An earthquake doesn't announce itself all at once. It radiates energy outward from its source as several distinct kinds of seismic wave, travelling at different speeds — and one type reliably outruns the others. Earthquake early warning systems exist entirely because of that gap: by detecting the fast, mostly harmless wave the instant it arrives, a system can issue a warning before the slower, damaging wave shows up.
P-waves arrive first, S-waves do the damage
Earthquakes generate primary waves (P-waves) and secondary waves (S-waves), which travel outward from the rupture at different speeds through the Earth's crust. P-waves are compressional, pushing and pulling the ground back and forth in the direction they're travelling, and they move fastest — typically around six kilometres per second through crustal rock — but they're generally weak and rarely cause significant damage on their own. S-waves move the ground side to side, perpendicular to their direction of travel, at roughly half the speed of P-waves, and it's this slower, shaking motion that does most of an earthquake's structural damage. Because P-waves consistently arrive before S-waves, and the gap between them grows the further you are from the earthquake's origin, that gap is a genuine, physically guaranteed head start.
Turning a physics fact into seconds of warning
Earthquake early warning networks place sensitive seismometers across a region and use them to detect the P-wave the instant it arrives at any given station, then rapidly estimate the earthquake's location and likely magnitude from that early signal, before the more damaging S-wave has propagated outward. That estimate gets sent out as an alert to more distant locations, arriving via radio or cellular networks that travel far faster than the seismic waves themselves — giving people, trains, elevators, and automated systems anywhere from a few seconds to, at greater distances from the epicentre, up to a minute or more of advance warning before the shaking actually arrives. It's not a prediction of when an earthquake will happen; it's a race between an electronic signal and a slower physical wave, run only after the earthquake has already begun.
What we're still unsure about
The physics behind the P-wave and S-wave speed difference is completely settled, and operational early warning systems in places like Japan, Mexico, and the western United States have demonstrably given real, useful seconds of notice during actual earthquakes. What remains genuinely difficult is the trade-off inherent in the method itself: locations very close to an earthquake's epicentre get almost no warning at all, because the P-wave and S-wave haven't had time or distance to separate, which means the warning is systematically weakest exactly where the shaking will often be strongest — a limitation seismologists describe as fundamental to the physics, not something better sensors or faster processing can fully eliminate.
This sits inside Natural Hazards & Disaster Risk Reduction, one of seven topics in Environmental Geography, one of five domains in Geography, one of seventeen subjects the app can quiz you on.