The Initiation of Dynamic Rupture on a 3-m Laboratory Earthquake Experiment

Date: 4/26/2019

Time: 01:30 PM

Room: Elliott Bay

Repeating earthquakes and low-frequency tremor-like signals have been observed prior to the initiation of some earthquakes suggesting an extended nucleation process. I describe sequences of laboratory earthquakes generated on a 3-meter laboratory rock experiment that provide insights into how earthquakes can initiate and have implications for variability in repeating earthquake sequences. The granite sample was loaded to 2-12 MPa stress levels and then sheared, while slip on the laboratory fault was measured at 16 locations along its 3-m length. If the initial along-fault stress distribution was sufficiently heterogeneous, stick-slip events that ruptured the entire fault were preceded by a sequence of smaller, M -2.5 earthquakes that repeatedly ruptured the same subsection of the fault, with successive events increasing in rupture length and stress drop. More uniform stress distributions produced stick-slip events that initiated more abruptly, often with a meter-sized zone of slow slip (nucleation zone) that expanded and accelerated until reaching seismic slip speeds (>0.1 m/s). The effects of fault strength heterogeneity and fault healing can cause variations in the size and speed of the nucleation process. When continuously loaded at a steady rate, the sample produced sequences of complete-rupture stick-slip events which might be analogous to repeating earthquake sequences observed in nature. The laboratory experiments show that the largest events in these sequences are the result of a smaller and more abrupt nucleation process, whereas weaker events nucleated more slowly and over a larger fault area. Finally I discuss different models for scaling up the laboratory results to larger, rougher natural faults, and describe which aspects are supported by laboratory and field observations.

Presenting Author: Gregory McLaskey

Authors