Is Fracking Causing Earthquakes? What Science Says

Fracking-related activities are unambiguously causing earthquakes in multiple regions around the world, and the scientific evidence linking underground fluid injection to seismic events has grown overwhelming over the past two decades. But the relationship is more layered than headlines suggest. The biggest driver of the most damaging induced quakes is not the hydraulic fracturing process itself but the massive volumes of salty wastewater that get pumped deep underground afterward. Fracking proper has also triggered earthquakes directly, including events strong enough to be felt at the surface, though typically smaller in magnitude. Understanding which part of the process is doing what, and why some areas shake while others do not, matters for how the risk gets managed.

How Underground Injection Triggers Fault Slip

Earth’s crust is laced with faults, many of which sit under enough tectonic stress that they are close to slipping on their own. When large volumes of fluid are pumped underground, the added pressure in the rock’s pore spaces reduces what geophysicists call the effective normal stress clamping a fault shut. Think of it like adding water under a heavy block resting on a tilted surface: the block was already close to sliding, and the water lifts it just enough to let gravity do the rest. This mechanism is well established in the literature and is the primary way injection triggers seismicity.¹

Pore pressure changes alone do not tell the whole story, though. The standard model of injection-induced quakes focuses solely on pore pressure, but researchers have shown that full poroelastic coupling also matters: injecting fluid physically deforms the surrounding rock, and those stress changes can reach faults that the fluid itself has not yet touched.² In a shale gas field in China, researchers demonstrated that stress transfer traveling ahead of the zone of fluid infiltration could trigger quakes at distances much farther than pore pressure diffusion alone would explain, and that these stress signals can travel almost ten times faster than the creeping front of pressurized rock.³ This helps explain why some induced earthquakes show up at surprising distances from the injection point.

A critical geological factor is the presence of faults in the crystalline basement rock beneath the softer sedimentary layers where injection typically takes place. Most of the largest induced earthquakes occur on these deeper basement faults. Even when a fault is hydraulically isolated from the injected fluid, poroelastic stresses transmitted downward can trigger slip without elevated pore pressure ever reaching the fault itself.⁴ This is one reason that apparently distant or deep earthquakes can still be linked to shallow injection operations.

Wastewater Disposal Is the Bigger Culprit

Hydraulic fracturing produces enormous quantities of water that flows back to the surface mixed with salts, naturally occurring radioactive materials, and hydrocarbons. The cheapest way to deal with this “produced water” has traditionally been to reinject it into deep rock formations, often the porous Arbuckle Group in Oklahoma or similar disposal zones elsewhere. This saltwater disposal happens at far greater cumulative volumes and over much longer periods than the fracturing itself, and it is the activity most strongly linked to the surge in induced earthquakes across the central United States.

Oklahoma’s experience is the most dramatic example. In areas that saw five- to ten-fold increases in saltwater disposal rates, earthquake counts shot up in lockstep. Adjacent areas with little disposal activity stayed quiet.⁵ In Alfalfa County, the number of magnitude 2.5 or greater earthquakes went from essentially zero in 2010 to over 300 in 2015 as disposal volumes climbed.⁶ The state went from experiencing a couple of magnitude-3 events per year to hundreds, becoming one of the most seismically active regions in the country.

When Oklahoma regulators ordered volume reductions and required some wells to be plugged back to shallower depths so that fluid could no longer reach the deeper Arbuckle formations, earthquake rates dropped substantially. Modeling suggests that without those plugbacks, seismicity rates today would be several times higher than what has actually been observed.⁷ The Oklahoma experience is essentially a natural experiment: ramp up disposal, quakes surge; scale it back, quakes decline. The correlation held across multiple counties and was strong enough to convince regulators, courts, and even skeptical industry players.

Fracking Itself Can Also Cause Earthquakes

While wastewater disposal drives the highest-profile earthquake surges, the hydraulic fracturing operation itself directly triggers seismicity in certain geological settings. The distinction matters because the two activities involve different volumes, pressures, and durations, and because fracking-induced quakes tend to be smaller (though not always small enough to ignore).

