Stress history controls the spatial pattern of aftershocks: case studies from strike-slip earthquakes

Abstract

Earthquake ruptures perturb stress within the surrounding crustal volume and as it is widely accepted now these stress perturbations strongly correlates with the following seismicity. Here we have documented five cases of the mainshock-aftershock sequences generated by the strike-slip faults from different tectonic environments of world in order to demonstrate that the stress changes resulting from large preceding earthquakes decades before effect spatial distribution of the aftershocks of the current mainshocks. The studied mainshock-aftershock sequences are the 15 October 1979 Imperial Valley earthquake (Mw = 6.4) in southern California, the 27 November 1979 Khuli-Boniabad (Mw = 7.1), the 10 May 1997 Qa'enat (Mw = 7.2) and the 31 March 2006 Silakhor (Mw = 6.1) earthquakes in Iran and the 13 March 1992 Erzincan earthquake (Mw = 6.7) in Turkey. In the literature, we have been able to find only these mainshocks that are mainly characterized by dense and strong aftershock activities along and beyond the one end of their ruptures while rare aftershock occurrences with relatively lower magnitude reported for the other end of their ruptures. It is shown that the stress changes resulted from earlier mainshock(s) that are close in both time and space might be the reason behind the observed aftershock patterns. The largest aftershocks of the mainshocks studied tend to occur inside the stress-increased lobes that were also stressed by the background earthquakes and not to occur inside the stress-increased lobes that fall into the stress shadow of the background earthquakes. We suggest that the stress shadows of the previous mainshocks may persist in the crust for decades to suppress aftershock distribution of the current mainshocks. Considering active researches about use of the Coulomb stress change maps as a practical tool to forecast spatial distribution of the upcoming aftershocks for earthquake risk mitigation purposes in near-real time, it is further suggested that the background mainshocks along the neighbouring faults should be taken into account in producing the stress change maps for commenting on aftershock occurrences.