Seismology: How Earthquakes Are Detected and Measured

Seismology sits at the intersection of physics, geology, and engineering, providing the foundational science behind every earthquake alert, hazard map, and building code on the planet. From the first crude instrument designed to detect distant tremors in ancient China to today's globe-spanning digital networks that pinpoint earthquake locations within seconds, the field has undergone a remarkable transformation — yet the core challenge remains the same: sensing, recording, and interpreting the waves that radiate outward when rock breaks beneath the surface.

This article traces the history and mechanics of seismology, explains how modern instruments work, describes the global networks that monitor earthquakes around the clock, and explores the cutting-edge technologies poised to revolutionize earthquake detection in the coming decades.

Understanding seismology is essential not only for scientists but for anyone living in earthquake-prone regions. The data seismologists collect drives everything from earthquake magnitude calculations to real-time alert systems that protect communities worldwide.

What Is Seismology?

Seismology is the scientific study of earthquakes and the propagation of elastic waves through the Earth. The term derives from the Greek seismos (earthquake) and logos (study). While most people associate seismology with earthquake monitoring, the field encompasses a much broader range of inquiry.

Seismologists study the generation and propagation of seismic waves — the energy released when rocks fracture along faults, when volcanoes erupt, when meteorites impact the surface, or even when large explosions occur. By analyzing how these waves travel through Earth's interior, seismologists have mapped the planet's internal structure, identifying the crust, mantle, outer core, and inner core long before any drill could reach those depths.

The discipline divides broadly into two branches. Earthquake seismology focuses on understanding where, why, and how often earthquakes occur, developing hazard assessments, and improving early warning capabilities. Exploration seismology uses artificially generated seismic waves to image subsurface structures for oil, gas, mineral exploration, and engineering site characterization. Both branches depend on the same fundamental physics: measuring how waves move through rock and other materials.

Seismology also provides critical data for nuclear test monitoring. The Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) operates a global network of seismic stations specifically to detect clandestine nuclear explosions, distinguishing their signatures from natural earthquakes.

A Brief History of Seismology

The Ancient Origins

The earliest known instrument designed to detect earthquakes was the seismoscope invented by Chinese polymath Zhang Heng in 132 AD during the Han Dynasty. His device, called the Houfeng Didong Yi (roughly "instrument for measuring the seasonal winds and the movements of the Earth"), was a large bronze vessel approximately two meters in diameter. Eight dragon heads arranged around the exterior each held a bronze ball in its mouth. When seismic waves arrived, an internal pendulum mechanism displaced one of the balls, which dropped into the open mouth of a corresponding bronze toad below, indicating the direction from which the earthquake originated.

According to the historical text Book of the Later Han, the instrument successfully detected an earthquake in Longxi (roughly 600 km from the capital Luoyang) in 138 AD — an event that was only confirmed by messengers days later. While the exact internal mechanism remains debated by historians, Zhang Heng's device represents the first documented effort to instrumentally detect seismic activity.

For centuries afterward, earthquake observation remained largely anecdotal. European natural philosophers in the 17th and 18th centuries debated whether earthquakes were caused by underground fires, compressed air, or chemical explosions. The 1755 Lisbon earthquake — which devastated the Portuguese capital and generated a catastrophic tsunami — spurred the first systematic scientific investigations. John Michell, an English clergyman and natural philosopher, published a paper in 1760 proposing that earthquakes were caused by waves of elastic compression traveling through the Earth, a remarkably prescient insight.

The Birth of Modern Seismology (1850s–1900s)

The mid-to-late 19th century saw the emergence of seismology as a quantitative science. Robert Mallet, an Irish engineer, conducted the first controlled seismic experiment in 1849, detonating gunpowder charges and measuring the speed of the resulting waves through sand and rock. He coined the terms "seismology," "epicenter," and "isoseismal" and produced the first seismicity map of the world in 1857.

