The state of stress in the Earth's crust is a key to understanding crustal dynamics. Recent seismological observations in NE Japan clearly show spatial variations of stress orientations in the crust; the stress field is basically consistent with relative plate motions as a first-order approximation, but it rotates significantly beneath mountain ranges or in and around the source regions of large earthquakes. Such observations suggest that the deviatoric stress field produced by relative plate motions is small enough to be disturbed by other stress sources such as topography and earthquake faulting. According to quantitative modeling studies, the deviatoric stress magnitude in the seismogenic zone in NE Japan should be in the order of tens of MPa, which is too small to overcome laboratory-derived fault strength. This suggests that fault weakening mechanisms, including the low frictional coefficient of the fault surface and overpressurized fluids, play an important role in the generation of earthquakes. Deviatoric stress magnitude in seismogeneic zones can be lower than in stable continents if the fault weakening mechanism is related to the subduction process. Given that deviatoric stress magnitude is generally small in the seismogeneic zones of island arcs, various stress sources can lead to heterogeneous states of stress. It is important to take account of stress heterogeneity when modeling crustal dynamics.
Knowledge of the strength of faults in the continental upper crust is critical to our understanding of crustal stress states, coseismic faulting, and lithospheric deformation. In this paper, we investigate time- and displacement-dependent fault-zone weakening (softening) over geological time caused by the hydrothermal alteration of rock, the development of faulting-related structure and fabric, and changes in the relevant deformation mechanisms. In the shallow portion of the continental seismogenic zone (< 5 km), hydrothermal alteration induced by comminution and fluid flow along fault zones progressively enriches weak phyllosilicates. The development of phyllosilicate-aligned fabric with increasing shear strain leads to an effective weakening with increasing cumulative fault displacement. In the deep portion of the seismogenic zone (> 5 km), frictional–viscous flow occurs in combination with friction contributed by phyllosilicates and the dissolution–precipitation of clasts after the introduction of water, phyllosilicates and anastomosing fabrics all increasing with greater fault displacement. In addition, the water weakening of quartz and feldspar is an important softening process in the deeper portion of the seismogenic zone (> 10 km). The smoothing of fault-zone topography by the shearing of irregularities and asperities, as well as the thickening of the fault zone, leads to a reduction over time in the bulk frictional resistance of a fault as displacement increases. These time- and displacement-dependent weakening processes of fault zones give rise to diverse strength and stress states of the crust depending on its maturity and may provide clues to reconciling the stress–heat flow paradox of crustal faults.
Development of suitable algorithms and an increase in computational capability have enabled dynamic earthquake cycle simulations (ECS) to be conducted in which both coseismic rapid slip associated with inertial effects and interseismic quasistatic processes are simulated in a single framework. The rate- and state-dependent friction (RSF) law is a very useful tool in ECS, because it is able to reproduce a spectrum of fault behaviors including steady slip, aseismic transient, and earthquakes when coupled with an elastic medium. The RSF law has, however, been developed with a rather narrow range of experimental conditions where cataclasis dominates. Recent experimental and theoretical studies have developed fault constitutive laws that are applicable to different conditions where different deformation mechanisms are important. The ECS is useful for realizing and quantifying fault motions based on different hypotheses of fault slip deformation mechanisms, such as brittle-plastic transition at a deeper extent of a seismogenic fault, pressure solution creep, and remarkable weakening at a coseismic high slip rate. In the field of structural geology, conceptual fault models, which represents distribution of dominant deformation mechanisms and strength along the depth of a major fault, have been proposed and updated since the 1970s on a basis of field observations of fault rocks and laboratory experiments. Since the development of ECS, it has become possible to build them as objective numerical models once fault constitutive laws are formulated, and to compare behaviors under different hypotheses. Recent studies on the significance of changes in deformation mechanisms of a fault slip for fault behavior and reviewed and perspectives are discussed.
One important factor controlling crustal strength is fault zone development. A fault can mature through multiple stages where small faults are generated by rupturing homogeneous media, and are partially connected each other, with large deformations locally concentrated along a major fault. Fault geometry, stress field, and seismic activity, which are detectable with seismic observations, may depend on the stage of fault evolution. At the same time, fault distribution and slip properties observed in a geological survey also provide important information on constraints to fault development. In particular, comparing fault distribution and direction between seismological analyses and geological observations plays an important role in understanding the history of fault zone evolution. A bottleneck in the seismological approach to investigating fault zone development is low spatial resolution due to low spatial density and number of seismic stations. A key element for understanding fault zone development in the crust based on seismic observations is discussed. “0.1 manten” hyper dense seismic observations carried out in the focal area of the 2000 Western Tottori earthquake (M 7.3) are introduced along with the preliminary results of a data analysis.
GNSS data analyses reveal that recent large inland earthquakes in the Northeast (NE) Japan occurred in strain concentration zones. Seismic low-velocity anomalies, indicative of mechanically weak materials (weak zones), are estimated below the strain concentration zones at depths corresponding to the lower crust. Such crustal structures with weak zones have been formed as an accumulation of tectonic movements and igneous activities since early Miocene. Volcanic activity in the NE Japan during the Late Cenozoic Era can be subdivided into three prominent stages: continental margin volcanism stage, back-arc basin opening stage, and island-arc volcanism stage. The crustal structure of the NE Japan arc is characterized by many rift structures and large transcurrent faults formed during the back-arc basin opening stage, and by many large caldera volcanoes formed during the island-arc volcanism stage. The relationships among mechanically weak crustal structures, present strain localizations, earthquake distributions, and geological characteristics including rift structures, large transcurrent faults, volcanic belts, and caldera volcanoes, are clarified using various geophysical data such as gravity anomalies, seismic velocity structures, strain rates, and epicenter distributions. The results show that strain concentration zones and inland earthquake epicenters have close spatial relationships with geological structures such as rift boundary faults, large transcurrent faults, caldera structures, and volcanic belts. It can be interpreted that fluids migrating upwards from lower crustal weak zones below rifts, volcanic belts, or calderas, effectively weakened the crust due to its high pore fluid pressure, and caused earthquake ruptures under horizontal compression.
The 2016 Kumamoto earthquake occurred in the tectonically complex central Kyushu area where several forcing factors such as the subducting Philippine Sea plate, the Median Tectonic Line and the Nankai forearc sliver, the spreading Okinawa trough, and the migrating volcanic front are involved. Neogene–Quaternary tectonics of central Kyushu are revisited by integrating geological, seismological, and geodetical approaches. Deformation histories of the Futagawa and Hinagu fault zones, the source faults of the Kumamoto earthquake, are also established in an attempt to explain the relationship between geologic structures and rupture processes of the earthquake. The results show that present-day tectonics surrounding central Kyushu are considered to have originated in the last 1 Ma or younger, as a transtensional tectonic zone (Central Kyushu Shear Zone) characterized by combined dextral faults and rift zones (or volcanoes). Reflecting spatiotemporal variations of the crustal stress field and rift activity, the Futagawa and Hinagu fault zones show multi-stage deformation throughout the Neogene–Quaternary periods: normal faulting to dextral faulting for the Futagawa fault zone and sinistral to dextral faulting for the Hinagu fault zone. Those diverse histories of stress and strain fields in central Kyushu possibly led to the complexities of fault geometry and rupture process of the Kumamoto earthquake.