Modeling Silent Slip Events along the Unstable-stable Transition in Subduction Zones

Recent high-resolution GPS observations of crustal movements revealed that silent slip events occurred in and below the deeper part of the seismogenic zone. For example, Ozawa et al. (2001) detected a silent thrust slip event in the deeper part of the Tokai subduction zone, which is a well-known seismic gap along the Suruga-Nankai trough. This event started from the beginning of 2001 and is migrating at the speed of about 20 km/year. Slip velocity of silent events is roughly estimated to be 10^-8-10^-9 m/s that is little larger than the velocity of relative plate motion. To investigate the mechanism of silent slip events, we simulate the earthquake generation processes of a thrust fault in 3-D elastic half-space. First, we use the ordinal Dieterich/Ruina rate- and state-dependent friction law with cut-off velocity to the rate-dependence. We give depth distribution of constitutive law parameters based on experimental studies. Below the depth of the unstable-stable transition, critical weakening displacement begins to increase due to an increase of plastic deformation between micro-asperities. Around the transition zone, slow slip events are expected to occur due to large critical weakening displacement and small stress drop. The results of numerical simulations show that stationary slips proceed at the deeper part of the fault region, and then, they proceed gradually upward due to the stress concentration along the unstable-stable transition zone. In some cases, eventual silent slips occur at the deeper part of the seismogenic fault 10-50 years before instability. The slip velocity of this event is 10^-9 m/s that is close to the observed value of slip velocity. A few years before instability, slip acceleration occurs along the transition zone. In some cases, silent slip events migrate horizontally along unstable-stable transition zone at the speed of about 10 km/year. Finally slip is localized into a narrow area at the deeper part of the seismogenic fault. Silent slip events can be interpreted as events caused by the stress concentration along the unstable-stable transition zone and the transitional behavior of the friction law. Around the unstable-stable transition zone, actual frictional behavior is thought to be very complex. Shimamoto (1986) performed experiments using Halite to investigate frictional behavior around the unstable-stable transition. The results show the complex behavior where strength depends on strain rate: at a very low strain rate, strain rate strengthening occurs, at an intermediate strain rate, strain rate weakening occurs, and, at a high strain rate, strain rate strengthening occurs. These results suggest that at a very low strain rate, flow law is at work, at an intermediate strain rate, frictional instability can occur, and, at a high strain rate, strong viscous dissipation is at work. Using a spring and slider block system, we can reproduce slow slip events with this kind of friction law.