The pivotal advances occurred in Japan. After the devastating 1880 Yokohama earthquake, British scientists John Milne, James Alfred Ewing, and Thomas Gray — all working at the Imperial College of Engineering in Tokyo — developed the first modern seismographs. Milne's horizontal pendulum seismograph, refined throughout the 1880s and 1890s, became the standard instrument deployed worldwide. By 1900, Milne had established a global network of approximately 30 stations reporting to a central observatory on the Isle of Wight in England.

In 1897, Richard Dixon Oldham identified the distinct arrivals of P-waves (primary), S-waves (secondary), and surface waves on seismograms, establishing the wave classifications still used today. His 1906 paper demonstrating that S-waves do not penetrate Earth's core provided the first seismological evidence for a liquid outer core.

The 20th Century: Networks and Nuclear Testing

The 20th century saw explosive growth in seismological capability. Beno Gutenberg and Charles Richter at Caltech developed the first widely adopted magnitude scale in 1935, providing a standardized way to compare earthquake sizes. The Richter scale (technically the local magnitude scale, ML) became a household term, though it has since been largely superseded by the moment magnitude scale (Mw) for scientific use.

The Cold War transformed seismology. The need to monitor Soviet nuclear tests drove massive investment in seismic instrumentation. The World-Wide Standardized Seismograph Network (WWSSN), established in the early 1960s with over 120 stations, represented a quantum leap in global coverage and standardization. This network not only detected nuclear tests but also provided the seismological data that helped confirm plate tectonics.

The shift from analog to digital recording in the 1970s and 1980s fundamentally changed the field. Digital broadband seismometers — capable of recording ground motion across a wide range of frequencies — replaced the narrow-band analog instruments of earlier decades. These instruments, combined with satellite communications and computer processing, enabled the real-time seismic monitoring networks that operate today.

How Seismographs Work

The Fundamental Principle

A seismograph (the recording system) relies on a seismometer (the sensor) to detect ground motion. The underlying principle is inertia: a mass suspended on a spring or pendulum tends to remain stationary while the frame holding it moves with the ground during an earthquake. The relative motion between the stationary mass and the moving frame is what the instrument measures.

Consider a simple thought experiment: hang a heavy weight from a spring attached to a rigid frame bolted to the ground. When an earthquake shakes the frame, the weight's inertia causes it to lag behind the frame's motion. This relative displacement — frame moving while the mass stays relatively still — can be converted into a measurable signal.

Seismograph vs. Seismometer vs. Seismogram

These three terms are often confused but have distinct meanings:

• Seismometer: The sensor itself — the physical device that detects ground motion. Modern seismometers are electromagnetic instruments where a coil moves relative to a magnet (or vice versa), generating an electrical signal proportional to ground velocity.
• Seismograph: The complete recording system, including the seismometer, data acquisition electronics, timing system (usually GPS-synchronized), and data storage or transmission equipment.
• Seismogram: The recorded output — the actual waveform data showing ground motion over time. Historically this was a paper trace; today it is a digital time series.

Modern Broadband Seismometers

Today's workhorse instruments are broadband seismometers, which record ground motion across a very wide frequency range — typically from 0.003 Hz (periods of 360 seconds) to 50 Hz or higher. This broad bandwidth allows a single instrument to record everything from slow tectonic deformation signals to the sharp, high-frequency arrivals of nearby earthquakes.

The Streckeisen STS-2, Nanometrics Trillium, and Güralp CMG-3T are among the most widely deployed broadband seismometers in global networks. These instruments use force-feedback electronics: rather than allowing the mass to swing freely, a feedback system applies a force to keep the mass nearly centered, and the applied force — which mirrors the ground motion — becomes the output signal. This approach dramatically extends the frequency range and dynamic range compared to passive instruments.

Modern broadband seismometers can detect ground displacement as small as fractions of a nanometer while also recording the violent shaking from a nearby large earthquake without going off-scale — a dynamic range exceeding 140 dB (a factor of 10 million in amplitude).

Strong-Motion Accelerometers

For engineering applications and near-source recording of large earthquakes, strong-motion accelerometers complement broadband seismometers. While broadband instruments measure ground velocity and are optimized for sensitivity, accelerometers measure ground acceleration and are designed to stay on-scale during violent shaking that would clip a broadband sensor.

