Seismic activity is any vibration that travels through the Earth’s rocks, whether triggered by shifting tectonic plates, volcanic eruptions, landslides, or even human industrial operations. Most people equate the term with earthquakes, and earthquakes do account for the lion’s share of seismic energy released each year, but the category is broader than that. Measuring all of this shaking involves a surprisingly diverse toolkit, from century-old principles of wave timing to deep-learning algorithms running on pocket-sized computers.
Why the Ground Shakes in the First Place
Earth’s outer shell is broken into tectonic plates that drift, collide, and slide past one another. Where plates meet, friction locks their edges together even as deeper forces keep pushing. Stress accumulates over decades or centuries until the locked zone can no longer hold, and the rock snaps forward in a sudden slip. That abrupt release radiates energy outward as seismic waves. The process of stress buildup involves several preparatory changes: off-fault yielding, shifts in underground fluid pressure, and sometimes small foreshocks that partially relieve local stress before the main rupture. When conditions reach a tipping point, rapid shear failure follows, with the speed and intensity of the slip varying from point to point along the fault.
After the main rupture, the story isn’t over. Laboratory-scale models of subduction-zone earthquakes show that the overriding plate snaps back quickly, while the downgoing slab rebounds more slowly and with less displacement. That combination of fast upper-plate rebound and delayed slab response can actually help re-lock the fault, resetting the clock for the next earthquake cycle.1Journal of Geophysical Research: Solid Earth. Upper Plate Response to a Sequential Elastic Rebound and Slab Acceleration During Laboratory‐Scale Subduction Megathrust Earthquakes Meanwhile, the stress changes and fluid pressures that set up the initial rupture involve processes like hydraulic fracturing of rock and lubrication by fluids or even melt along the fault surface.2Interpretation. Crustal stress buildup/relaxation and pore pressure in preparation and sequential brittle-shear fault rupture
Volcanoes and Other Non-Tectonic Sources
Not all seismic activity comes from faults slipping. Volcanoes generate their own distinctive tremors as magma and superheated gases force their way through rock. At Sakurajima volcano in Japan, researchers classified volcanic earthquakes by depth and found that the shallow ones (less than about two kilometers deep) cluster beneath and just south of the active craters, tracking closely with the known magma supply system.3Geophysical Journal International. Classification of volcanic tremors and earthquakes based on seismic correlation: application at Sakurajima volcano, Japan At the Iwo-Yama vent in Japan’s Kirishima volcanic complex, continuous tremor turned out to be generated by gas and hot water resonating inside crack-like fissures beneath a boiling pool. The tremor amplitude at certain frequencies tracked with the amount of hot water in the pool, which itself rose and fell with rainfall and evaporation.4Journal of Volcanology and Geothermal Research. Volcanic tremor associated with successive gas emission activity at a boiling pool
Other natural causes include landslides, glacial calving, and meteorite impacts, all of which send detectable waves through the ground. Even ocean waves constantly pound the seafloor hard enough to create a permanent background hum (more on that later). The point is that “seismic activity” covers any event that makes the solid Earth vibrate, whether the cause is geological, atmospheric, oceanic, or biological.
When Humans Cause Earthquakes
Induced seismicity has moved from a footnote in geophysics textbooks to front-page news. It has long been recognized that earthquakes can be triggered by filling reservoirs, mining, pumping fluids out of the ground, or injecting fluids back in.5PubMed. Injection-induced earthquakes The modern surge in concern centers on oil and gas operations, especially the disposal of enormous volumes of wastewater produced alongside hydrocarbons.
In the Delaware Basin of west Texas, widespread deep earthquakes have been linked mainly to shallow wastewater injection. The injected fluid doesn’t need to physically reach the fault; it changes the stress field through the surrounding rock, even when impermeable shale layers sit between the injection zone and the deep faults.6PubMed Central. Widespread deep seismicity in the Delaware Basin, Texas, is mainly driven by shallow wastewater injection In western Canada, hydraulic fracturing itself appears to be the dominant trigger rather than wastewater disposal, which is the reverse of the pattern seen in the central United States. A detailed catalog from one Canadian shale play showed earthquakes tightly clustered around fracking sites in both space and time, with the largest event reaching magnitude 3.9. That quake struck weeks after injection ended, along a fault extending from the injection zone down into the crystalline basement rock, and seismicity persisted for months.7PubMed. Fault activation by hydraulic fracturing in western Canada
These findings matter because the mechanism determines how you manage the risk. If wastewater disposal is the culprit, reducing injection volumes or spreading them across more wells can help. If hydraulic fracturing directly activates a fault, you need to know where faults are before you drill. The lag between injection and the largest earthquake, sometimes weeks or months, also makes monitoring tricky.
