Every earthquake you see on Tremr — the M2.1 off the coast of Japan, the M4.8 in Turkey, the M6.0 in Peru — was detected by an instrument buried in the ground somewhere on Earth. That instrument is a seismograph, and the chain of events from ground tremor to the number on your screen is a remarkable piece of engineering that has been refined over more than a century.
The Basic Idea
The core principle of a seismograph is surprisingly simple: inertia. When the ground moves, a heavy mass suspended inside the instrument tends to stay still while the casing around it moves with the earth. The relative motion between the stationary mass and the moving casing is what gets recorded.
Think of it like a pen hanging from a spring inside a box. When the box shakes, the pen stays roughly in place (due to inertia) while the paper attached to the box moves beneath it. The result is a wavy line — a seismogram — that captures the motion of the ground over time. That wavy line is the raw material of seismology.
From Pendulum to Digital
The earliest seismographs, built in the late 19th century, used exactly this pendulum-and-drum mechanism: a heavy weight on a wire, a rotating drum of paper, and an ink pen tracing the movement. They were enormous instruments, sometimes weighing several tonnes, housed in specially built vaults to isolate them from vibration.
Modern seismographs are very different. Today's instruments use a coil of wire suspended inside a magnetic field. When the ground shakes, the coil moves relative to the magnet, generating a tiny electrical current proportional to the ground's velocity. That current is amplified, filtered, digitised, and recorded — continuously, 24 hours a day, 365 days a year — onto a computer hard drive or transmitted directly to a data centre over the internet.
Modern instruments also record motion in three directions simultaneously: up-down, north-south, and east-west. This three-axis data allows seismologists to fully characterise the motion of the ground at any point, which is essential for locating an earthquake's source and understanding its mechanism.
Reading the Waves
An earthquake generates several types of seismic waves, and a seismogram shows them arriving at different times. The first to arrive are P-waves (primary or compressional waves) — they travel fastest, pushing and pulling rock like a spring. Next come S-waves (shear or secondary waves) — slower, but they shake rock side to side and carry more destructive energy. Finally, surface waves roll across the outer layer of the earth like ripples on water; they're the slowest but often the most destructive at distant locations.
The time difference between P and S wave arrivals at a single station tells seismologists how far away the earthquake was. With readings from at least three stations, they can triangulate the earthquake's epicentre. With dozens or hundreds of stations, they can pinpoint location to within a few kilometres and calculate depth, fault geometry, and the amount of energy released.
The Global Network
No single seismograph can detect all earthquakes. What makes modern earthquake monitoring possible is the global network — thousands of stations spread across every continent, from Iceland to Antarctica, from the bottom of the Pacific Ocean to the summit of Hawaiian volcanoes.
The USGS operates the Global Seismographic Network (GSN), a backbone of around 150 high-quality stations worldwide, supplemented by thousands of regional and national networks. Every station streams data continuously to processing centres, where algorithms scan incoming waveforms around the clock for earthquake signatures. When a potential event is detected at multiple stations simultaneously, an automated system triggers an alert and begins computing a location and magnitude — often within 5–10 minutes of the earthquake occurring.
From Sensor to Tremr
Once USGS algorithms have processed a detection and computed a preliminary magnitude and location, the event is published to the USGS public GeoJSON feed — the same feed that powers Tremr. Tremr queries that feed every five minutes and displays new events on the map and in the list, usually within 10–20 minutes of the earthquake occurring.
The magnitude you see on Tremr is a USGS estimate that may be updated as more stations report in and seismologists refine their analysis. Large earthquakes often have their magnitudes revised in the hours and days after the event as data accumulates. The "reviewed" status flag in the detail panel indicates whether a human analyst has confirmed the automated estimate.
Next time you tap an earthquake dot on Tremr, picture the chain behind it: ground vibrates → coil moves in magnet → current flows → data streams to servers → algorithm detects pattern → magnitude computed → feed updated → your screen refreshes. That whole chain, from shaking ground to your fingertip, typically takes less time than it takes to brew a coffee.