Accelerograph data directly informs building codes, providing the peak ground acceleration (PGA) and spectral acceleration values that engineers use to design earthquake-resistant structures. Networks of strong-motion instruments — such as the USGS National Strong-Motion Project and Japan's K-NET and KiK-net — have recorded thousands of earthquakes, building the empirical ground-motion databases that underpin modern seismic hazard analysis.

Global Seismic Networks

Continuous earthquake monitoring depends on networks of seismic stations distributed across continents and ocean islands. Each network serves specific purposes, from global earthquake location to regional hazard monitoring and nuclear test detection.

Major Networks

IRIS — Incorporated Research Institutions for Seismology USGS ANSS

The IRIS Global Seismographic Network

The IRIS GSN represents the backbone of global seismology. Its approximately 150 stations, uniformly distributed around the globe, provide the data used to locate virtually every earthquake of magnitude 5.0 or larger worldwide. Each station hosts a broadband seismometer, a strong-motion accelerometer, and associated data acquisition and telemetry systems, typically in a vault or borehole to minimize noise from wind, temperature changes, and human activity.

GSN stations transmit data in near-real-time via satellite to data centers where automated algorithms continuously scan for earthquake signals. The network's global coverage ensures that any significant earthquake anywhere on Earth is recorded by multiple stations from different directions — essential for accurate location determination.

How Regional Networks Add Density

While global networks provide worldwide coverage, they lack the station density needed to locate small earthquakes or provide rapid alerts in specific regions. Regional networks fill this gap. In California, for example, the ANSS integrates data from the California Integrated Seismic Network (CISN), which combines stations operated by the USGS, Caltech, UC Berkeley, and other institutions. This dense instrumentation allows the network to locate earthquakes as small as magnitude 1.0 and to issue ShakeAlert warnings within seconds of rupture initiation.

Japan operates the densest seismic network on Earth. NIED's Hi-net alone includes approximately 800 borehole stations at depths of 100-200 meters, spaced roughly 20 km apart across the entire country. This extraordinary density enables detection of extremely small earthquakes and provides the high-resolution data that drives Japan's world-leading earthquake early warning system.

How Earthquakes Are Located

The P-Wave / S-Wave Method

Locating an earthquake — determining its epicenter (the point on the surface directly above the rupture) and hypocenter (the actual point of rupture at depth) — relies on the difference in arrival times of P-waves and S-waves.

P-waves travel faster than S-waves — roughly 6-8 km/s versus 3.5-4.5 km/s through crustal rock. This means P-waves always arrive first at a seismic station, followed by S-waves. The time difference between the two arrivals (the S-P interval) is directly proportional to the distance between the station and the earthquake.

A single station can determine how far away an earthquake occurred, but not in which direction. With two stations, the possible location is narrowed to two points where the distance circles intersect. A minimum of three stations is needed to uniquely determine an epicenter by triangulation — finding the single point consistent with all three S-P distance estimates.

In practice, modern networks use data from dozens or even hundreds of stations simultaneously, employing least-squares inversion algorithms to find the location (latitude, longitude, depth) and origin time that best fits all the observed arrival times. Depth determination requires either a station very close to the epicenter or the identification of depth phases — waves that reflect off the Earth's surface near the epicenter before traveling to the recording station.

Real-Time Processing

When the USGS detects an earthquake, the process from first P-wave detection to a published location typically takes just minutes. Here is the general sequence:

1. P-wave detection: Automated algorithms at each station detect the onset of a P-wave above the background noise level. A short-term average / long-term average (STA/LTA) algorithm is the most common trigger method.
2. Association: A central processing system collects P-wave detections from multiple stations and determines whether they are consistent with a single earthquake source using arrival time patterns.
3. Preliminary location: Using the associated P-wave arrivals, an initial hypocenter and origin time are calculated. This preliminary location is often available within 2-4 minutes for well-instrumented regions.
4. Magnitude estimation: Initial magnitude is calculated from the amplitude and frequency content of the recorded waves. Multiple magnitude types may be computed; for significant earthquakes, the moment magnitude (Mw) is determined through waveform modeling.
5. Refinement: As more data arrive — including S-wave picks, surface wave data, and reports from more distant stations — the location and magnitude are refined. For significant earthquakes, the USGS issues updates over hours and days as analysis continues.

