A fault breaks when the stress pushing rock along the fault plane exceeds the friction holding it in place. This tipping point depends on how fast stress accumulates, how strong the rock’s grip is, and whether outside forces like fluid pressure weaken that grip. The process can take centuries of slow buildup before releasing in seconds.
How Stress Builds Until Failure
Earth’s tectonic plates move constantly, but faults between them often stay locked by friction. As plates grind past or push against each other, the surrounding rock bends and stores energy like a stretched rubber band. The USGS describes this as the elastic rebound theory: rock on either side of a locked fault gradually deforms over decades or centuries, and when the stored stress finally overcomes friction, the fault snaps and the rock springs back to its original shape. That snap is the earthquake.
A classic way to visualize this: imagine a straight fence built across the San Andreas Fault. Over time, plate motion warps the fence into an S-shape. When the fault finally breaks, the two halves of the fence straighten out again, but they’re now offset from each other. That offset represents the slip, the distance the two sides moved during the quake.
The Breaking Point: Friction vs. Force
Whether a fault breaks comes down to a simple contest between two forces. Shear stress is the sideways force trying to slide one block of rock past another. Frictional resistance is what holds the fault together, and it depends on how hard the two surfaces are pressed against each other (the normal stress) and the friction coefficient of the rock.
At depths relevant to most earthquakes, rock friction follows a surprisingly universal pattern. Lab experiments show that at pressures below about 2 kilobars (typical of the upper crust), the shear stress needed to cause sliding is roughly 85% of the normal stress, regardless of rock type. At higher pressures, the relationship shifts slightly, but the key finding is that rock type barely matters. Granite, sandstone, limestone: once you’re deep enough, they all resist sliding by about the same amount. What changes from fault to fault is how much stress is being applied and how quickly it accumulates.
What Weakens a Fault’s Grip
A fault doesn’t have to wait for stress alone to push it past the breaking point. Fluid pressure inside the fault zone can do half the work. Water and other fluids naturally exist along faults at depth, produced partly by chemical reactions that release water from minerals. When the fluid pressure inside pore spaces rises, it effectively pushes the two sides of the fault apart, reducing the normal stress clamping them together. Less clamping force means less friction, and less friction means the fault can slip under lower shear stress than it otherwise would.
This is also why injecting fluid underground (during wastewater disposal or certain industrial operations) can trigger earthquakes on faults that were already near their breaking point. The added fluid pressure tips the balance. Research published in PNAS confirms that increasing pore pressure reduces frictional resistance and promotes slip, though higher fluid pressure can also increase the size of the zone that needs to fail before a full rupture begins, sometimes resulting in slow, stable sliding rather than a sudden quake.
Where Faults Can Break: The Depth Window
Faults only break in a brittle, snapping fashion within a limited depth range called the seismogenic zone. Near the surface, rock is cool and brittle enough to fracture. But as you go deeper, rising temperatures cause rock to behave more like putty, flowing slowly instead of snapping. Lab experiments on fault materials place this brittle-to-ductile transition at roughly 350 to 510°C, which corresponds to depths of about 10 to 25 kilometers in most continental settings, depending on the local heat flow.
Below that temperature threshold, rock deforms by flowing rather than breaking, so earthquakes become rare. Above it, in the shallowest few kilometers, rock may be too fractured and weak to store much elastic energy. The most dangerous earthquakes nucleate in the middle of the seismogenic zone, where rock is strong enough to lock and store stress but brittle enough to fail catastrophically.
Slow Slip vs. Sudden Rupture
Not every fault failure is violent. Some faults release their stored stress gradually over days, weeks, or even months in what geologists call slow slip events. The difference comes down to how friction behaves as the fault starts to move. On faults that produce earthquakes, friction drops sharply once sliding begins, allowing the rupture to accelerate. On faults that creep, friction actually increases at higher slip speeds, or fluid pressure changes within the fault gouge act as a brake, preventing runaway acceleration.
Some fault zones do both. The Cascadia subduction zone, for instance, produces slow slip events roughly every 14 months along its deeper portions while its shallower, locked sections continue building toward a future large earthquake. A fault’s behavior can vary along its length depending on local temperature, rock composition, and fluid conditions.
Can One Earthquake Trigger Another Fault to Break?
Yes, and it happens in two distinct ways. When a fault ruptures, it permanently reshuffles stress in the surrounding crust. Nearby faults that were already close to failure may get pushed over the edge by this static stress change. This effect is permanent: if the stress change brings a neighboring fault closer to failure, it stays closer even after the shaking stops. Conversely, if the stress shift moves a fault further from failure, it can actually delay its next earthquake.
The second mechanism is dynamic triggering, caused by the passing seismic waves themselves. These waves create temporary stress pulses that can cause nearly instantaneous failure on distant faults, but only if those faults are already very close to their breaking point and the pulse is large enough. Unlike static triggering, dynamic stress changes don’t have a lasting effect. They either cause immediate failure or nothing at all.
How Scientists Estimate When a Fault Is Due
Geologists use GPS networks to measure how fast the ground on either side of a fault is moving and, more importantly, how much of that motion is being stored as elastic strain rather than released by steady creep. The difference between the expected slip and the actual slip is called the slip deficit, and it grows over time on locked faults.
If GPS measurements show a fault is being loaded at about 1 millimeter per year, and a typical earthquake on that fault produces roughly 1 meter of slip, simple division gives a recurrence interval of about 1,000 years. For faster-moving faults like the San Andreas, loading rates of 20 to 30 millimeters per year mean large earthquakes recur on the order of every 100 to 300 years. These are rough averages, not predictions. Earthquakes are not periodic like clockwork, and a fault can break earlier or later than its average interval.
Warning Signs Before a Break
Some large earthquakes are preceded by smaller foreshocks near the eventual rupture point, occurring anywhere from weeks to mere seconds beforehand. But this pattern is unreliable. Not all large earthquakes have foreshocks, and not all swarms of small earthquakes lead to a larger one. There is currently no method that can look at a cluster of small quakes in real time and determine with certainty whether a bigger event is coming.
Other potential precursors, like changes in groundwater levels, shifts in ground tilt measured by sensitive instruments, or subtle changes in how seismic waves travel through a fault zone, have been observed before some earthquakes but not others. The fundamental challenge is that faults operate under enormous pressures miles underground, and the final moments before failure involve processes happening on scales too small and too deep to monitor directly.