What Causes Earthquakes at Transform Plate Boundaries?

Earthquakes at transform plate boundaries happen because two slabs of Earth’s crust grind horizontally past each other, and friction along their shared fault locks them together until the accumulated stress snaps the rock loose in a sudden jolt. This stick-slip process, where the fault alternates between being stuck and rapidly sliding, is the fundamental engine of transform-boundary seismicity. But the details beneath that simple summary are surprisingly varied: some parts of a transform fault creep along quietly while neighboring sections stay locked for centuries, the minerals lining the fault influence whether it slips gently or violently, and a single large earthquake on one segment can load stress onto the next segment and set the clock ticking toward its own rupture.

Stick-Slip Friction and Elastic Rebound

When two tectonic plates slide past each other along a transform fault, the motion is not smooth. Rough surfaces and enormous confining pressure keep the rock locked in place for years, decades, or centuries, while the plates continue to be pushed by deeper forces. The rock on each side of the fault bends and stores elastic energy, much like a spring being compressed. When the stress finally exceeds the frictional strength holding the fault together, the locked patch ruptures and the two sides lurch past each other, releasing that stored energy as seismic waves. This is the elastic rebound model, and laboratory experiments on rock samples show that the same stick-slip physics observed in small-scale friction tests successfully explains what happens during real earthquakes.

¹

The amount of energy released depends on how large the locked patch was and how much strain built up before failure. A small locked patch that breaks every few years might produce a moderate earthquake. A long segment that stays locked for centuries can release enough energy for a magnitude 7 or greater event. The critical variable is friction: specifically, how the friction on a given patch of fault changes as the rock starts to slide.

Locked Patches, Creeping Sections, and Frictional Behavior

Not every stretch of a transform fault behaves the same way. Some sections creep along almost continuously, accommodating the plate motion without producing large earthquakes. Others remain stubbornly locked and only release their strain in violent ruptures. The difference comes down to how the rock’s frictional resistance responds to motion. In sections where friction increases as sliding speeds up, the fault tends to slow itself down and creep steadily. Where friction drops as sliding speeds up, small movements accelerate into runaway slip, and earthquakes result.

Studies on the North Anatolian Fault in Turkey illustrate this nicely. Along the Ismetpasa segment, the shallow portion of the fault creeps because friction there strengthens with velocity, but at around three to six kilometers deep, conditions transition to a regime where friction weakens with velocity, and that deeper zone stays locked between earthquakes. On fully locked segments of the same fault, the friction-weakening behavior extends all the way to the surface.

²

The San Andreas Fault shows similar patchwork behavior. Around Parkfield, California, sections that creep steadily are separated by distinct locked patches that rupture in repeated moderate earthquakes. The creeping areas actually shape how and where ruptures start and stop, because they act as barriers or buffers between the zones that store energy.

³ A Bayesian analysis of the broader creeping section of the San Andreas identified discrete locked asperities within what had been assumed to be a uniformly sliding zone, and those hidden locked patches are consistent with magnitude 6 or greater earthquakes over the past century.

⁴

What Fault Minerals and Fluids Contribute

The rocks lining a transform fault are not inert bystanders. Over millions of years of grinding, the fault zone develops a core of finely crushed rock called gouge, and the minerals that crystallize within that gouge control how easily the fault slips. Core samples drilled directly from the San Andreas Fault at depth reveal thin films of smectite-group clay minerals coating the slip surfaces. These clays are weak and slippery, and their presence from near the surface down to roughly ten kilometers means they can promote creep behavior across a broad depth range.

⁵

In the actively deforming zones of that same borehole, magnesium-rich smectite clays are the dominant mineral phase, apparently formed by the interaction of fault-zone fluids with entrained serpentinite. Researchers propose that the development of interconnected networks of these low-friction clays, combined with stress-driven dissolution and re-precipitation of minerals, weakens the fault down to at least three kilometers.

⁶

Fluids play a second role beyond mineral chemistry. When pore-fluid pressure rises inside the fault zone, it effectively pushes the two sides apart and reduces the normal stress clamping them together. This makes the fault easier to break. In large strike-slip fault zones, the buildup of high-pressure fluids in compartmentalized cells within phyllosilicate-rich gouge can produce spatial weakening of the fault, favoring small earthquakes and fault creep rather than rare catastrophic ruptures.

⁷ On the San Andreas specifically, variations in pore pressure have been linked to episodic creep events that change both the timing and location of slip along the fault.