The USGS locates approximately 20,000 earthquakes per year — an average of about 55 per day. Of these, about 100 are large enough to cause damage (magnitude 6.0 or larger), and approximately 15 are classified as major (magnitude 7.0+). The agency publishes earthquake information on its Earthquake Hazards Program website, providing real-time maps, ShakeMaps showing estimated ground shaking intensity, and information pages for significant events.

Earthquake Early Warning Systems

Earthquake early warning (EEW) systems represent one of the most consequential applications of seismology. These systems exploit the fact that electronic signals travel much faster than seismic waves. By detecting the initial P-waves from an earthquake and rapidly estimating its location and magnitude, an EEW system can issue alerts to areas that have not yet experienced strong shaking.

The warning time depends on the distance from the epicenter. Areas close to the fault may receive only a few seconds of warning — or none at all — while areas 50-100 km away may get 10-30 seconds. Even a few seconds allows automated responses: slowing trains, opening fire station doors, shutting down gas lines, and triggering personal protective actions like "Drop, Cover, and Hold On."

How EEW Works

1. Detection: Dense seismic networks detect the initial P-wave from an earthquake.
2. Rapid characterization: Algorithms analyze the first 1-3 seconds of the P-wave to estimate the earthquake's location, depth, and magnitude.
3. Alert generation: If the estimated magnitude exceeds a threshold and predicted shaking at target areas exceeds a minimum intensity, an alert is generated.
4. Dissemination: Alerts are broadcast via dedicated communication channels, cell phone wireless emergency alerts, mobile apps, and automated systems.
5. Shaking arrives: Strong S-waves and surface waves arrive after the alert.

Worldwide EEW Systems

ShakeAlert — Earthquake Early Warning for the West Coast Japan Meteorological Agency EEW

ShakeAlert: The US System

The United States' ShakeAlert system, developed by the USGS in partnership with university partners and state agencies, became publicly operational in stages: California in October 2019 (via the MyShake app), Oregon in March 2021, and Washington in May 2021. The system was integrated with the Federal Emergency Management Agency's (FEMA) Wireless Emergency Alerts (WEA) in 2021, enabling alerts to be delivered directly to cell phones without requiring a specific app.

ShakeAlert uses approximately 1,675 seismic stations along the US West Coast. When an earthquake is detected, the system can issue an alert within approximately 5-10 seconds of the P-wave arrival at the nearest stations. The alert threshold for public notification is Modified Mercalli Intensity (MMI) IV or greater at the user's location — the level at which shaking becomes widely felt.

Japan's JMA System

Japan's EEW system, operated by the Japan Meteorological Agency, is widely considered the most advanced in the world. The system draws on an extremely dense network of seismometers operated by both JMA and NIED. Alerts are disseminated through an integrated network of television and radio broadcasts (which are automatically interrupted), mobile phone emergency alerts, public sirens, and automated controls for transportation and industrial systems.

The system proved its value during the March 11, 2011, Tohoku earthquake (M9.1). Although the system initially underestimated the earthquake's magnitude (a common challenge with very large events), it issued the first alert approximately 8 seconds after the initial P-wave detection, providing 15-30 seconds of warning to Tokyo, roughly 370 km from the epicenter. The warnings are credited with saving lives by triggering automatic train braking (Shinkansen bullet trains stopped safely) and alerting millions to take cover before the severe shaking arrived.

Emerging Technologies in Seismology

Distributed Acoustic Sensing (DAS)

Distributed Acoustic Sensing is one of the most promising new technologies in seismology. DAS uses existing fiber optic telecommunications cables as dense seismic sensor arrays. An interrogator unit at one end of a fiber sends laser pulses down the cable and measures the backscattered light. Tiny changes in the optical fiber caused by seismic waves alter the backscatter pattern, effectively turning every few meters of fiber into an individual seismic sensor.

