The global rate of large earthquakes has not meaningfully increased. A detailed statistical analysis of every magnitude 7 and above earthquake from 1900 to 2012 found that the record, once aftershock clusters are stripped out, is indistinguishable from a random process with a steady underlying rate. The rate of magnitude 8 and above quakes did spike around 2004, but similar bursts had happened before, and smaller earthquakes stayed close to their historical average. What has changed is how many earthquakes we detect, how quickly we hear about them, and how much human activity is now shaking the ground in places that used to be quiet. Untangling these threads explains the gap between what the data show and what it feels like.
What the Global Record Actually Shows
The impression that big earthquakes are becoming more frequent took hold after the devastating 2004 Indian Ocean earthquake and the cluster of major events that followed: the 2010 Haiti and Chile quakes, the 2011 Tōhoku earthquake in Japan, and several others. Researchers at the University of California and elsewhere tested whether this cluster represented a genuine change in the planet’s seismic behavior. They applied three different statistical tests across a range of magnitude thresholds and found that the timing of large earthquakes worldwide, after removing local aftershock sequences, fits a pattern you would expect from a process that is essentially random and unchanging over time. The rate of the very largest quakes (magnitude 8 and above) was elevated starting around 2004, but rates had been almost as high in earlier decades, and no known physical mechanism could explain a real shift in the global rate of large earthquakes.1PubMed Central. Global risk of big earthquakes has not recently increased
This is an important point that gets lost in headlines. A random process can produce clusters by chance alone. Roll a die enough times and you will get streaks of sixes. The Earth’s tectonic system is incomparably more complex than a die, but the statistical signature looks the same: bursts of activity followed by lulls, with no trend line creeping upward. The planet has roughly the same number of large earthquakes per decade now as it did a century ago.
Why It Feels Like More
Two things have changed dramatically since the mid-twentieth century. First, the global seismic monitoring network has expanded enormously. In the early 1900s, a moderate earthquake in a remote part of the ocean or in a sparsely populated mountain range might not be recorded at all. Today, thousands of seismometers blanket the planet, and even small tremors are cataloged in near-real time. The number of recorded earthquakes has gone up, but that reflects better instruments, not more shaking.
Second, the speed and reach of media coverage have exploded. A magnitude 6 earthquake in a developing country might once have taken days to appear in foreign newspapers, if it appeared at all. Now it is a push notification on your phone within minutes. Research on how digital media shapes earthquake risk perception bears this out. A study of urban youth in Indonesia found that frequent exposure to earthquake-related content on digital platforms was strongly associated with higher perceived seismic risk. The effect was substantial: digital risk amplification predicted risk perception with a moderate-to-strong relationship, though exposure alone did not translate directly into preparedness behavior. Instead, heightened perception had to develop first before people changed what they did.2AOSIS (Jàmbá: Journal of Disaster Risk Studies). Digital media, risk perception, and earthquake preparedness among Indonesian urban youth
In practical terms, this means that social media and 24-hour news create a feedback loop. Every earthquake gets amplified, every aftershock gets its own headline, and the cumulative impression is that something unprecedented is happening. The earthquakes are real, but the sense that they are accelerating is largely an artifact of living inside an information environment that never stops updating.
How Aftershocks Create Clusters
Even when the background rate of earthquakes stays flat, individual large earthquakes generate swarms of follow-on events that can dominate headlines for weeks or months. Aftershocks follow a well-known pattern: they are most frequent right after the main event and decay over time in a predictable way described by the Omori law. The physical explanation involves stress redistribution. When a fault slips in a large earthquake, it changes the stress on neighboring faults and on small subsidiary fractures alongside the main fault. Numerical simulations show that most aftershocks happen on these small subsidiary faults, triggered by the local stress increase caused by the irregular geometry of the main fault itself.3Journal of Geophysical Research: Solid Earth. Mainshock and Aftershock Sequence Simulation in Geometrically Complex Fault Zones
Beyond these nearby aftershocks, large earthquakes can also trigger seismicity at surprising distances. The 1992 Landers earthquake in California set off a series of smaller quakes across other parts of the state, well beyond what traditional aftershock zones would predict. Research into this phenomenon found that the seismic waves radiating from a large earthquake can cause elastic weakening and failure on distant faults if the waves produce dynamic deformations above a threshold of several microstrain. This can happen regardless of how far away the triggering event is.4Nature. Dynamic triggering of earthquakes
So a single large earthquake can produce hundreds of detected events in its aftershock zone, plus additional triggered events at greater distances. For weeks after a major quake, earthquake counts spike dramatically. If another large earthquake happens to occur somewhere else during the same window, the combined catalog looks alarming even though each sequence is behaving exactly as seismology predicts.
Human Activities That Genuinely Cause Earthquakes
While the natural background rate of earthquakes has stayed roughly constant, there is one category of seismicity that has genuinely and measurably increased: earthquakes caused by human industrial activity. This is not speculation. Several distinct mechanisms are well documented.