Near Fox Creek, Alberta, a sharp increase in earthquakes began in late 2013 tied to hydraulic fracturing in the Duvernay shale. Seismological analysis showed the earthquake sequences were triggered by pore pressure increases from fracturing operations reactivating pre-existing faults.⁸ Subsequent work found that the induced earthquakes were associated with completions that used larger injection volumes and that seismic productivity scaled linearly with injection volume.⁹ Additional earthquake clusters in the Duvernay East Shale Basin were later attributed to nearby fracking operations with greater than 99.7 percent confidence, including events reaching magnitude 4.18.¹⁰

In England, hydraulic fracturing at the Preston New Road site in Lancashire generated felt earthquakes during both of its well operations: a magnitude 1.6 event in 2018 and a magnitude 2.9 in 2019. Microseismic monitoring showed each operation activated different pre-existing faults with different orientations.¹¹ The 2019 event exceeded the UK regulator’s threshold and helped lead to a moratorium on fracking in England. The Preston New Road case is telling because it took place under intense monitoring and still produced events large enough to be felt and to cause minor cosmetic damage to nearby buildings.¹²

In some locations, the two activities work together. A case study of a magnitude 3.1 earthquake sequence in western Canada found that pore pressure diffusion from wastewater disposal and poroelastic stress transfer from nearby hydraulic fracturing likely acted collaboratively to trigger the event. The researchers concluded that long-term wastewater disposal stress should be considered even when analyzing earthquakes apparently triggered by fracking, especially when both operations are happening near the same faults.¹³

The Permian Basin and Multi-Cause Complexity

Texas’s Permian Basin illustrates how messy attribution can get in a region with decades of injection history and many overlapping operators. The basin has a long history of induced earthquakes, but rates have climbed steeply over the past two decades. In March 2020, the region experienced a magnitude 5.0 event likely induced by wastewater disposal. Analysis determined that the vast majority of seismicity near Pecos, Texas, since 2000 has been linked to an increase in disposal into wells injecting deeper than about 1.5 kilometers.¹⁴

The western sub-basin, the Delaware Basin, has at least three distinct earthquake systems running simultaneously: deep basement-fault earthquakes caused by nearby deep injection in New Mexico, basement-fault earthquakes caused by more distant injection in central Texas, and shallow earthquakes on faults confined within the sedimentary layers that appear to be driven by a combination of shallow injection, hydraulic fracturing, and even compaction of oil reservoirs during production.¹⁵ Researchers have shown that the widespread deep seismicity in the Delaware Basin is mainly driven by shallow wastewater injection transmitting poroelastic stresses downward through layers of shale that act as barriers to fluid flow but not to stress.¹⁶ In other words, the water never reached those deep faults, but the mechanical squeeze from all that injection did.

Why Earthquakes Can Continue After Injection Stops

One of the more unsettling aspects of induced seismicity is that earthquakes do not always stop when pumping does. In Oklahoma’s Alfalfa County, even after substantial reductions in wastewater injection, the average depth of earthquakes kept increasing at roughly 0.12 kilometers per year, suggesting pressure was still migrating downward into the basement rock.⁶ The earthquake count dropped, but the process did not simply switch off.

Studies of injection-induced seismicity across multiple settings find that the time lag between injection changes and seismicity response typically ranges from under 100 days to one to three years, depending on the distance to nearby faults and the permeability of the rock.¹⁷ In enhanced geothermal systems, multiple projects have observed the largest earthquake occurring after injection had already ceased, at magnitudes between roughly 1.6 and 5.5.¹⁸ This is a genuine hazard management challenge: you cannot simply shut the well and assume the risk is over.

The physics behind this trailing seismicity involves continued pore pressure diffusion after the pressure source is removed, as well as nonlinear interactions where the rock’s permeability changes in response to the fluid that was injected. Researchers have described how a “triggering front” of early, distant earthquakes and a “back front” of growing seismic quiescence both evolve according to predictable diffusion equations even after injection stops.¹⁹ The practical takeaway is that any mitigation strategy needs to account for a tail of seismicity that can last months to years.

Traffic Light Systems and How Risk Is Managed

The most widely used tool for managing fracking-induced seismicity is the traffic light system. Operators and regulators agree on magnitude thresholds: green means proceed normally, yellow means heightened monitoring or reduced injection rates, and red means stop immediately. The idea is simple, but the execution is tricky.

A global compilation of fracking-induced seismicity datasets found that magnitude jumps of less than two units account for the majority of cases, meaning quakes generally escalate somewhat gradually rather than leaping from nothing to a damaging event. About a quarter of cases experienced a post-injection magnitude increase, with the largest trailing increase being 1.6 magnitude units. Most trailing events occurred within days of stopping injection.²⁰ These findings mean that for a traffic light system to protect against, say, a magnitude 4.0 event, the red-light threshold should be set around magnitude 2.0 or lower to give enough buffer for potential jumps and trailing increases.

