Large earthquakes are not becoming more frequent. The global rate of magnitude 7 and above earthquakes has remained roughly steady since modern seismological records began around 1900, and statistical analyses consistently show that what looks like clustering is explained by random variability plus localized aftershock sequences. The perception that the ground is shaking more often comes from a combination of vastly improved detection networks, rapid global news coverage, and growing populations in earthquake-prone regions. That said, one category of earthquakes genuinely has surged in certain places, and the cause is human activity, not geology.
Why It Feels Like the Earth Is Shaking More
If you follow the news, earthquakes seem to be everywhere. A study examining over 320,000 news articles from 32 newspapers worldwide found that coverage is wildly uneven: about 71 percent of all earthquake-related news in 2015 was dedicated to just three events, out of more than 1,500 earthquakes of magnitude 5 or greater that year. Nearly half the articles used words like “devastating,” “catastrophic,” or “monster,” and roughly 77 percent focused on counting fatalities or describing destruction rather than providing context about how common such events actually are.1Copernicus Publications (Geoscientific Communication). Seismic risk: the biases of earthquake media coverage The result is a mental catalog skewed toward horror. When three or four destructive earthquakes happen in quick succession anywhere on the planet, the coverage creates an impression of escalation that the underlying data does not support.
Social media amplifies the effect further. Smartphone video of collapsing buildings reaches millions within minutes, something that was impossible even twenty years ago. The emotional weight of watching real-time destruction makes it hard to step back and ask whether there were just as many damaging earthquakes in, say, the 1950s. There were. The difference is that a lethal earthquake in a remote part of Iran or Peru in 1960 might have produced a single wire-service dispatch; today it generates thousands of posts and hours of footage.
What the Seismic Catalog Actually Shows
The most direct way to test whether large earthquakes are increasing is to take the full catalog of magnitude 7-plus events since 1900 and run it through statistical tests designed to detect non-random clustering. One widely cited analysis did exactly that, using three classes of statistical tests on the global record of M ≥ 7 earthquakes. The data could not reject the simplest possible explanation: that large earthquakes occur as independent random events at a roughly constant rate, with localized aftershock sequences layered on top.2Geophysical Research Letters. Random variability explains apparent global clustering of large earthquakes In plain terms, the earthquakes that seem to come in bunches are the same kind of bunching you would see if you flipped a coin and got five heads in a row. It feels meaningful, but it is within the range of what randomness produces.
A separate analysis focused specifically on great earthquakes, the truly enormous events above magnitude 8 or so. At a threshold around magnitude 8.4 to 8.5, the record does deviate slightly from a purely random process, with fewer events than expected in some windows. But even that deviation is not strong enough to conclude that great earthquakes cluster in any systematic way.3Geophysical Research Letters. Recurrence statistics of great earthquakes The bottom line from the statistics is unambiguous: if you look at the global rate of large earthquakes decade by decade, it wobbles, but it does not trend upward.
One reason the wobbling stands out is that the catalog itself has changed. In the early twentieth century, seismographs were sparse, poorly calibrated, and concentrated in Europe and North America. Many moderate earthquakes in Africa, Southeast Asia, and South America simply went unrecorded. As instruments spread across the globe and improved in sensitivity, the number of detected earthquakes climbed steeply, but that climb reflects better listening, not louder shaking.
How Better Instruments Changed the Count
The number of recorded earthquakes per year has risen dramatically since the mid-twentieth century, particularly for small and moderate events. This increase is almost entirely an artifact of detection. Modern seismic networks use thousands of stations around the world, many of them in boreholes or on quiet bedrock sites specifically chosen to minimize noise. Research on network performance has shown that a few quiet, well-placed stations improve detection far more than scattering many noisy ones across a region.4Seismological Research Letters. Modeling seismic network detection thresholds using production picking algorithms As networks gained these high-quality stations through the late twentieth century, the threshold for the smallest detectable earthquake dropped steadily. Events that would have gone unnoticed in 1960 now show up clearly in the data.