A single 10-km fiber optic cable can provide the equivalent of thousands of seismometers at meter-scale spacing — a density impossible to achieve with conventional instruments. Researchers at Stanford University demonstrated DAS on the Stanford campus fiber network, detecting local and teleseismic earthquakes with high fidelity. Submarine DAS experiments have also recorded earthquakes using existing trans-oceanic telecommunications cables, potentially opening vast ocean basins to seismic monitoring for the first time.

DAS has limitations: the data volumes are enormous (terabytes per day for a single cable), signal quality is lower than dedicated broadband seismometers, and the technique works best for relatively high-frequency signals. Nevertheless, DAS is rapidly maturing and has already been deployed operationally in several settings, including geothermal monitoring and urban seismic hazard assessment.

Ocean-Bottom Seismometers

Approximately 70% of Earth's surface is covered by oceans, yet the vast majority of seismic stations are on land. This creates significant gaps in global monitoring coverage, particularly for earthquakes along mid-ocean ridges and in subduction zones far from shore.

Ocean-bottom seismometers (OBS) address this gap by placing instruments directly on the seafloor. Modern OBS instruments can operate autonomously for a year or more, recording broadband seismic data that is retrieved when the instrument is recovered. Real-time ocean-bottom observatories — such as Japan's S-net (150 cable-connected ocean-bottom sensors along the Japan Trench) and the US Ocean Observatories Initiative (OOI) — transmit data continuously via submarine cables.

Japan's S-net, deployed after the 2011 Tohoku disaster at a cost of approximately $1 billion, is specifically designed to improve tsunami and earthquake early warning for subduction zone events along Japan's Pacific coast. The network can detect offshore earthquakes up to 25 seconds faster than land-based stations alone.

Smartphone Seismology

The MEMS (Micro-Electro-Mechanical Systems) accelerometers built into modern smartphones are sensitive enough to detect moderate-to-large earthquakes. Several initiatives have explored crowdsourced earthquake detection using smartphones.

The MyShake project, developed at UC Berkeley, turned smartphones into seismometers via an app that ran a detection algorithm in the background. When the phone detected earthquake-like shaking, it transmitted the data to a central server for analysis. MyShake demonstrated the feasibility of smartphone-based earthquake detection and contributed to the ShakeAlert early warning system's development. While the original MyShake research app was retired, the technology influenced Google's Android Earthquake Alerts System, which now uses accelerometers in Android phones worldwide to detect earthquakes and send alerts — effectively creating the world's largest seismic network by device count.

The Community Seismic Network (CSN), operated by Caltech, uses low-cost MEMS accelerometers placed in buildings to create dense urban seismic networks. These instruments provide building-specific shaking data that can help emergency responders assess structural damage after an earthquake.

Machine Learning in Seismology

Artificial intelligence and machine learning are transforming how seismologists process data. Deep learning algorithms have demonstrated superior performance in several critical tasks:

• Phase picking: Identifying the precise arrival times of P-waves and S-waves. Models like PhaseNet and EQTransformer can pick arrivals with accuracy matching or exceeding expert human analysts, processing data thousands of times faster.
• Earthquake detection: Machine learning models can detect small earthquakes buried in noise that traditional algorithms miss, dramatically increasing the completeness of earthquake catalogs. Studies applying these techniques to existing data have revealed hidden earthquake sequences and fault structures.
• Earthquake early warning: ML-based algorithms can estimate magnitude and location from shorter windows of initial P-wave data, potentially reducing EEW alert times by critical seconds.
• Induced seismicity monitoring: ML helps distinguish human-caused earthquakes (from fracking, wastewater injection, or mining) from natural seismicity, an increasingly important capability.

[MAP: Global distribution of major seismic monitoring networks and earthquake early warning system coverage areas] Data source: IRIS, USGS, FDSN, CTBTO Features: Station locations color-coded by network, EEW coverage zones shaded, tectonic plate boundaries shown