⁸

How One Earthquake Can Trigger the Next

Transform boundaries are not isolated faults; they are systems of interconnected segments. When a large earthquake ruptures one segment, it redistributes stress to neighboring segments. If that added stress pushes an already-strained segment closer to failure, it can trigger another earthquake, sometimes within months, sometimes decades later.

The North Anatolian Fault provides one of the clearest examples. Since 1939, a remarkable sequence of large earthquakes has marched progressively westward along the fault, each one apparently loading the next segment. Modeling of this sequence shows that the delayed triggering can be explained by slow-moving viscoelastic stress fronts in the upper mantle that gradually increase the failure stress at adjacent sites. The stress boosts delivered by this mechanism can exceed the values from the initial elastic jolt by as much as 177 percent, and the apparent threshold stress at which successive earthquakes occurred in the sequence ranged from roughly 0.02 to 3.15 megapascals.

⁹

This cascading behavior means the history of past earthquakes on a transform boundary directly affects where the next one is most likely. It also complicates forecasting, because the clock on each segment depends not just on how fast the plates are moving but on what neighboring segments have done recently.

Fault Geometry and Rupture Jumping

Transform faults are rarely perfectly straight lines. They have bends, jogs, and step-overs where the fault trace shifts laterally. These geometric complications matter because they can either stop a rupture in its tracks or, under the right conditions, allow it to jump from one fault segment to the next.

Where a step-over exists, the gap between two parallel fault strands can create a pull-apart basin, a depression formed by the extension between the offset segments. Recent three-dimensional dynamic rupture simulations based on conditions at the Dead Sea Basin show that these basins greatly enhance the ability of a rupture to jump across a step-over. The basin geometry reduces the stress concentration needed to nucleate rupture on the second fault, and the effect is especially strong for basins extending to about eight kilometers deep. Super-shear patches, where the rupture moves faster than the rock’s shear wave speed, were commonly observed on the second fault segment during these simulations.

¹⁰

On the other hand, damage in the rock surrounding a fault can have the opposite effect. Pre-existing fractures and off-fault deformation absorb energy from a propagating rupture, reducing the amount of slip on the main fault itself. There is a statistically significant inverse relationship between the degree of damage in the surrounding rock and the amount of slip during an earthquake.

¹¹ In soft sediments near the surface, this energy dissipation can slow or even halt rupture propagation entirely, which helps explain why shallow earthquakes in these settings tend to release less radiated energy.

¹²

Continental Versus Oceanic Transform Faults

Transform boundaries come in two broad flavors: continental ones like the San Andreas and Alpine faults, and oceanic ones that offset mid-ocean ridges. The two behave quite differently. Oceanic transforms, especially at fast-spreading ridges, tend to have simple thermal structures, short earthquake cycles, and high levels of foreshock activity. Continental transforms are messier. They do not neatly follow the simple rules of rigid plate motion, and their earthquakes can reach depths of around 20 kilometers or more in shortening segments where the geometry compresses the crust.

¹³

Oceanic transforms also extend seismicity to unexpected depths. Along the Romanche mega transform fault in the equatorial Atlantic, a band of microseismicity ten to fifteen kilometers thick was detected at depths of ten to thirty-four kilometers, within mantle rock at temperatures of 700 to 900 degrees Celsius. That is well below where rock is supposed to be too hot to fracture in the classic brittle sense. The finding suggests that while large earthquakes occur in the upper brittle crust, a substantial amount of deformation is accommodated by semi-brittle behavior in the deeper mantle.

¹⁴

This means our picture of transform earthquakes as purely shallow, brittle-crust phenomena is incomplete. Some fraction of the seismic energy budget at oceanic transforms involves hot, ductile rock breaking in ways that blur the textbook boundary between brittle and plastic deformation.

Slow Slip Events and Non-Volcanic Tremor

Between the dramatic earthquakes, transform faults are not always quiet. Sensitive instruments detect a range of subtle, slow-motion slip events that sit in a gray zone between full earthquakes and steady creep. Non-volcanic tremor is a weak, long-duration seismic signal observed episodically on some major faults, often in conjunction with slow slip events.

¹⁵

On the Parkfield segment of the San Andreas Fault, researchers have found that low-frequency earthquakes increase following slow slip events, pointing to a continuum between aseismic and seismic slip rather than a clean division. Transient deformation in creeping segments can perturb stress in adjacent locked areas, potentially promoting seismic activity.

¹⁶

The practical significance is that these quiet signals may carry clues about future large earthquakes. Temporary accelerations in creep rate have been proposed as a mechanism for triggering earthquakes on locked patches. Monitoring them is now a priority in earthquake-hazard science, with tools including tiltmeters, strainmeters, satellite-based radar, and fiber-optic sensing networks.

