Triggered Aseismic Fault Movements From Nearby Earthquakes, Static or Dynamic Effect?

Observations show that the occurrence of an earthquake can affect either the seismic or aseismic slip behavior of nearby faults. Two stress triggering models are often used in the studies of changes in seismicity after a significant event. One is the static stress triggering mechanism that is associated with the faulting. The other is the dynamic stress triggering model attributed to the ground shaking from the passage of seismic waves. These two mechanisms are also used to explain the phenomenon of triggered fault slip, which is a form of aseismic fault movement (or creep) coinciding closely in time with a large nearby seismic event while being distinct spatially from the primary rupture. We evaluate the possible triggering role of the static stress changes by looking into 14 observed cases of triggered aseismic slip following seven large shocks in California. The changes in static shear stress, normal stress and Coulomb Failure stress (CFS) from each main shock are resolved on those nearby fault segments, which are known to exhibit creep. We examine the signs of static stress changes relative to whether the triggered slip took place on the faults or not. Most of the slipped nearby fault segments were encouraged to move by the imposed changes in static CFS, but there are three discrepancies with this general picture. Two involve the southern part of the San Andreas fault after the 1987 Elmore Ranch and Superstition Hills earthquake sequence; the other involves the southern Calaveras fault segment after the 1989 Loma Prieta event. These three misfits imply that static stress triggering is not the sole mechanism responsible for causing the observed triggered slip. We also examine the possible triggering role of transient loading numerically using a one-dimensional massless spring-slider system. The temporal evolution of the slider (fault) movement is governed by the two-state-variable rate- and state-dependent frictional law proposed by Ruina. Both forms of the ``fault creep'' phenomenon, i.e. ``secular creep'' and ``creep events'' are simulated. We then put transient loads, which are modeled by sine waves, into the system to examine whether the timing of the next expected ``creep event'' is affected. Preliminary modeling results show that certain types of transient loading can cause a large time advance of the next ``creep event'', which starts shortly after the transient load is applied. While our work examines triggered creep events near the surface, it may well have implications for the occurrence of similar events near the bottom of the seimogenic zone of faults where a transition occurs from velocity-weakening to velocity-strengthening behavior.