← Back to Sessions

How Do Stick-slip Events Initiate on a Rough Fault?

Description: 

Laboratory experiments have demonstrated that the slip behavior of a smooth fault is governed by the relationship between the system stiffness k and fault length L relative to the critical stiffness k_c and critical nucleation length L_c. However, as natural faults exhibit roughness at multiple scales, the validity of this stability criterion for rough faults remains to be evaluated. Here, we conducted direct shear experiments on 760-mm-long granite blocks under a normal stress of 10 MPa. The rough interface between the two blocks was created by a natural fracture and is characterized by ~4 mm of topographic variation over a wavelength of ~100 mm. Gouge material was removed to promote instability. When the sample was deformed near the rough fault’s “mated” position, it mostly exhibited stable creep with hundreds of accompanying acoustic emissions ranging from magnitude -7 to -5, but it also produced occasional stick-slip events whereby the entire fault slipped rapidly (~30 mm/s). Our analysis focused on the stick-slip events and their foreshocks. The results show that the slip behavior of our rough fault cannot be explained by the critical nucleation length L_c in the above criterion. On one hand, the entire fault kept creeping without a discernible nucleation process before stick-slip events, and foreshocks were distributed across the fault surface instead of clustering around the epicenters of stick-slip events. This is similar to the behavior of a rigid block with L≪L_c. On the other hand, stick-slip events produced dynamic rupture fronts that were clearly captured by both slip sensors and high-frequency seismograms. The ruptures were observed to initiate near one end of the sample, propagate toward the opposite end, and then propagate backward, resembling the slip behavior observed on smooth faults with L≫L_c. Additionally, the initial 100 μs of a stick-slip event exhibited waveforms that were nearly identical to those of a smaller acoustic emission. Therefore, the behavior of our rough fault aligns more closely with the “cascade up” model described in McLaskey (2019) rather than the “preslip” model.

Session: SSJ-SSOC-SSA Joint Session: From Slow to Fast Earthquakes: Bridging the Spectrum of Fault Slip [Poster]

Type: Poster

Date: 4/16/2026

Presentation Time: 08:00 AM (local time)

Presenting Author: Han Chen

Student Presenter: Yes

Invited Presentation: 

Poster Number: 145

Authors