Supershear Ruptures on Straight Faults

Most earthquake ruptures propagate at speeds below the shear wave velocity of the surrounding rock. But on transform faults with exceptionally straight, continuous traces, ruptures sometimes accelerate past the shear wave speed in what is called supershear propagation. All documented cases of supershear rupture are associated with faults that have simple geometry: remarkably linear, narrow, and lacking segmentation features.

¹⁷

A vivid example is the 2020 magnitude 7.7 earthquake along the Oriente Fault in the Cayman Trough. The rupture began at a sub-shear speed for the first 40 to 50 kilometers, then accelerated to supershear speed and maintained it for roughly 200 kilometers along a very linear, unsegmented portion of the fault. Interestingly, supershear ruptures produce unusually low ground accelerations in the near field and low aftershock productivity, because the energy radiates in a different pattern, somewhat analogous to a sonic boom. The Oriente Fault segment had not experienced significant seismic activity for at least a century and had a shallow locking depth, both factors that may have set the stage for such a long supershear event.

¹⁸

The paradox of supershear earthquakes is that while the rupture speed is extreme, the shaking at the surface can actually be less intense close to the fault than for a comparable sub-shear event. The danger shifts to locations farther away, where the concentrated wave front, the seismic equivalent of a sonic boom, can arrive with more force than expected.

How Regular Are Transform-Fault Earthquakes

One of the persistent questions about transform boundaries is whether their earthquakes recur on a predictable schedule. The answer is complicated, and probably less reassuring than you would hope. Paleoseismic trenching along the East Anatolian Fault’s Palu-Lake Hazar segment yielded an estimated recurrence interval for surface-rupturing magnitude 7 or greater earthquakes of between roughly 65 and 360 years.

¹⁹ That is a wide range, and it reflects real variability in how quickly stress re-accumulates and how interactions with neighboring segments speed up or slow down the process.

Along the Dead Sea Transform, a two-thousand-year record of surface-rupturing earthquakes shows that the time between events varies by a factor of two to four. The pattern is not time-predictable, meaning you cannot simply add a fixed interval to the last earthquake and know when the next one is due. Instead, the interplay between adjacent fault segments, which can produce earthquakes of varying sizes, creates a recurrence pattern that looks irregular over human timescales even though the long-term slip rate stays roughly constant.

²⁰

New Zealand’s Alpine Fault is sometimes cited as a counterexample because its paleoseismic record suggests a relatively regular recurrence interval. Research there has progressed from establishing that large surface-rupturing earthquakes do occur at all, to estimating recurrence behavior, rupture extents, and the probability of the next event.

²¹ Even so, “relatively regular” for a fault still means substantial uncertainty in any given century.

Human Activity and Transform-Boundary Seismicity

Transform faults are not immune to human influence. In parts of the central United States, unprecedented rates of seismicity have followed increases in the volume of wastewater being injected underground from oil and gas operations. While the region’s faults are intraplate rather than plate-boundary features, the physics applies anywhere pre-existing faults exist under stress: pumping fluid underground raises pore pressure, reduces effective friction, and can push a fault past its failure threshold. The relationship between injected volume and the number of induced events is roughly linear.

²²

Near major continental transform faults, where tectonic stress is already high, the margin of extra fluid pressure needed to trigger slip could be quite small. This makes careful management of subsurface fluid injection especially important in regions close to active transform boundaries. The fluid-weakening mechanism is the same one that operates naturally inside faults, as described earlier for phyllosilicate-rich gouge zones; human activity just accelerates the pressure change.

Strike-Slip Tectonics Beyond Earth

Transform-style faulting is not unique to our planet. On Jupiter’s moon Europa, surface features called Ménec Fossae have been interpreted as products of transtensional (strike-slip with an extensional component) tectonic activity. The orientation and relationships of the tectonic features are consistent with horizontal shearing above a possible shallow subsurface water reservoir.

²³

Europa lacks plate tectonics in the terrestrial sense, but tidal forces from Jupiter stretch and compress its icy shell in ways that produce shearing stresses remarkably similar to those at transform boundaries on Earth. Whether the slip on Europa’s faults is seismic, aseismic, or something entirely different remains an open question. But the fact that the same geometric and stress relationships show up on a moon made of ice, driven by tidal forces rather than mantle convection, suggests that the basic mechanics of strike-slip faulting are universal wherever you have a rigid shell being sheared.