Aftershock triggering by complete Coulomb stress changes

We examine the correlation between seismicity rate change following the 1992, M7.3, Landers, California, earthquake and characteristics of the complete Coulomb failure stress (CFS) changes (ΔCFS(t)) that this earthquake generated. At close distances the time-varying “dynamic” portion of the stress change depends on how the rupture develops temporally and spatially and arises from radiated seismic waves and from permanent coseismic fault displacement. The permanent “static” portion (ΔCFS) depends only on the final coseismic displacement. ΔCFS diminishes much more rapidly with distance than the transient, dynamic stress changes. A common interpretation of the strong correlation between ΔCFS and aftershocks is that load changes can advance or delay failure. Stress changes may also promote failure by physically altering properties of the fault or its environs. Because it is transient, ΔCFS(t) can alter the failure rate only by the latter means. We calculate both ΔCFS and the maximum positive value of ΔCFS(t) (peak ΔCFS(t)) using a reflectivity program. Input parameters are constrained by modeling Landers displacement seismograms. We quantify the correlation between maps of seismicity rate changes and maps of modeled ΔCFS and peak ΔCFS(t) and find agreement for both models. However, rupture directivity, which does not affect ΔCFS, creates larger peak ΔCFS(t) values northwest of the main shock. This asymmetry is also observed in seismicity rate changes but not in ΔCFS. This result implies that dynamic stress changes are as effective as static stress changes in triggering aftershocks and may trigger earthquakes long after the waves have passed.