The effects of friction on earthquake triggering and fault zone evolution

Open Access
Savage, Heather M.
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
May 25, 2007
Committee Members:
  • Chris Marone, Committee Chair
  • Sridhar Anandakrisnan, Committee Member
  • Terry Engelder, Committee Member
  • Michele Cooke, Committee Member
  • Peter E Schiffer, Committee Member
  • friction experiments
  • earthquake triggering
  • transient stress
  • fault damage zone
  • fracture
Variations in frictional fault strength caused by oscillating stresses and slip are analyzed. Laboratory experiments of oscillating shear stress, equivalent to oscillating stresses that trigger seismicity on natural faults, highlight changes to stick-slip instabilities such as clock advance of failure, strength of the fault at failure and pre and coseismic slip. I test both continuous and transient oscillations. Correlation between oscillations and stick-slip events are determined by the timing of dynamic failure with respect to the oscillation phase or by the change in recurrence interval from background rates. The correlation of failure with the shear stress oscillation depends on oscillation frequency and amplitude at low frequencies and predominantly on oscillation amplitude at high frequencies. The frequency boundary between these two regimes is analogous to the inverse of the time needed to displace the frictional critical slip distance. The role of fault zone architecture in determining a fault’s susceptibility to triggering is also investigated by varying gouge layer thickness (2 - 6 mm) and studying bare granite surfaces. Varying fault zone architecture shows that stick-slip triggering depends on both properties of the transient oscillation, as already noted, and fault zone properties. In these experiments, the load point displacement must be greater than the critical slip distance (Dc) in order for triggering to occur. Faults with large Dc creep throughout the interseismic period and the triggering threshold is therefore a function of oscillation amplitude and frequency as well as fault state. Faults with small Dc are locked throughout most of the interseismic period, so that the triggering threshold is not a function of fault state. Frequency can inhibit failure when Dc is achieved as velocity is increasing and encourages failure if Dc is achieved during velocity reduction. Increasing velocity temporarily strengthens the fault whereas a velocity reduction further weakens and promotes failure. In addition to laboratory experiments, numerical simulation of strong and weak faults, as well as slip-weakening faults, are compared to determine differences in damage zone patterns. Tensile fractures form along the fault where slip gradients between elements are high. The presence of off-fault fracturing changes fault zone properties, which in turn influence slip patterns and work partitioning in the fault system. The weak and slip-weakening fault systems demonstrate that after seismic radiation and fracture propagation energy are accounted for, excess energy is available for further damage around the fault. This damage is most likely asymmetric and concentrates in areas that have already fractured.