events is also still being researched. In the 2000s, researchers discovered slow slipping along plate boundaries that unfolds over days to months, along with faint tremors that accompany it. Their relationship to major earthquakes draws a lot of attention, but whether they act as a precursor remains unconfirmed.
- Predicting long-period shaking is also difficult. High-rise buildings far from the epicentre can sway strongly, and this depends heavily on the structure of underground sediment layers. Because our knowledge of underground structure is limited, uncertainty in these predictions remains.
Connection to textbooks (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| MS | Science — how earthquake shaking travels / initial tremor duration | P and S waves, intensity vs. magnitude, seconds × 8 |
| HS | Basic Physics — waves (longitudinal and transverse) / Basic Earth Science — Earth's interior | Difference between the two waves, deriving the Omori formula |
| HS+ | Physics — elasticity and wave speed (often treated as advanced) | v = √(elastic modulus/density), why S waves can't pass through liquid |
| Univ | Seismology, elasticity theory, geophysics | S-wave shadow, discovery of outer and inner core, tomography |
| Research | Seismology (unresolved) | Limits of prediction, rupture growth, slow slip |
| ― | Disaster-preparedness education | Protect your head, don't go outside, act once shaking stops |
- Japan Meteorological Agency (気象庁) explanatory materials on earthquakes and on the workings and limits of earthquake early warning.
- Shearer, P. M., Introduction to Seismology (a standard seismology textbook).
- Lehmann, I., P', Publications du Bureau Central Séismologique International, 1936 (discovery of the inner core).
- Obara, K., Nonvolcanic deep tremor associated with subduction in southwest Japan, Science 296, 1679–1681, 2002 (discovery of deep low-frequency tremor).
- Materials from Japan's Cabinet Office (内閣府) and local governments on earthquake response and preparedness.
※ Seismic wave speeds and coefficients vary with underground structure. This article presents commonly used approximations.
※This article is a general-audience science explainer. In the event of an earthquake, follow the instructions of the Japan Meteorological Agency, fire services, local government, and your area's disaster-preparedness plan. In areas at risk of tsunami, evacuate to high ground immediately without waiting for a warning. The figures given here are approximations meant to illustrate the underlying mechanism.