The problem early warning solves
Earthquakes strike without obvious precursors. Unlike hurricanes that build over days, a fault rupture transitions from quiet to violent in under a second. That leaves zero preparation time for anyone near the epicenter unless technology can outrun the shaking.
Earthquake early warning systems exploit one physical fact: electronic signals travel far faster than seismic waves. A sensor near the fault can detect the initial P-wave, estimate magnitude and location, and broadcast an alert that reaches distant cities before the destructive S-wave and surface waves arrive.
P-waves versus S-waves: the detection window
When a fault ruptures, it releases energy as two primary body wave types. P-waves are compressional, travel at roughly 6 kilometers per second through crustal rock, and arrive first. S-waves are shear waves, travel at about 3.5 kilometers per second, and carry most of the destructive energy.
The gap between P and S arrival times grows with distance from the epicenter. At 100 kilometers, the difference is roughly 15 seconds. At 200 kilometers, about 30 seconds. An EEW system races to identify the P-wave, classify the event, and push an alert before the S-wave arrives at the target population.
How ShakeAlert works in the United States
ShakeAlert is the operational EEW system covering the West Coast of the United States, operated jointly by the USGS, Caltech, UC Berkeley, the University of Washington, and the University of Oregon. It uses a dense network of over 1,600 seismic stations along the Pacific plate boundary.
When multiple stations detect P-wave arrivals consistent with a single source, ShakeAlert's algorithms estimate the event's location, magnitude, and expected shaking intensity at surrounding points. If the estimated shaking exceeds a threshold, an alert message propagates through the Wireless Emergency Alert system, smartphone apps, and institutional feeds.
- Detection to alert latency: typically 5–15 seconds depending on station density near the epicenter
- Coverage: California, Oregon, Washington — roughly 50 million people
- Alert distribution: WEA, MyShake app, ShakeAlert-integrated building systems
- Limitations: the blind zone near the epicenter may receive the alert after shaking has already begun
Japan's JMA system: the global benchmark
Japan's Meteorological Agency has operated its EEW system since 2007, making it one of the most mature deployments in the world. The system uses over 4,000 seismic stations — a density unmatched by any other country — which allows initial detection within 2–3 seconds of rupture onset.
JMA EEW alerts reach the public through television broadcasts, smartphone push notifications, railway automatic braking systems, factory shutoff triggers, and municipal loudspeakers. The 2011 Tohoku earthquake demonstrated both the system's strength and its limits: coastal cities received 15–60 seconds of warning, but the initial magnitude estimate underestimated the event's true scale.
What happens in those few seconds
Seconds of warning may sound trivial, but they enable specific protective actions. Individuals can drop, cover, and hold. Surgeons can withdraw instruments. Elevators can stop at the nearest floor and open doors. Gas valves can shut automatically. Fire station doors can open before frames warp.
Industrial systems gain the most value. A natural gas pipeline network that receives 10 seconds of warning can isolate segments and prevent post-quake fires. A rail network that receives 8 seconds can begin emergency braking and prevent derailments. These automated responses are pre-programmed — the system does not wait for human judgment.
The blind zone problem
No EEW system can warn the area closest to the epicenter. The alert propagation time plus the processing latency means there is always a radius — typically 20–30 kilometers — where shaking arrives before the alert. This is called the blind zone.
Reducing the blind zone requires denser station networks, faster telemetry, and faster algorithms. On-site warning approaches, where a single sensor at the user's location estimates the incoming shaking from the very first P-wave cycles, can supplement network-based systems but trade accuracy for speed.
Newer approaches: MEMS sensors and smartphones
Traditional seismic stations use high-precision broadband instruments that cost thousands of dollars each. A newer approach supplements these with MEMS accelerometers — the same sensors inside smartphones and IoT devices — deployed at much higher density for much lower cost.
Google's Android Earthquake Alerts System turns every Android phone into a potential seismic sensor. When multiple phones in a region detect shaking simultaneously, the system generates an alert. This has already issued warnings in countries without any traditional EEW infrastructure, including Turkey, the Philippines, and parts of Central America.
False alerts and the trust tradeoff
Every EEW system faces a fundamental tradeoff between speed and accuracy. Issuing an alert quickly based on minimal data reduces warning time loss but increases the chance of false or overestimated alerts. Waiting for more sensor confirmations improves accuracy but costs precious seconds.
A single false alert in a densely populated area can cause panic, injuries during unnecessary evacuations, and erosion of public trust. Systems must calibrate their thresholds so that missed detections are rare but false positives do not become routine. Most operational systems now use tiered alerting: a fast preliminary estimate followed by rapid updates as more data arrives.
Countries building new systems
Beyond the US and Japan, operational or developing EEW systems exist in Mexico (SASMEX, operational since 1991 for the Mexico City corridor), Taiwan (CWA), South Korea (KMA), China (ICL), Turkey (AFAD pilot), Italy (INGV pilot), and several others. Each faces different challenges based on tectonic setting, station density, population distribution, and communication infrastructure.
The global trend is toward integration of traditional seismometers, MEMS networks, GNSS displacement sensors, and smartphone crowdsourcing into hybrid detection systems that can issue faster, more accurate warnings across diverse geographies.
What EEW cannot do
Early warning systems do not predict earthquakes. They detect them after rupture has begun and race to deliver alerts faster than the destructive waves travel. They cannot tell you an earthquake will happen next Tuesday. They cannot prevent damage at the epicenter. They cannot substitute for structural engineering, land-use planning, or community preparedness.
Their value is narrowly defined but deeply impactful: converting zero warning time into a few seconds of actionable notice. For populations in seismically active zones, those seconds are the difference between being under a desk or standing next to a window when the shaking hits.