Deep Earthquakes

Abstract

Below the Earth’s crust, earthquakes delineate subducting lithosphere down to nearly 700 km depth. Many properties of deep earthquakes are similar to those of shallow events, for example, their primarily double-couple nature and frequency–magnitude statistics. The most obvious difference lies in the much lower rates of seismicity and aftershock occurrence for earthquakes deeper than 70 km. As with shallow earthquakes, the rupture process is highly variable. Stress drops and rupture velocities differ subtly, rather than markedly, from those of shallow events. Some properties, including aftershock productivity, rupture duration and complexity, although highly variable, appear to change systematically with depth, undergoing abrupt changes near 550 km depth. The inferred thermal state of subducting slabs suggests that temperature controls the location of seismicity in different ways above and below about 350 km depth. A consideration of stress in subducting slabs points toward multiple sources of stress heterogeneity, and indicates that the stress drops of most deep earthquakes are incomplete. Spatial structures delineated by earthquakes in subducting slabs include double seismic zones, large shear zones, repeating earthquakes, outboard earthquakes, and remnant slabs. A number of physical mechanisms to generate deep earthquakes have been proposed, including dehydration embrittlement, transformational faulting, and thermal shear instability. Above about 350 km, accumulating evidence supports dehydration embrittlement on both reactivated and new faults as a plausible mechanism to generate intermediate-depth seismicity. Below 350 km it is not clear how earthquakes are generated. There is some evidence that more than one mechanism operates to generate seismicity below ∼350 km.