The most prominent involves wastewater disposal from oil and gas operations. When water is injected deep underground at high pressure, it changes the fluid pressure on nearby faults and can trigger them to slip. In Oklahoma, the connection became impossible to ignore during the 2010s, when the state went from experiencing a handful of magnitude-3-plus earthquakes per year to hundreds. Research showed that high-volume wastewater injection not only triggered earthquakes but caused them to migrate deeper over time at roughly half a kilometer per year, driven by density differences between the injected fluid and the surrounding rock. Even after injection rates were substantially reduced, the density-driven pressure front continued migrating downward, meaning earthquakes could persist for a decade or more after the injection slowed.5Nature Communications. High density oilfield wastewater disposal causes deeper, stronger, and more persistent earthquakes
Hydraulic fracturing itself, distinct from wastewater disposal, also triggers earthquakes, though typically smaller ones. The process drives fluid under high pressure into rock formations, and that pressure can reactivate existing faults.6Reviews of Geophysics. Hydraulic Fracturing‐Induced Seismicity The concern has grown as the technique has expanded. A review of the science found that hydraulic fracturing can trigger seismic events through a variety of direct and indirect mechanisms, and in some cases could result in injuries and damage to infrastructure.7Nature Reviews Earth & Environment. Developments in understanding seismicity triggered by hydraulic fracturing
Large reservoirs represent another source. When a dam fills, the weight of the water column exerts pressure on the underlying crust and changes the pore-fluid pressure at depth. A study of the Atatürk Dam in southeastern Turkey found that while the immediate gravitational load of the water actually stabilized local faults in the short term, pore-pressure diffusion over decades gradually increased the effective stress in the seismogenic zone, enhancing background seismicity over time.8Frontiers in Earth Science. Reservoir-Triggered Earthquakes Around the Atatürk Dam (Southeastern Turkey) The overpressure from a typical reservoir is modest, on the order of a tenth of a megapascal, but if a fault is already near its breaking point, even a small nudge can trigger failure.9Scientific Reports. Anthropogenic Triggering of Large Earthquakes
Geothermal energy projects add a more recent chapter. Enhanced geothermal systems inject fluid underground to extract heat, and the process has induced earthquakes in several locations. In the Strasbourg region of France, seismicity induced by geothermal drilling became enough of a public issue to generate significant political debate and community opposition.10The Extractive Industries and Society. Unlocking the subsurface through knowledge controversy Similar episodes have occurred in South Korea and Switzerland.
If you live in a region that has seen a genuine uptick in felt earthquakes over the past fifteen years, human-caused seismicity is the most likely explanation. Central Oklahoma, parts of Texas, western Canada, and several European sites have all experienced this. The earthquakes are real, they are new, and they have a clear cause. This is one area where the public intuition that “there are more earthquakes than there used to be” is correct for specific places, even though it is not correct globally.
Groundwater, Rainfall, and the Surprising Ways Water Moves Faults
Beyond industrial injection, the natural movement and removal of water can also affect earthquake activity. In central California, a century and a half of groundwater pumping from the Central Valley has removed enough mass to cause the surrounding landscape to rise by one to three millimeters per year. GPS measurements show that this uplift extends into the Coast Ranges and the Sierra Nevada, and the unloading reduces the effective normal stress on the San Andreas Fault, bringing it closer to failure. Researchers found a viable mechanism linking groundwater depletion to observed seasonality in small earthquakes near Parkfield, with seismicity peaking in late summer and autumn when the valley is driest and the flex of the crust is greatest.11Nature. Uplift and seismicity driven by groundwater depletion in central California A separate modeling study confirmed that the stress changes on the San Andreas Fault correlate positively with both shallow seismicity and low-frequency earthquakes, for both short-term drought cycles and the long-term drawdown since the 1860s.12Geophysical Research Letters. San Andreas Fault Stress Change Due To Groundwater Withdrawal in California’s Central Valley, 1860‐2010
On a much shorter timescale, heavy rainfall can trigger small earthquakes. Below Mount Hochstaufen in southeastern Germany, a densely monitored seismic zone showed that recorded earthquakes were highly correlated with calculated pore-pressure changes caused by rainwater diffusing down into the crust. The faults in the area are so close to their failure threshold that even the tiny pressure variations associated with precipitation are enough to trigger them at several kilometers depth.13Geophysical Research Letters. Evidence for rainfall‐triggered earthquake activity A similar relationship was documented in the central Swiss Alps after an extreme rainfall event in 2005, when the infiltration of large amounts of rain was followed by a measurable increase in local seismicity.14Geophysical Journal International. Locally triggered seismicity in the central Swiss Alps following the large rainfall event of August 2005
Groundwater extraction on the other side of the world tells a parallel story. Along the Dead Sea Fault in Jordan, pumping from the Wadi Al-Arab basin has changed Coulomb failure stresses by several megapascals, enough to trigger seismicity on nearby fault segments.15Journal of Geophysical Research: Solid Earth. Induced Seismicity by Groundwater Extraction at the Dead Sea Fault, Jordan These cases illustrate a broader principle: faults that are already stressed close to their breaking point can be tipped over by remarkably small changes in the water load above or within the crust. As climate change intensifies both droughts and extreme rainfall events, these water-related triggers could become more frequent in some regions.
