Earthquakes are caused primarily by the sudden release of stored elastic strain energy within the Earth when rocks rupture or when an existing geological fault suddenly slips. Most earthquakes occur because tectonic plates are continually moving, but friction prevents many parts of their boundaries from sliding smoothly. Stress therefore accumulates in surrounding rocks until the forces acting on a fault overcome friction and the rocks move abruptly.
The resulting movement generates seismic waves that travel through the Earth and along its surface. The fundamental process is commonly described by elastic rebound theory, which states that rocks on either side of a locked fault can gradually deform as tectonic forces act on them, storing elastic energy. When rupture occurs, the rocks partially return towards a less strained configuration and the released energy propagates through the surrounding material.
Plate tectonics provides the principal large scale explanation for earthquake distribution. Earths lithosphere is divided into tectonic plates that move relative to one another over geological timescales. At convergent boundaries, plates move towards each other, producing powerful earthquakes associated with subduction zones and continental collision. At divergent boundaries, plates move apart and earthquakes commonly occur along spreading centres and normal faults.
At transform boundaries, plates slide horizontally past one another and earthquakes can occur when friction temporarily locks the fault before sudden movement. Earthquakes are not restricted to plate boundaries, however. Intraplate earthquakes occur within plate interiors when ancient faults are reactivated or stresses are transmitted through the rigid lithosphere. The distribution of earthquakes therefore reflects both present day plate motions and inherited geological structures.
Fault friction is a major factor controlling whether rocks remain locked or undergo repeated earthquakes. Fault surfaces are irregular and may be pressed together by the weight of overlying rocks and by tectonic forces. Friction can temporarily prevent sliding even when tectonic stress continues to increase. Fluids within fractures can alter effective pressure and influence fault strength, while temperature, rock type, mineral composition and depth also affect deformation.
Some faults creep relatively slowly and continuously, whereas others remain locked for long periods and then produce sudden earthquakes. The distinction between stable sliding and unstable rupture is important in understanding why earthquakes vary in frequency and magnitude. Fault behaviour can also change along a single fault, with some sections accumulating strain while neighbouring areas deform more gradually.
Earthquakes can also be triggered or influenced by processes other than ordinary tectonic plate motion. Volcanic activity can generate earthquakes when magma moves through fractures, changes pressure or causes surrounding rocks to crack. Large underground fluid injections can alter pore pressure and reduce effective friction on faults, producing induced seismicity in suitable geological settings. Activities associated with geothermal operations, wastewater disposal, mining and reservoir changes have all been associated with earthquake sequences in particular circumstances.
The filling of large reservoirs can alter stresses and pore pressures within the crust and has occasionally been linked to earthquakes. Underground explosions can also generate seismic signals, although their physical source mechanism differs from natural fault rupture. Human activity therefore can modify stresses and fault conditions, but most earthquakes worldwide remain natural tectonic events.
The depth and geometry of a fault influence how an earthquake develops. Shallow earthquakes can produce particularly strong shaking at the surface because seismic energy has less distance to travel before reaching buildings and infrastructure. Deeper earthquakes can occur within subducting slabs and may be felt over very large areas because their waves propagate through relatively competent rock. Some earthquakes rupture faults for hundreds of kilometres, while others involve relatively small patches of rock.
Rupture can propagate at several kilometres per second, causing different parts of a fault to fail sequentially rather than simultaneously. The direction and speed of rupture, fault orientation, depth, surrounding geology and amount of slip all influence the resulting seismic waves. These factors help explain why two earthquakes with similar overall energy can produce different patterns of shaking and damage.
Stress accumulation is commonly described using concepts from rock mechanics. Stress represents force distributed over an area, while strain describes deformation produced by stress. Rocks can behave elastically under suitable conditions, meaning they deform but can recover much of their original shape when stress is removed. If stress exceeds the strength of the rock or fault interface, brittle failure can occur. At greater depths and temperatures, rocks may deform more plastically, although earthquakes can still occur under particular conditions. The transition between brittle and ductile behaviour depends on temperature, pressure, rock composition, strain rate and the presence of fluids. Earthquakes therefore reflect the complex mechanical behaviour of the crust and upper mantle rather than simply the presence of movement between plates.
After a major earthquake, the surrounding crust can continue adjusting through aftershocks and other forms of postseismic deformation. Aftershocks occur because the main rupture changes the stress distribution around the fault and neighbouring faults. Some sections may become closer to failure, while others become less stressed. Slow afterslip can also occur when parts of a fault continue moving after the main rupture without producing a comparable sudden earthquake.
In some geological settings, one earthquake can alter the probability of subsequent earthquakes on nearby faults through stress transfer. This does not mean that every earthquake can be precisely predicted from another earthquake, because the crust contains complex three dimensional structures and uncertain frictional conditions. Earthquake triggering is therefore a statistical and mechanical concept rather than a simple deterministic chain.
The ultimate cause of most earthquakes is therefore the interaction between tectonic forces, rock deformation and frictional failure within the Earth. Plate movements provide the large scale driving forces, while local geology, fault geometry, stress state, rock properties and fluids influence exactly where and when rupture occurs. Earthquakes range from microscopic events that are detectable only with sensitive instruments to great earthquakes that rupture enormous areas of fault and release immense amounts of energy.
Their causes are studied through seismology, geology, geodesy, laboratory rock mechanics and numerical modelling. Although scientists understand the physical mechanisms that generate earthquakes and can identify regions of elevated seismic hazard, the exact time, location and magnitude of a future earthquake generally cannot be predicted deterministically. Hazard assessment therefore relies heavily on understanding faults, historical earthquakes, crustal deformation and probabilistic seismic behaviour.
