Millions of earthquakes happen every year across the globe, but the vast majority are too small to feel. Of the roughly 500,000 that seismic networks detect annually, about 100,000 are strong enough for people nearby to notice, and only around 100 cause any damage. Between 1980 and 2009, an average of about 25 earthquakes per year were significant enough to directly affect human populations.1PubMed Central. The Human Impact of Earthquakes: a Historical Review of Events 1980-2009 and Systematic Literature Review The gap between “millions” and “25 that matter” tells you most of what you need to know about earthquake frequency: the planet is constantly rumbling, but the shaking that reshapes lives is comparatively rare.
Why Small Earthquakes Vastly Outnumber Large Ones
Earthquake frequency follows a pattern that seismologists have observed for over a century. For every step down in magnitude, roughly ten times as many earthquakes occur. So if there are about 15 magnitude-7 events in a given year, there will be something like 150 magnitude-6 events, around 1,500 magnitude-5 events, and so on down the scale. This relationship holds remarkably well across shallow, intermediate, and deep earthquakes alike, with frequency increasing roughly exponentially as magnitude decreases down to at least magnitude 2.2Quarterly Journal of the Geological Society of London. The energy of earthquakes Below that threshold, the earthquakes keep happening but most seismic networks cannot reliably detect them.
This scaling pattern means that the planet is never seismically quiet. On any given day, thousands of tiny earthquakes ripple through the crust. Most register only as wiggles on sensitive instruments. The ones that make headlines are statistical outliers, but they are entirely expected outliers. A magnitude-8 earthquake is not a freak occurrence that defies the pattern. It sits right on the predicted curve, just at the rare end of it.
Where Most Earthquakes Happen
About 90 percent of the world’s earthquakes strike along the Pacific Ring of Fire, a horseshoe-shaped belt of tectonic plate boundaries stretching from New Zealand up through Japan, across the Aleutian Islands, and down the western coasts of the Americas.3Physica A: Statistical Mechanics and its Applications. Earthquake occurrences in the Pacific Ring of Fire exhibit a collective stochastic memory for magnitudes depths and relative distances of events The concentration is not coincidental. These boundaries are where oceanic plates dive beneath continental plates or collide with each other, generating enormous stress that releases as earthquakes.
A second major seismic belt runs from the Mediterranean through the Himalayas and into Southeast Asia, accounting for most of the remaining plate-boundary earthquakes. Together these two belts cover the vast majority of global seismicity. But they are not the whole story.
Earthquakes Far From Plate Boundaries
Intraplate earthquakes, those occurring in the interior of tectonic plates well away from any boundary, are less common but can be devastating precisely because the regions they strike are usually unprepared. These earthquakes tend to happen where old faults and zones of weakness in the crust are reactivated under the broad compressional stress that plate motion imposes across entire continents. In the eastern United States, eastern China, and western Europe, this reactivation tends to produce strike-slip motion on steep faults. In places like eastern Canada, Scandinavia, Australia, and peninsular India, the faulting style shifts to thrust or normal motion.4Tectonophysics. Some seismological and geometric features of intraplate earthquakes
The causes behind these different styles are varied. In eastern Canada and Scandinavia, the crust is still slowly rebounding from the weight of ice sheets that melted thousands of years ago. In Australia and India, resistance from distant plate collisions stresses the continental interior. Along the Gulf Coast of the United States, the sheer weight of accumulated sediment plays a role. Continental margins can also see elevated seismicity because lateral differences in the thickness of the underlying rock create temperature contrasts that concentrate strain along the transition zone between oceanic and continental crust.5Tectonophysics. Enhanced intraplate seismicity along continental margins: Some causes and consequences
Earthquakes at Every Depth
Not all earthquakes happen near the surface. Seismologists classify them into three depth ranges: shallow (down to about 60 kilometers), intermediate (60 to 300 kilometers), and deep (300 to roughly 690 kilometers). Shallow earthquakes are by far the most numerous and the most destructive because they release energy close to where people live. But earthquakes at hundreds of kilometers below the surface have puzzled scientists for decades. At those depths, rock should be too hot and under too much pressure to crack the way it does near the surface.
