Laboratory Friction Experiments and Modeling Reveal the Mechanism of Shallow Slow Slip Events Observed in the Nankai Trough, Southwest Japan

Slow slip events (SSEs) are the slowest type of discrete slip within the full spectrum of fault-slip behaviors and have been confirmed by both geodetic (e.g., Dragert et al., 2001) and laboratory data (e.g., Ikari, 2019). They have attracted considerable attention due to their mutual interaction with earthquake processes, and multiple approaches have been employed to investigate different aspects of SSEs. Here, we present a study that combines laboratory friction experiments and numerical modeling to explore the mechanisms of SSEs observed through geodetic and borehole data.

We conducted velocity-stepping friction experiments on intact core samples retrieved from the major reverse fault zones of the Nankai Trough, southwest Japan. These experiments were performed under both in-situ effective stress conditions and at 10 MPa, with slip velocities ranging from 1.6 nm/s (plate tectonic driving rates) to 30 μm/s. Our results reveal that fault zone samples transition from velocity-weakening to velocity-strengthening behavior as slip velocities increase, and some rate-and-state friction (RSF) parameters exhibit a dependence on sliding velocity. Numerical models (Zhang and Ikari, 2024) using velocity-dependent RSF parameters, constrained by our experimental data, successfully replicate SSEs comparable to those observed in the Nankai Trough (Araki et al., 2017; Yokota and Ishikawa, 2020) by assuming fault patches at depth ranges and sizes consistent with observational data. In contrast, models based on non-transitional frictional behavior (constant RSF parameters) or near-neutral stability (constant RSF parameters with extremely small velocity weakening) generate slip events that are several orders of magnitude faster than observed SSEs. We therefore propose that the transitional frictional behavior with increasing slip velocity is a key mechanism of shallow SSEs in the Nankai Trough.

Our study demonstrates that laboratory data obtained from centimeter-scale samples can be used to predict the frictional behavior of real faults on the scale of tens of kilometers. By integrating methodologies from multiple disciplines, we can achieve a more comprehensive understanding of the dynamics governing fault slip behavior.