Counter-intuitive Behavior of Subduction Zones: Weak Faults Rupture, Strong Faults Creep

Abstract:

Subduction interfaces that produce great earthquakes are often said to be “strongly coupled”, and those that creep are said to be “weakly coupled”. However, the relation between the strength and seismogenic behavior of subduction faults is far from clear. Seismological and geodetic observations of earthquake rupture usually provide information only on stress change, not fault strength. In this study, we infer fault strength by calculating frictional heating along megathrusts and comparing results with heat flow measurements. We find that stick-slip megathrusts that have produced great earthquakes such as at Japan Trench and northern Sumatra have very low apparent friction coefficients (~ 0.02 – 0.03), but megathrusts that creep such as at Northern Hikurangi and Manila Trench have higher values (up to ~0.13). The differnce cannot be explained by coseismic dynamic weakening of the stick-slip megathrusts, because the average stress drop in great earthquakes is usually less than 5 MPa, equivalent to a coseismic reduction of apparent friction coefficient by less than ~0.01. Therefore our results indicate differences in the static strength of different subduction faults. Why are the creeping faults stronger? We think it is related to their creeping mechanism. Very rugged subducting seafloor tends to cause creep and hinder great earthquake rupture (Wang and Bilek, 2014). In contrast, all giant earthquakes have occurred at subduction zones with relatively smooth subducting seafloor. Large geometrical irregularities such as seamounts generate heterogeneous structure and stresses that promote numerous small and medium size earthquakes and aseismic creep. The creeping is a process of breaking and wearing of geometrical irregularities in a deformation zone and is expected to be against relatively large resistance (strong creep). This is different from the creeping of smooth faults due to the presence of weak fault gouge (weak creep) such as along the creeping segment of the San Andreas fault. The general correlation between subducting seafloor ruggedness, creeping, and greater heat dissipation, if further verified, provides a new perspective in assessing earthquake and tsunami hazards for risk mitigation.

Wang, K., and S. L. Bilek. Tectonophysics 610, 1–24 (2014).