Spatial correlation of aftershock locations and on-fault main shock properties

Abstract

We quantify the correlation between spatial patterns of aftershock hypocenter locations and the distribution of coseismic slip and stress drop on a main shock fault plane using two nonstandard statistical tests. Test T1 evaluates if aftershock hypocenters are located in low-slip regions (hypothesis H1), test T2 evaluates if aftershock hypocenters occur in regions of increased shear stress (hypothesis H2O) In the tests, we seek to reject the null hypotheses H0: Aftershock hypocenters are not correlated with (1) low-slip regions or (2) regions of increased shear stress, respectively. We tested the hypotheses on four strike-slip events for which multiple earthquake catalogs and multiple finite fault source models of varying accuracy exist. Because we want to retain earthquake clustering as the fundamental feature of aftershock seismicity, we generate slip distributions using a random spatial field model and derive the stress drop distributions instead of generating seismicity catalogs. We account for uncertainties in the aftershock locations by simulating them within their location error bounds. Our findings imply that aftershocks are preferentially located in regions of low-slip (u ≤ 1/3umax) and of increased shear stress (Δσ < 0). In particular, the correlation is more significant for relocated than for general network aftershock catalogs. However, the results show that stress drop patterns provide less information content on aftershock locations. This implies that static shear stress change of the main shock may not be the governing process for aftershock genesis. Copyright 2006 by the American Geophysical Union.