Some researchers advocate moving beyond one-size-fits-all thresholds. A study of the Eagle Ford shale in Texas developed a risk-informed strategy using spatially varying red-light thresholds ranging from magnitude 2.0 to 5.0, depending on the local population density and building vulnerability. This approach is fairer and safer than applying a single blanket threshold across a broad area, because a magnitude 3.0 event under a city is a very different problem than the same event under empty rangeland.²¹

Multi-well injection schemes may also help. Simulations show that distributing the same total injection volume across multiple wells, rather than pumping it all through one, can reduce the maximum earthquake magnitude.²² This makes intuitive sense: spreading the pressure perturbation over a larger area means less concentrated stress on any single fault.

Telling Induced Earthquakes From Natural Ones

Not every earthquake near a fracking operation is caused by that operation, and regulators need ways to tell the difference. In Ohio, researchers used waveform template matching to distinguish induced from natural seismicity. They found that induced earthquake sequences tend to behave like swarms, with many similar-sized events clustered tightly in space and time, as opposed to the more random occurrence patterns typical of natural quakes. Regions with previously documented seismicity but no nearby injection wells showed far fewer template matches and far less of this swarm-like clustering.²³

The shallow depth of induced earthquakes also helps. Induced events tend to occur at shallower depths than typical tectonic earthquakes, which has practical consequences for ground shaking: a shallow magnitude 3.0 event can produce stronger shaking at the surface than a deeper magnitude 3.0, meaning the felt impact of induced quakes is disproportionate to their size on paper.²⁴ Ground-motion models designed for natural earthquakes can underestimate shaking from induced ones if they do not account for this depth difference.

How Maximum Magnitude Relates to Injection Volume

One of the more pressing questions is whether we can predict how large an induced earthquake might get. Empirical models suggest a relationship between the volume of injected fluid and the maximum magnitude, and physics-based simulations confirm this scaling holds when the background stress on faults is moderate. When stresses are already close to failure, however, the scaling breaks down and larger-than-expected events become possible.²⁵

The injection rate, not just total volume, also matters. High flow rates produce more frequent seismic events that start at relatively low injected volume, which actually gives operators warning signs they can track. Low flow rates allow larger volumes to be injected without any seismic events at first, but when an earthquake does occur it can be abruptly large with no precursory activity.²⁶ This is a counterintuitive finding: pumping slowly and steadily might seem safer, but it can actually create conditions for a sudden large event. Faster pumping generates a stream of small quakes that act as an early warning system.

The Same Problem Shows Up in Geothermal Energy and Carbon Storage

Induced seismicity is not unique to the oil and gas industry. Enhanced geothermal systems, which inject water at high pressure to fracture hot rock and create underground heat exchangers, face the same challenge. Increased pore fluid pressure can reactivate pre-existing faults, and in some cases the resulting earthquakes have been severe enough to cause building damage and force project cancellation, as happened in Basel, Switzerland, and Pohang, South Korea.²⁷

Carbon capture and storage faces similar concerns. Proposals to inject large volumes of captured CO₂ into deep geological formations have raised warnings that earthquakes are likely to be triggered, and even small to moderate events could threaten the seal integrity of the storage reservoir.²⁸ The physics is somewhat different from single-phase water injection: because CO₂ migrates more slowly than the pressure pulse it creates, stress can build up on deep faults in a delayed fashion that may not respond immediately to shutting off injection.²⁹ These concerns do not mean carbon storage is impossible, but they do mean the seismic risks require the same careful site selection, monitoring, and pressure management that the oil and gas industry has been learning the hard way.

Reducing Disposal as a Path Forward

Since wastewater disposal is the dominant driver of the largest induced earthquakes, one strategy is simply to dispose of less water underground. Produced-water recycling and treatment technologies are advancing, and optimization frameworks now exist for planning how much water to reuse versus dispose of, balancing cost against injection risk. Work on the Permian Basin has modeled this as a multi-objective problem: maximize economic value, minimize disposal volumes, and minimize an injection-risk indicator, all while meeting regulatory and quality constraints.³⁰

Oklahoma’s experience already demonstrated that reducing injection volumes and moving injection to shallower zones away from basement faults works.⁷ The open question is whether industry and regulators in newer hotspots like the Permian Basin will act quickly enough. The Denver earthquakes of the 1960s, triggered by fluid waste injection at the Rocky Mountain Arsenal, first demonstrated the link between underground injection and seismicity over half a century ago.³¹ The science has been clear for a long time. What has lagged is the willingness to treat underground disposal as a seismic risk factor from the start, rather than waiting for the earthquakes to arrive and then scrambling to respond.