Long-Term Seismic Cycles Most People Have Never Heard Of
Earthquakes do not occur on a perfectly even schedule, even on a single fault. Beyond the short-term clustering of aftershocks, faults can exhibit much longer cycles of activity and quiet that span thousands of years. Research on the Wasatch Fault in Utah, one of the most studied fault systems in the western United States, found evidence of earthquake clusters lasting thousands of years, separated by comparable periods of relative calm. The researchers modeled these as “supercycles,” long-period stress cycles that contain multiple individual earthquake cycles within them.16Journal of Geophysical Research: Solid Earth. Earthquake supercycles as part of a spectrum of normal fault slip styles
Parallel findings came from the Fucino faults in central Italy, where dating of fault surfaces using cosmogenic isotopes revealed that strain accumulated and released in cycles lasting roughly three to six thousand years. Each supercycle included a long quiet phase of slow strain buildup lasting several thousand years, followed by a cluster of three to four large earthquakes that released most of the strain in less than one to two thousand years.17Journal of Geophysical Research: Solid Earth. Earthquake synchrony and clustering on Fucino faults (Central Italy) as revealed from in situ 36Cl exposure dating
These supercycles are invisible on the timescale of a human life or even a civilization. A fault that last ruptured four thousand years ago might be in its quiet accumulation phase or might be nearing the end of it. The instrumental seismic record, which covers only about 120 years with any reliability, is a tiny window into processes that play out over millennia. This makes it almost impossible to tell, from modern data alone, whether any particular region is in a “busy” or “quiet” phase of a longer cycle.
Ice Sheets, Glaciers, and the Seismic Hangover From the Last Ice Age
One of the more counterintuitive connections in seismology involves the relationship between ice sheets and earthquakes. During ice ages, kilometers-thick glaciers press down on continental crust, suppressing fault activity beneath them. When the ice melts, the crust rebounds, and that unloading can increase slip rates on active faults or reactivate dormant ones.18Geology Today. Response of faults to climate‐induced changes of ice sheets, glaciers and lakes Scandinavia is still rising measurably from the loss of its ice sheet more than ten thousand years ago, and the associated strain field dominates the region’s seismic behavior to this day.
Modeling of the North American ice sheet’s retreat showed that glacial unloading could trigger earthquakes both within the former ice margin, such as the Charlevoix seismic zone in Quebec, and beyond it, such as the Wabash Valley along the Illinois-Indiana border.19Geophysical Research Letters. Can deglaciation trigger earthquakes in N. America? A broader study of this phenomenon across both northern Europe and eastern North America found that the interference between tectonic strain and radial rebound strain from ice sheet removal creates alternating zones of elevated and suppressed seismicity around former rebound centers. This helps explain why some seemingly stable continental interiors experience occasional moderate earthquakes that are hard to account for using plate tectonics alone.20Quaternary Science Reviews. Deglaciation Seismotectonics: a principal influence on intraplate seismogenesis at high latitudes
With modern glaciers and ice sheets shrinking under warming temperatures, there is a reasonable question about whether this process is creating new seismic activity in places like Greenland and Alaska. The timescales are very different from the massive post-ice-age rebounds, but the underlying physics is the same. The crust responds to changes in the load sitting on top of it, and those responses include changes in which faults are closer to or farther from failure.
How Incomplete Records Warp Our Sense of What Is Normal
Judging whether earthquake activity is unusual requires knowing what “usual” looks like, and that knowledge has limits. The instrumental seismic record with global coverage goes back only to around 1900, and the data before the 1960s are significantly less complete than what came after. For older periods, researchers rely on historical accounts and archaeological evidence. Destruction layers in ancient settlements, collapsed walls, displaced foundations, and other physical traces can fill in the seismic record going back thousands of years. But as one review noted, the further back you go, the more the historical record fades and the archaeological record has to take over, with all of its ambiguities.21ScienceDirect. Earthquakes and archaeology
This matters because a 120-year record is statistically thin for drawing conclusions about rare events. A magnitude 9 earthquake might happen on a given subduction zone once every few hundred years, meaning the entire instrumental era might capture zero or one of them. A cluster of two or three large quakes in a decade looks extraordinary against that short record but might be unremarkable against the backdrop of geologic time. Paleoseismology, the study of ancient earthquakes preserved in sediments and fault scarps, consistently shows that past earthquake activity was at least as variable as what we see today, with clusters and quiet periods alternating in ways that our short modern record cannot fully capture.
The mismatch between what we can measure and what we need to know is one of the deepest challenges in earthquake science. It feeds public anxiety because there is no simple way to tell a concerned person, “This is normal.” The honest answer is that we do not always know what normal looks like on a fault whose last major rupture predates the seismograph.