Competing thermal pressurization and dilatancy hardening realizes coexistence of fast and slow slip on the shallow plate boundary fault

Abstract:

Understanding seismogenesis on the shallow plate boundary fault is important for seismic and tsunami risk assessment. Recent geophysical observations document repeating slow earthquakes on the shallow plate boundary fault. It also hosts coseismic rupture during megathrust earthquakes, which cause devastating tsunami. Investigations on fault rock samples acquired during scientific drilling project at the Nankai trough suggest that both fast and slow slip occurred on the same shear zone along the shallow décollement. This study investigates the physical mechanism, which realizes the coexistence of fast and slow slip on the same fault, using numerical simulations.

We use the model of Suzuki and Yamashita (2010, 2014), in which interactions between fault slip and fluid are considered. There are two competing mechanisms; thermal pressurization (TP) and dilatancy hardening (DH). TP enhances fault slip through increasing pore fluid pressure by shear heating, whereas DH suppresses fault slip through decreasing pore fluid pressure by dilation induced by slip. We conducted numerical simulations under the physical condition of shallow décollement, which was estimated using X-ray Computed Tomography (XCT) data for the fault rock samples, to understand how these two mechanisms work there.

In theoretical considerations, in which heat and fluid diffusion are neglected, the model shows that TP does not work for local seismic events on the shallow décollement. However, our model also shows that if the fault is forced to slip by a certain amount by external energy input, such as propagating coseismic rupture from deeper plate interface during megathrust earthquakes, TP is, in turn, invoked and coseismic rupture can occur. Numerical simulations with heat and fluid diffusion confirm our theoretical considerations. Calculated temperature and porosity increase during megathrust earthquakes are on the same order as those estimated from geological analysis to the rock samples acquired at the Nankai trough. Our model implies that all the shallow plate boundary faults in subduction zones can host tsunamigenic slip. Rather, it is important for tsunami risk assessment to evaluate whether strain energy accumulated on the deeper plate interface is large enough to invoke TP on the shallower plate interface.