Consider what this means for someone scanning a catalog. In the 1930s, the global network could reliably detect earthquakes down to about magnitude 5. Today, dense regional networks in places like California or Japan can pick up events below magnitude 1. The catalog for small earthquakes has exploded in size. For large events, roughly magnitude 7 and up, the catalog is nearly complete going back to 1900 because those earthquakes send waves strong enough to register on even primitive instruments anywhere on the planet. This is why the “are earthquakes increasing” question has a clear answer at the top end of the magnitude scale and a more complicated answer at the bottom: the big ones have always been well counted, and their rate is flat.
The Exception That Proves the Rule: Human-Caused Earthquakes
There is one place where earthquakes genuinely did increase, and the cause was not tectonic. Oklahoma went from averaging one or two magnitude 3-plus earthquakes per year before 2009 to hundreds per year by 2014 and 2015. The sharp rise was tied directly to the injection of wastewater from oil and gas operations deep underground. Research found that the depth of injection relative to the crystalline basement rock was the strongest predictor of seismic activity, with the combination of injection depth and volume being critical. Restricting injection to 200 to 500 meters above the basement could reduce annual seismic energy release by a factor of roughly 1.4 to 2.8.5PubMed. Oklahoma’s induced seismicity strongly linked to wastewater injection depth After regulations tightened around injection volumes and depths in 2015 and 2016, the seismicity rate dropped substantially, confirming the link.
Oklahoma is the most dramatic example, but induced seismicity happens wherever humans inject or extract large volumes of fluid underground. The mechanism involves changing pore pressure in rock that is already close to its breaking point. Interestingly, the relationship between injection and earthquakes is not linear over time. Research on long-term, low-pressure fluid injection found that seismicity tends to be highest during the early phases of injection, when the rate of pore pressure change is greatest, and then decreases over time even as injection continues.6PubMed Central. Growth and stabilization of induced seismicity rates during long-term, low-pressure fluid injection This pattern suggests that the reservoir adjusts, and the earthquake risk front-loads itself rather than building indefinitely.
Reservoirs behind large dams can also trigger earthquakes through a related mechanism. The weight of impounded water creates elastic stress changes on nearby faults, and water seeping into the subsurface alters pore pressure at depth.7Geophysical Journal International. Reservoir related seismicity changes around the Gotvand Dam (south west of Iran) These reservoir-induced earthquakes are typically small, but they have occasionally reached damaging magnitudes near large dams in seismically active areas. As more dams are built in developing countries and geothermal energy extraction expands, induced seismicity is a growing concern, even though tectonic seismicity itself is not increasing.
How Earthquakes Trigger Each Other
Even without human interference, earthquakes do not happen in isolation. A large earthquake changes the stress field on surrounding faults, sometimes pushing them closer to failure. This process, known as Coulomb stress transfer, helps explain why earthquakes sometimes arrive in sequences that feel like escalation. The 2016-2017 central Italy sequence is a vivid example: a magnitude 6.0 event in August 2016 was followed by a magnitude 6.5 event in October and several magnitude 5-plus events into January 2017, all on a connected fault system where each rupture loaded the next segment.8Geophysical Journal International. Coulomb stress transfer and fault interaction over millennia on non-planar active normal faults: the Mw 6.5–5.0 seismic sequence of 2016–2017, central Italy
A similar pattern played out during the devastating 2023 Turkey-Syria earthquake sequence, where analysis of the East Anatolian Fault Zone showed that stress transferred between mainshocks contributed to the sequence’s spread. Shallow areas near the source zones received positive stress changes of around 1.0 bar, enough to accelerate failure on neighboring fault segments.9Turkish Journal of Earth Sciences. Investigation of earthquake sequence and stress transfer in the Eastern Anatolia Fault Zone by Coulomb stress analysis These cascades are not evidence of a global increase. They are the normal mechanical process by which stress redistributes through the crust. But when they happen in populated areas and produce a drumbeat of headline events over weeks or months, they contribute powerfully to the feeling that earthquakes are accelerating.