Magnitude Scales and What They Actually Tell You
When an earthquake hits the news, the first number you hear is its magnitude. That single digit is doing a lot of heavy lifting: it describes the total energy released at the source. The scale most seismologists rely on today is the moment magnitude scale, developed in the late 1970s. It relates magnitude to a physical quantity called the seismic moment, which combines how much rock slipped, the area of the fault that broke, and the stiffness of the rock. The original paper showed that moment magnitude lines up well with the older local-magnitude and surface-wave-magnitude scales in their respective useful ranges, but unlike those older scales it doesn’t “saturate” for the very largest earthquakes.8Journal of Geophysical Research: Solid Earth. A moment magnitude scale
Saturation is the practical reason the older Richter scale fell out of professional use. For earthquakes above about magnitude 7, the Richter scale underestimates the true size because the specific type of seismic wave it measures maxes out. Moment magnitude doesn’t have that ceiling. A recent theoretical treatment confirmed that the scale’s mathematical foundation rests on relationships between radiated energy, stress drop during faulting, and the seismic moment itself.9Earth, Planets and Space. A theoretical basis of the moment magnitude scale
One thing that catches people off guard is that the scale is logarithmic. Each whole number step represents roughly 32 times more energy. A magnitude 7 earthquake releases about a thousand times more energy than a magnitude 5. So the numerical gap between a 5.0 and a 7.0 looks small on paper but is enormous in destructive potential.
Intensity Is Not the Same as Magnitude
Magnitude describes what happened at the earthquake’s source. Intensity describes what you actually felt where you were standing. Two people living at different distances from the same earthquake will report very different intensities, even though the magnitude was fixed at the source. The most widely used intensity scale in the United States is the Modified Mercalli Intensity (MMI) scale, which runs from I (not felt at all) to XII (total destruction).
MMI ratings come from human observations and damage reports, but they can also be estimated from instrument readings. Research using eight significant California earthquakes found that lower intensities (roughly V through VII) correlate best with peak ground acceleration, which is the sharp jolt you feel, while higher intensities (above VII) correlate better with peak ground velocity, which relates more closely to structural damage.10Earthquake Spectra. Relationships between peak ground acceleration, peak ground velocity, and Modified Mercalli Intensity in California This makes intuitive sense: whether you notice an earthquake at all depends on how sharply the ground jerks, but whether a building collapses depends more on how fast the ground moves back and forth.
Local soil conditions also matter. Soft, water-saturated sediment can amplify shaking compared to hard bedrock, which is why two neighborhoods at the same distance from a fault can experience very different damage. Studies of strong-motion data in the western United States have shown that the relationship between peak ground motion and intensity changes depending on whether the site sits on rock, stiff soil, or soft sediment, with the amplification patterns actually reversing at the highest intensities.11Soil Dynamics and Earthquake Engineering. A note on scaling peak acceleration, velocity and displacement of strong earthquake shaking by Modified Mercalli Intensity (MMI) and site soil and geologic conditions
Pinpointing Where an Earthquake Happened
An earthquake generates several types of seismic waves. The two most important for locating the source are P-waves (compressional waves that arrive first because they travel fastest) and S-waves (shear waves that arrive second). Because P-waves and S-waves travel at known but different speeds through a given type of rock, the time gap between their arrivals at a seismometer tells you how far away the earthquake was. Get that distance from three or more stations and you can triangulate the earthquake’s location.
The principle is simple enough that students have reproduced it in a schoolyard. In one educational experiment, students used a hammer strike as a seismic source, identified P-waves, S-waves, and surface waves by hand on their recordings, then applied two-dimensional triangulation. Using a statistically averaged velocity model for their site, they pinpointed the source with an average error of just 0.6 meters.12Physics Education. Earthquake in the courtyard: an experiential learning activity for epicentre triangulation and wave phase identification Professional networks use the same underlying logic but with thousands of stations and sophisticated three-dimensional models of how wave speed varies with depth and rock type.