A Brief History: From Wineglasses to Broadband
The earliest earthquake detectors were crude — Renaissance-era Italian scientists noticed that pendulums or wineglasses would tip over or spill during earthquakes. But detecting an earthquake is not the same as measuring one.
The first true seismograph was built in 1875 by Italian physicist Filippo Cecchi, who devised a pendulum-based system that left a mark on a moving cylinder of soot-covered paper. When the earth shook, the pendulum would stay relatively stationary due to inertia while the paper beneath moved, leaving an ink trace. The principle was sound, but the instruments were enormous and temperamental.
John Milne, a British geologist working in Japan in the late 19th century, revolutionised seismograph design. His horizontal pendulum seismograph — deployed across Japan — could detect earthquakes across entire islands. By the 1890s, Japan had a network of Milne seismographs that proved earthquake source location was possible using multiple instruments. This was the birth of quantitative seismology.
For most of the 20th century, mechanical seismographs ruled. They were refined and improved, with mechanical amplification allowing even small motions to produce visible traces. But they were slow: the data was on paper, and someone had to physically retrieve and analyse it, sometimes days later. The earthquake was already fading from memory before the seismogram was even interpreted.
Digitisation changed everything. By the 1980s, electronic seismometers with digital recording appeared. By the 1990s, broadband seismometers — instruments sensitive to a wide range of frequencies — became standard. These modern instruments are roughly 100 times more sensitive than their mechanical predecessors and can record continuously without degradation.
Modern Digital Networks: Real-Time Data Streaming
Today's seismometers are often described as "broadband" or "short-period," depending on their frequency response. A broadband instrument responds to both rapid vibrations (like the snap of fault rupture) and slow oscillations (like surface waves). This versatility is crucial for modern seismology — one instrument can capture information about the earthquake's source mechanism, its depth, its magnitude, and its effects.
The real revolution in modern seismology is data transmission. A seismometer in rural Iceland streams data in real-time to processing centres in the United States. The continuous data flow — 100+ samples per second from each instrument — means that algorithms can detect and locate an earthquake within seconds of rupture, often before the strongest shaking has even reached nearby cities.
The USGS Global Seismographic Network uses a standard protocol to transmit data from stations worldwide. Each station has redundant internet connections, so if one fails, data continues to flow. Some stations in remote locations use satellite internet links, ensuring that even the most isolated instruments contribute to the global picture.
From 150 Stations to Thousands: The Full Monitoring Picture
The USGS Global Seismographic Network's ~150 stations are just the backbone. Supplementing these are thousands of regional and national networks operated by governments, universities, and research institutions worldwide.
Japan alone operates over 1,000 seismic stations — the densest network in the world. The array is so dense that most earthquakes are recorded at multiple stations with short distances, allowing precise location to within a few kilometres. California's network has hundreds of stations. Europe, South America, Indonesia, and New Zealand each maintain extensive networks.
Dense networks have an unexpected benefit: they detect much smaller earthquakes. A single station might detect magnitude 4 earthquakes from a nearby fault. A dense network can detect magnitude 1-2 earthquakes. These small earthquakes reveal details about fault geometry and stress state that inform understanding of hazard.
The downside of all this data is volume. A global network of 10,000+ stations, each sampling 100+ times per second, generates petabytes of data annually. Storage, transmission, and processing of this data requires industrial-scale computing infrastructure. But the payoff is extraordinary: seismology has transformed from a slow, qualitative science into a real-time, high-precision discipline.
Beyond the Network: Complementary Technologies
GPS (Global Positioning System) has become an unexpected partner to seismology. GPS receivers can detect millimetre-scale ground displacement. During a major earthquake, GPS stations can record permanent ground shifts — the amount the earth actually moved at that location. This permanent displacement, combined with seismic waves recorded at the same location, provides a complete picture of the earthquake rupture.
Seafloor seismometers and pressure sensors on the ocean floor detect submarine earthquakes directly. These instruments are crucial for understanding the 1960 Chile earthquake (M9.5, the largest ever recorded) and other megathrust earthquakes that rupture beneath the ocean. Without seafloor instruments, whole classes of earthquakes would be poorly understood.