Several competing ideas explain how deep earthquakes are possible. One leading hypothesis involves water. As tectonic plates are pushed deep into the mantle, minerals within them release water at high pressures, and this dehydration can trigger sudden failure.6PubMed. Deep-Focus Earthquakes and Recycling of Water into the Earth’s Mantle Another explanation centers on a mineral called olivine, which can persist in a metastable state as it is carried to great depths inside cold slabs. When this mineral abruptly transforms into a denser phase under stress, the sudden volume change can produce a type of faulting that mimics ordinary earthquakes.7PubMed. Mantle phase changes and deep-earthquake faulting in subducting lithosphere A third possibility is thermal runaway, where frictional heating in a narrow zone escalates so quickly that the rock melts and fails catastrophically. Each of these mechanisms has strong evidence behind it, but each also has gaps, and the full picture likely involves some combination of all three.8Annual Review of Earth and Planetary Sciences. Mechanisms and Implications of Deep Earthquakes
Aftershocks and How Earthquakes Cluster in Time
A large earthquake is rarely an isolated event. It is typically followed by a swarm of aftershocks that can continue for months or years, sometimes even a decade or longer. The rate at which aftershocks taper off follows a well-known decay pattern: a burst of activity immediately after the mainshock, then a gradual decline that can be described mathematically. This decay pattern, combined with the same magnitude-frequency scaling that governs earthquakes in general, gives seismologists a framework for estimating how many aftershocks of a given size to expect after a large event.9Geophysical Research Letters. A generalized Omori’s law for earthquake aftershock decay
What surprises many people is how far-reaching these triggered earthquakes can be. After a major earthquake, elevated seismicity is observed not only in the immediate rupture zone but across distances of up to about 1,000 kilometers, and this triggered activity can persist for 7 to 11 years.10USGS Publications Warehouse. Global Omori law decay of triggered earthquakes: Large aftershocks outside the classical aftershock zone The mechanism for this remote triggering involves the passage of seismic waves from the mainshock, which can nudge distant faults that are already close to failure.11Annual Review of Earth and Planetary Sciences. EARTHQUAKE TRIGGERING BY STATIC, DYNAMIC, AND POSTSEISMIC STRESS TRANSFER
There is an important caveat, though. While microearthquakes can be triggered at great distances almost immediately, the triggering of larger, more damaging earthquakes appears to be confined to within a few rupture lengths of the mainshock. Analysis of 30 years of global seismic data shows no significant increase in the rate of magnitude-5-or-larger earthquakes at truly global distances after big mainshocks.12Nature Geoscience. Absence of remotely triggered large earthquakes beyond the mainshock region In other words, a large earthquake in Chile does not make a large earthquake in Japan more likely. Regional hazard goes up after a mainshock, but global hazard does not.
Swarms Versus Aftershock Sequences
Not every cluster of earthquakes is an aftershock sequence. Earthquake swarms are groups of events that lack a clear dominant mainshock and instead build and wane without the characteristic decay pattern. Fluids moving through the crust play a key role in many swarms. Research in seismically active regions like the western Alps has found that both aftershock sequences and swarms can be driven by underground fluid pressure, but the two differ in how that pressure evolves over time. In a classic aftershock sequence, fluid overpressure spikes after the mainshock and then gradually decays. In a swarm, overpressure fluctuates at high levels without a clear declining trend.13Geophysical Journal International. Swarms and mainshock–aftershocks sequences are both triggered by fluids in the Ubaye Region (Western Alps) This distinction matters for hazard assessment: a swarm does not necessarily signal that a bigger earthquake is imminent, but its unpredictable timing makes it harder to model.
Human Activities That Create Earthquakes
One of the more striking developments in seismology over the past two decades is the recognition that human activities now generate a measurable fraction of the earthquakes detected worldwide. The most dramatic example came from Oklahoma, where wastewater disposal from oil and gas operations drove the state’s earthquake rate from a handful of felt events per year to hundreds. The link between injection volumes and earthquake counts is direct enough that researchers have built models to estimate how many induced events a given volume of injected fluid will produce.14Bulletin of the Seismological Society of America. Earthquakes Induced by Wastewater Injection, Part I: Model Development and Hindcasting
Wastewater injection is the primary culprit, but it is not the only one. Hydraulic fracturing itself can trigger small earthquakes, though these tend to be much smaller than those caused by high-volume wastewater disposal.15PubMed. Injection-induced earthquakes Reservoir impoundment, the filling of large dams, can also trigger seismicity by changing the stress on underlying faults. At the Irapé dam in Brazil, the weight of 136 meters of added water caused an undrained pressure buildup in the low-permeability rock below, eventually triggering a magnitude-3.0 earthquake nearly four kilometers deep.16Solid Earth. Earthquakes triggered by the subsurface undrained response to reservoir impoundment at Irapé, Brazil Near the Atatürk Dam in Turkey, a different dynamic was observed: the gravitational load of the water actually stabilized nearby faults in the short term, but pore-pressure slowly diffusing over decades gradually increased stress and elevated background seismicity over time.17Frontiers in Earth Science. Reservoir-Triggered Earthquakes Around the Atatürk Dam (Southeastern Turkey)
Mining, geothermal energy extraction, and carbon dioxide storage are other activities that can induce seismicity. The earthquakes produced are rarely large enough to cause structural damage, but they have occasionally reached magnitudes in the mid-4s and even low-5s in the case of wastewater disposal, which is strong enough to crack walls and alarm communities. The phenomenon has forced regulators in several countries to develop traffic-light protocols that scale back injection operations when seismicity exceeds certain thresholds.
Are Big Earthquakes Becoming More Common?