Using Seismic Waves to See Inside the Earth
Seismologists realized decades ago that earthquakes are not just hazards but also free illumination sources. Every large earthquake sends waves crisscrossing through the planet’s interior. By comparing the expected and actual travel times of those waves at stations worldwide, researchers build three-dimensional images of the Earth’s interior in a process called seismic tomography, conceptually similar to a medical CT scan. Over the last few decades, this technique has revealed fundamental features of the mantle and core, including plumes of hot rock rising from the deep mantle and slabs of cold, dense oceanic plate sinking at subduction zones.13PubMed Central. Mantle dynamics and seismic tomography
Understanding what happens at great depth matters for earthquake science, too. Deep earthquakes, those occurring hundreds of kilometers below the surface in subducting slabs, are puzzling because at those pressures and temperatures rock should flow rather than snap. Researchers have proposed three main mechanisms: a mineral phase transformation that localizes stress, release of water from hydrated minerals that weakens rock, and thermal runaway where frictional heating softens rock so fast that it concentrates slip. Each mechanism explains part of what is observed, but none single-handedly accounts for everything, and the current thinking is that more than one process likely works together.14Annual Review of Earth and Planetary Sciences. Mechanisms and Implications of Deep Earthquakes
Early Warning Systems and Artificial Intelligence
Because P-waves travel faster than the more destructive S-waves and surface waves, there is a narrow window between when an earthquake begins and when the worst shaking arrives at a given city. Earthquake early warning systems try to detect the P-wave automatically and issue an alert before the damaging waves reach populated areas. Even a few seconds of warning can be enough to slow trains, open fire station doors, or trigger an automatic shutdown of gas pipelines.
The bottleneck has always been speed and accuracy of detection. Deep-learning models are now being deployed to solve this. In Indonesia, a deep-learning system was trained to pick out P-wave arrivals in real time, with the goal of keeping the detection delay to no more than about two seconds after the actual wave arrival.15Applied Computing and Geosciences. Deep learning for real-time P-wave detection: A case study in Indonesia’s earthquake early warning system In New Zealand, a lightweight neural network was designed to run on a Raspberry Pi, a credit-card-sized computer. Despite having only around 38,000 trainable parameters and needing less than seven milliseconds to process each snippet of data, it correctly identified P-wave segments 98 percent of the time.16Scientific Reports. Lightweight convolutional neural network for real-time earthquake P-wave detection on edge devices in New Zealand Running detection on cheap hardware at the sensor itself, rather than streaming data to a remote server, shaves precious fractions of a second off the alert time.
Separate from early warning, machine-learning classifiers are also being used to distinguish genuine earthquake signals from noise in massive archives of past data. One framework achieved better than 99 percent classification accuracy on its home dataset and still managed about 95 percent when transferred to a completely different seismic network, suggesting these tools generalize well across regions.17Oxford Academic. Earthquake signal detection using a multiscale feature fusion network with hybrid attention mechanism
Reading Earthquakes from the Geologic Record
Instrumental seismology only stretches back about 120 years, which is a blink in geological time. To understand how often big earthquakes happen on a given fault, scientists dig trenches across the fault trace and look for layers of sediment that have been offset or disturbed by past ruptures. This field, paleoseismology, uses radiocarbon dating and other techniques to pin ages on ancient earthquakes.
Along the Longmenshan fault zone in China, which produced the devastating 2008 Wenchuan earthquake, trench studies identified two to three earlier events with similar vertical offsets. The oldest of those dates back somewhere between roughly 3,300 and 7,700 years ago.18Tectonophysics. Paleoseismic evidence and repeat time of large earthquakes at three sites along the Longmenshan fault zone In Japan, trenching across a fault that ruptured during the 2016 Kumamoto earthquake revealed evidence of seven surface-rupturing events stretching back more than 14,000 years, with an average interval of roughly 2,600 to 2,900 years between events. Crucially, the intervals were not regular: the gap between two particular events was much longer than the average.19Island Arc. Detailed paleoseismic history of the Hinagu fault zone revealed by the high‐density radiocarbon dating and trenching survey
That irregularity is a recurring headache for hazard forecasting. Faults do not operate on a strict schedule. An average recurrence interval gives you a rough sense of how active a fault is, but any individual gap between earthquakes can be much shorter or longer than the average.
