| Abstract: | Seismicity, deformation, state of stress, and abundance of fluids along subducting plate boundaries are reviewed, and the origin of large or great thrust-type earthquakes is discussed based on the recent experimental results on the slip behavior of halite and serpentine gouges.Shallow subducting plate boundaries above 20–25 km in depth are characterized by low seismicity, low tectonic stress, inter-plate decoupling, ductile deformation associated with the formation of metamorphic schistosity (except at very shallow depths), metamorphism suggesting solution processes on massive scale, and presence of abundant H2O. It is argued that these unique features are due to pressure-solution processes, to high fluid pressure, to low strength and stable behavior of clayey sediments under wet environments, and/or to the deformation of soft, unconsolidated sediments at very shallow depths. The low seismicity in this zone is in marked contrast with major strike-slip faults along which large earthquakes occur at depths shallower than 15–20 km. It is emphasized that these unique features are expected only for restricted regions where there is constant supply of H2O due to progressive metamorphism or where fluids in the rocks are trapped and cannot escape to the surface.Large or great thrust-type earthquakes in subduction zones initiate at depths of 30–50 km, below the shallow decoupled zone. In this focal depth range, the supply of H2O during progressive metamorphism perhaps diminishes downwards, the overriding and subducting plates are coupled and stick to each other during much of the inter-seismic period, and the resistance to slip (or shear stress) is presumably high. It is suggested that these earthquakes begin to occur at a depth where the plate-boundary zone becomes fairly dry. Deformation at these depths appears to be predominantly ductile, so that the earthquakes cannot be regarded simply as a brittle phenomenon. (1) Creep instability i.e., instability associated with plastic deformation, and (2) dehydration-induced instability are the most likely mechanisms for initiating the earthquakes, and both have some experimental support. Stick-slip of halite gouge while undergoing ductile deformation primarily by intracrystalline gliding is described and discussed as a supporting evidence for (1). Shear resistance of halite gouge increases with increasing confining pressure in stick-slip regimes. Hence the observed stick-slip may be a semi-brittle phenomenon with respect to the pressure dependence of the shear resistance, although the deformation texture cannot be distinguished from that formed by pressure-insensitive flow. Serpentine gouge exhibits violent stick-slip upon its decomposition under dry, not wet, environments, supporting the mechanism (2) above. Exact mechanisms which lead to the unstable fault motion are poorly understood as yet, but stick-slip of both halite and serpentine gouges is recognized only when the slip-rate dependence of friction is negative i.e., lower friction at faster slip rate, consistent with the theoretical prediction of Rice and Ruina (1983). There is a possibility that the thrust-type earthquakes can be explained essentially within the framework of fault constitutive laws developed by Dieterich (1979) and Ruina (1983). |