Could Climate Change Make Earthquakes More Common
This question comes up regularly, and the honest answer is that the connections are real but tiny compared to tectonic forces. Faults in the Earth’s crust are already under enormous stress, and very small perturbations can theoretically nudge one over the edge. Research has found that elastic stress changes as low as 0.07 kilopascals and pore-fluid pressure changes as low as 0.5 kilopascals have been proposed to naturally trigger earthquakes, with rainfall, snowpack, and atmospheric pressure all acting as potential modulators at local scales.10Geological Society of America Special Papers. Earthquake weather and climate change: Should we stress about the forecast? These are extraordinarily small stresses. To put them in perspective, tectonic stress on a major fault is millions of times larger. The climate-related effects are more like the last straw on a camel whose back was already about to break.
Over longer timescales, two climate-related mechanisms get more attention. The first is postglacial rebound: as ice sheets melt, the land they compressed slowly rises, changing stress patterns in the crust. In Greenland, for instance, earthquakes have been linked to the flexural stress created by ongoing deglaciation. A magnitude 5.1 event in 1971 and a notable earthquake in 1987 were both interpreted as normal faulting triggered by deglaciation-related stress, and the sparse general seismicity of Greenland can be related to the same rebound process.11Bulletin of the Seismological Society of America. The Greenland earthquake of 11 July 1987 and postglacial fault reactivation along a passive margin As modern ice sheets in Greenland and Antarctica continue to lose mass, this mechanism could produce more low-level seismicity in those regions, though the events are expected to remain small.
The second mechanism involves rising sea levels. Changes in eustatic sea level since the last ice age created a differential load across coastlines worldwide. The plate bending that results from this ocean loading alters the state of stress in the lithosphere within a certain distance of the coast.12Journal of Geophysical Research: Solid Earth. Ocean loading effects on stress at near shore plate boundary fault systems As sea level continues to rise, this effect will slowly evolve. But slowly is the key word. These stresses accumulate over centuries and millennia, not years. No credible seismologist argues that climate change will produce a noticeable increase in damaging earthquakes within our lifetimes. The tectonic engine that drives major earthquakes operates on forces that dwarf anything the atmosphere or oceans can deliver.
Why the Damage Keeps Getting Worse
If earthquakes are not increasing, why does the toll seem to grow? The answer is straightforward: more people live in earthquake-prone areas, and how they build matters enormously. A global analysis found that more than two-thirds of population growth in earthquake-prone areas, representing about 70 percent of the total population in those zones as of 2015, occurred in developing countries where the urbanization ratio sits between 20 and 60 percent. Nearly three-quarters of earthquake-related deaths, roughly 308,000, occurred in those same countries.13International Journal of Disaster Risk Science. A Global Analysis of the Relationship Between Urbanization and Fatalities in Earthquake-Prone Areas
The relationship between urbanization and earthquake fatalities is not simple, though. The same study found that once a region crosses an urbanization ratio of about 40 to 50 percent, earthquake fatalities actually start to decline even as population continues to grow. The reason is that more urbanized societies tend to enforce building codes, invest in infrastructure, and develop emergency response systems. Raw population size drives fatalities up, but the organized infrastructure that comes with mature urbanization pushes them back down. The deadliest scenarios are in countries that are urbanizing rapidly but have not yet developed the regulatory capacity to ensure buildings can withstand shaking, places like parts of Turkey, Iran, Nepal, and Haiti.
This means the rising death tolls from earthquakes are not a seismological problem. They are a development problem. An earthquake of identical magnitude can kill tens of thousands in one country and almost nobody in another, depending entirely on the quality of construction and emergency preparedness. Japan’s magnitude 9.0 earthquake in 2011 killed most of its victims through the tsunami, not building collapse, because Japanese construction standards are among the world’s strictest. The 2010 Haiti earthquake, at magnitude 7.0, killed over 200,000 people, largely because of unreinforced concrete buildings. The earthquakes did not differ in kind. The societies did.
How Scientists Establish Long-Term Earthquake Rates
The instrumental record only goes back about 120 years, which is a blink in geological time. To understand whether the current rate of large earthquakes is normal, seismologists turn to paleoseismology: digging trenches across faults and dating the layers of disrupted soil to determine when past earthquakes occurred. New Zealand, for example, recently produced a publicly available database of maximum-likelihood recurrence intervals and their uncertainties for 80 paleoseismic sites across the country, forming the largest such dataset in New Zealand to date.14Seismological Research Letters. Paleoseismic Earthquake Recurrence Interval Derivation for the 2022 Revision of the New Zealand National Seismic Hazard Model This kind of work reveals that individual faults have recurrence intervals ranging from hundreds to tens of thousands of years, with significant uncertainty.