After a cluster of devastating earthquakes, it is natural to wonder whether the planet’s seismicity is accelerating. The answer, based on rigorous statistical testing, is no. When researchers analyzed the global catalog of large earthquakes with local clusters (like aftershock sequences) removed, the occurrence of big earthquakes was statistically indistinguishable from a random process with a constant average rate.18PubMed Central. Global risk of big earthquakes has not recently increased In plain language, the rate of major earthquakes has not gone up. Periods that seem unusually active are within the range of what random chance produces.
The perception that earthquakes are becoming more frequent is largely an artifact of improved detection. Modern seismic networks record vastly more small earthquakes than networks from even a few decades ago. Mid-ocean earthquakes, for instance, were historically undercounted because land-based stations are too far away to pick them up reliably. Autonomous underwater hydrophone arrays have revealed an order of magnitude more seismicity along mid-ocean ridges than traditional networks detected.19Journal of Geophysical Research: Solid Earth. Comparison of mid‐oceanic earthquake epicentral differences of travel time, centroid locations, and those determined by autonomous underwater hydrophone arrays The earthquakes were always there; the instruments just were not.
How Faults Behave Over Centuries
While the global rate of big earthquakes appears steady over decades, individual faults have their own rhythms. Paleoseismology, the study of ancient earthquakes through geological evidence like displaced sediment layers and offset landforms, has revealed that major faults tend to produce ground-rupturing earthquakes at characteristic intervals. These intervals can span hundreds of years in some seismic zones to thousands of years in others.20Developments in Quaternary Sciences. Earthquake recurrence inferred from paleoseismology
A global compilation of 80 long-term earthquake records found that most well-studied, frequently active faults produce large earthquakes more regularly than a purely random process would predict, consistent with the idea that stress builds up at a roughly steady rate and releases when it exceeds a threshold. But faults with low activity rates, those that rupture less often than about once every 5,000 years, show a different pattern. Their earthquakes tend to cluster in time, with bursts of activity separated by long quiet periods, which does not fit the steady-buildup model.21Geophysical Research Letters. Periodicity and Clustering in the Long‐Term Earthquake Record
For hazard forecasting, the challenge is that recurrence intervals carry substantial uncertainty. At 38 paleoseismic sites in California, revised recurrence estimates that accounted for uncertainty in interpreting geological evidence were on average 16 percent longer than conventional estimates, with disproportionately wider confidence intervals.22Seismological Research Letters. Revised Earthquake Recurrence Intervals in California, U.S.A.: New Paleoseismic Sites and Application of Event Likelihoods The longer the geological record available for a given fault, the more useful the recurrence estimates become, but even the best records leave room for surprise.
When Weather Shakes the Ground
One of the more unexpected findings in recent seismology is that extreme weather events can alter earthquake rates in a measurable way. In Taiwan, Typhoon Morakot in 2009 triggered massive landslides that stripped enormous volumes of rock from mountain slopes. In the years that followed, shallow earthquake frequency in the landslide zone roughly doubled, climbing from about 0.8 events per day above the detection threshold to about 2 per day. When all recorded earthquakes were included, the jump was from about 5 to about 10 per day.23PubMed Central. Earthquake statistics changed by typhoon-driven erosion The effect persisted for at least two and a half years after the typhoon. The mechanism is straightforward in principle: removing a heavy load of rock from the surface changes the stress state in the crust below, effectively unclamping faults that were previously held in place by the weight above them.
This is orders of magnitude smaller than the aftershock spike following a major tectonic earthquake, but it demonstrates that the crust is sensitive to changes at the surface in ways that would have seemed implausible a generation ago. Glacial retreat, seasonal groundwater fluctuations, and large-scale mining operations all redistribute mass in ways that can, at least in theory, nudge seismicity rates. The planet’s earthquake clock is not purely tectonic; it responds, subtly, to what happens on and near the surface.
Detecting Earthquakes Beyond Earth
Seismometers have now operated on three bodies in the solar system: Earth, the Moon, and Mars. The Apollo missions deployed seismometers on the lunar surface between 1969 and 1977 and recorded thousands of moonquakes. NASA’s InSight lander did the same on Mars beginning in 2018, detecting over a thousand marsquakes before the mission ended. One analysis of these datasets proposed that seismicity on all three bodies shows periodic patterns linked to the roughly 154-day oscillation cycle of the solar wind, suggesting that even seismically quiet worlds experience tremors modulated by external forces.24Journal of Geophysics. Global coupling mechanism of Sun resonant forcing of Mars, Moon, and Earth seismicity This is a provocative claim that remains far from consensus, but the basic observation that seismicity exists on worlds without plate tectonics confirms that earthquakes (or their extraterrestrial equivalents) are not unique to our planet. Thermal contraction, meteorite impacts, and tidal forces all produce quakes on other worlds, though at far lower rates and energies than what Earth’s plate boundaries generate.