Evolution of stress in Southern California for the past 200 years from coseismic, postseismic and interseismic stress changes

Summary The seismicity of southern California results from stresses that arise from the relative motion of the Pacific and North American Plates being accommodated along the San Andreas Fault (SAF) system and the Eastern California Shear Zone (ECSZ). Here we calculate how the stress field in southern California has evolved over the past two centuries due to interseismic loading, as inferred from current GPS observations of surface velocities, from redistributions of static stress induced by large (Mw ≥ 6.5) earthquakes since the 1812 Wrightwood quake, and postseismic viscoelastic relaxation associated with these events that serves to transfer coseismic stresses from the deep, warm, lower crust and upper mantle to the overlying seismogenic upper crust. We calculate Coulomb stress changes on vertical strike-slip faults striking parallel to the SAF and at the hypocenters on the rupture planes of all Mw ≥ 6 events over the past two centuries. Our results suggest that the 1857 Mw = 8.2 Fort Tejon earthquake, by far the largest event to have occurred in the region over the past two centuries, had a profound influence on the state of stress in Southern California during the 19th century, inducing significant stress increases to the north (Parkfield region and adjoining creeping SAF) and south (southern SAF and San Jacinto fault), and stress relief across the southern ECSZ. These stress changes were then greatly magnified by postseismic relaxation through the early part of the 20th century. Slow interseismic build-up of stress further loads all major strike-slip faults and works to reload the areas of the ECSZ where stress was relieved by the 1857 quake. Our calculations suggest that only 56% of hypocenters were pushed closer to failure by preceding coseismic stress changes, suggesting that the occurrence of large earthquakes is not strongly determined by coseismic Coulomb stress changes. This percentage rises to 70% when postseismic stress changes are also considered. Our calculations demonstrate the importance of postseismic viscoelastic relaxation in the redistribution of stress following large earthquakes. We find, however, that postseismic processes associated with events more than about a decade old are near completion and thus do not significantly influence the regional velocity field presently observed in southern California.