The paleoseismic record consistently shows that large earthquakes have always come in clusters and gaps when viewed from a human timescale. A fault might produce three large events over two centuries and then go quiet for a thousand years. Across many faults globally, these clusters average out to the roughly constant rate visible in the modern instrumental catalog. What changes is which faults are active in any given century. The 2004 Sumatra and 2011 Japan megathrust earthquakes felt like an unprecedented era, but the paleoseismic record shows comparable clusters in previous centuries, long before anyone was keeping count.
Deep Earthquakes and Other Oddities
Most earthquake discussions focus on shallow events in the upper 20 kilometers of the crust, where almost all damaging earthquakes originate. But earthquakes also occur at depths of several hundred kilometers, within subducting slabs of oceanic crust diving into the mantle. These deep-focus earthquakes have their own puzzles. Research on 19 large deep-focus earthquakes near the Izu-Bonin subduction zone found evidence for two separate mechanisms at work, depending on the earthquake’s size: smaller events appear driven by transformational faulting of a mineral called metastable olivine, while larger events involve thermal runaway in surrounding material.15Geophysical Journal International. Moment-dependent rupture properties of deep-focus earthquakes in the Izu-Bonin subduction zone
One intriguing study looked for seasonal patterns in large deep-focus earthquakes and found a surplus of events during northern hemisphere summer. Taken at face value, the pattern was statistically significant. But the researchers themselves cautioned that the hypothesis was developed after looking at the data, that smaller deep earthquakes showed no such seasonality, and that no known physical mechanism could explain how surface seasons would affect processes hundreds of kilometers down. They could not rule out that the pattern was simply random chance.16Geophysical Research Letters. Possible seasonality in large deep‐focus earthquakes This kind of finding illustrates how seismology is full of suggestive patterns that dissolve under scrutiny, a useful reminder when evaluating claims about earthquake trends.
Deep earthquakes rarely cause significant damage at the surface because the energy dissipates over the long path from source to ground level. But they matter for understanding how the Earth works. They also contribute to earthquake catalogs, and a rising count of detected deep events, made possible by improved global networks, can inflate the total number of recorded earthquakes without implying any change in tectonic behavior.
What Earthquake Statistics Can and Cannot Tell You
Earthquake occurrence follows well-known statistical patterns. For any region, the relationship between earthquake magnitude and frequency is remarkably consistent: for every tenfold increase in the number of small earthquakes, you get roughly one-tenth as many large ones. This scaling relationship, established in the mid-twentieth century, holds across wildly different tectonic settings.17Journal of Geophysical Research: Solid Earth. Derivation of the complete Gutenberg‐Richter magnitude‐frequency relation using the principle of scale invariance The distribution of earthquake counts in catalogs can be modeled using standard probability distributions, with the best fit depending on the time window and magnitude range being considered.18Geophysical Journal International. Statistical distributions of earthquake numbers: consequence of branching process
What these statistical regularities tell you is that the earthquake machine operates with a kind of steady-state randomness. It does not speed up or slow down on human timescales. What it does do is fluctuate. A decade with an unusual number of magnitude 8-plus events is followed by a quieter decade, and the long-term average holds. The 2000s and early 2010s happened to include the 2004 Indian Ocean earthquake, the 2010 Chile earthquake, the 2010 Haiti earthquake, and the 2011 Japan earthquake. That cluster was terrifying and deadly, but it was not anomalous when measured against the full century of data. A similarly busy period occurred in the 1950s and 1960s, including events in Kamchatka, Chile, and Alaska that rank among the largest ever recorded.
The practical upshot is that earthquake preparedness cannot be based on trend-watching. You cannot look at a quiet decade and conclude the risk has decreased, and you cannot look at a busy one and conclude it has increased. The risk on any given fault is determined by its geology, its stress state, and its recurrence history, not by what has been happening globally. For anyone living in an earthquake-prone area, the question is never “are earthquakes increasing” but “is my building designed to survive the earthquake that my region’s geology says is overdue.”