Earthquakes are measured with networks of seismometers. Each station records ground motion over time as a seismogram. Software combines observations from multiple stations to estimate when and where rupture began, how deep it was and how large the source was. These are calculated solutions, so early reports can change.
Quick answer
- A sensor records motion in one or more directions.
- Processing identifies seismic phases, especially P and S arrivals.
- Arrival times from many stations constrain origin time and hypocenter.
- Amplitudes, durations or full waveforms provide magnitude estimates.
- Automatic and analyst-reviewed solutions are published and may be revised.
What instruments measure
A seismometer is the sensing instrument. A seismograph is the recording system, and a seismogram is the record it produces. Digital stations sample motion continuously, attach precise time and transmit records to a network center.
The trace contains earthquake signals, other natural motion, human activity and instrument noise. Analysts and algorithms identify phases by their timing, frequency, polarization and coherence across stations. P waves normally arrive before S waves. Surface waves often arrive later and can dominate long-period motion.
Learn how P, S and surface waves differ →
From wave arrivals to a location
A single station can estimate distance from the gap between P and S arrivals, but not a unique direction. A network compares observations from several stations with predicted travel times through a velocity model.
The locator varies four main unknowns: latitude, longitude, depth and origin time. It seeks the combination with the smallest useful mismatch between observed and predicted arrivals. Station geometry matters: a source surrounded by nearby instruments is usually better constrained than one far outside a network.
Depth is often the weakest coordinate because depth, velocity structure and origin time can trade off. Special reflected phases can improve a suitable event. The earthquake depth guide describes those limits.
How magnitude and review work
Magnitude is calculated after the source and distances are estimated. ML uses calibrated local amplitudes; mb and Ms use selected body or surface waves; Mw estimates seismic moment from broader waveform or source analysis. A network can publish several values before selecting a preferred one.
An automatic solution is machine-produced before analyst confirmation. A reviewed solution has received human examination, but it is not guaranteed never to change. New stations, corrected timing, better phase picks, a different event classification or merged duplicate records can revise a catalog entry.
Limits and common mistakes
The nearest point on a public map is not necessarily the nearest station. Marker precision should not be read as exact fault geometry. A hypocenter is the rupture start, not the full ruptured surface.
More decimal places do not remove uncertainty. Network operators can publish uncertainty fields, but compact feeds may omit them. “Reviewed” means examined, not immutable. Earthquake24 displays source reports and their timestamps rather than claiming independent detection.
Common questions
Can one seismometer locate an earthquake?
It can estimate aspects such as direction or distance under some conditions, but a robust hypocenter normally uses arrivals from a network.
How quickly is a magnitude available?
Automatic estimates can appear rapidly. The time varies with station coverage, telemetry and processing; more complete values can follow later.
Why are two agencies different?
They may use different stations, velocity models, phase picks, magnitude methods and review timing. Convergence can improve as data are exchanged.