The Earth’s Constant Background Hum
Even between earthquakes, seismometers are never truly silent. The ground vibrates continuously at very low frequencies, driven primarily by ocean waves. This ambient noise falls into three bands: a very low-frequency “seismic hum” generated by ocean infragravity waves pressing on the deep seafloor; primary microseisms at slightly higher frequencies, caused by ocean swell interacting with the gentle slopes of continental shelves; and secondary microseisms at higher frequencies still, produced by opposing ocean waves creating pressure pulses at the sea surface.20PubMed Central. Ambient seismic wave field
Far from being a nuisance, this noise has become a powerful research tool. Because ambient microseisms travel through the Earth continuously, seismologists can cross-correlate recordings at pairs of stations to extract information about the rock between them, effectively using ocean storms as a free, always-on seismic source. Observations around the remote island of Tristan da Cunha in the South Atlantic showed that secondary microseisms there are strong and closely correlated with local wave heights, while clear signals from distant storm sources also emerge when the data are processed.21Journal of Geophysical Research: Solid Earth. Characteristics of the Oceanic Ambient Seismic Noise Around Tristan da Cunha in the South Atlantic From OBS Data This technique lets researchers image the Earth’s crust and upper mantle even in places where earthquakes are rare.
Liquefaction and Other Secondary Hazards
The shaking itself is only part of the danger. One of the most dramatic secondary effects is soil liquefaction, in which waterlogged, loose soil loses its strength during shaking and behaves like a thick fluid. Buildings can sink, tilt, or topple not because the earthquake cracked their walls but because the ground beneath them turned to quicksand. The triggering mechanism involves the earthquake’s horizontal acceleration driving up pore-water pressure in the saturated soil until it exceeds the effective stress holding grains together. The depth and severity depend on features of the soil profile, especially the width of the water-saturated zone, and on the intensity of the bedrock motion underneath.22ScienceDirect. Mechanisms to explain soil liquefaction triggering, development, and persistence during an earthquake
Liquefaction risk maps are now standard products of seismic hazard assessments. Areas near rivers, harbors, and reclaimed land are particularly vulnerable because their soils tend to be young, loose, and saturated. In some past earthquakes, liquefaction damage extended well beyond the zone of severe shaking, catching communities off guard.
Moonquakes and Seismology Beyond Earth
Seismic measurement is not confined to our planet. The Apollo missions installed seismometers on the lunar surface that recorded data for years, revealing that the Moon has its own quakes. These moonquakes come in several flavors: shallow ones originating within about 200 kilometers of the surface, deep ones centered around 900 kilometers down, vibrations from meteoroid impacts, and signals from deliberate crashes of spent rocket stages. Moonquakes have a distinctive signature where energy builds slowly to a peak and then decays even more slowly, a consequence of the bone-dry, heavily fractured lunar crust scattering seismic waves rather than absorbing them.23Space Science Reviews. Lunar Seismology: A Data and Instrumentation Review
NASA’s InSight lander did the same for Mars, detecting hundreds of marsquakes between 2018 and 2022 and using them to map the planet’s interior layering. Future missions to Europa and Titan have proposed deploying seismometers on icy moons. In each case, the logic is the same one seismologists have used on Earth for over a century: listen to how waves move through a body and you can figure out what it’s made of, how thick its layers are, and whether anything is molten inside.
Can Animals Sense Earthquakes Before They Happen?
Reports of unusual animal behavior before earthquakes go back centuries: dogs howling, fish leaping, snakes emerging from burrows in winter. The question is whether animals are actually sensing something real or whether the stories are just selective memory. A review comparing these reports with laboratory studies of animal sensory thresholds concluded that some animals are substantially better than humans at detecting certain geophysical stimuli that may precede earthquakes, including ground tilts, electric-field changes, and low-frequency vibrations.24Reviews of Geophysics. Unusual animal behavior before earthquakes: A review of possible sensory mechanisms That makes the reports plausible, but plausible is a long way from useful. The signals are inconsistent, the lead times vary wildly, and no one has managed to build a reliable forecasting system around animal behavior. For now, seismometers and GPS networks remain far more dependable.