[S-SS10] Statistical seismology and underlying physical processes

[SSS10-P14] Shear strain energy changes caused by large crustal earthquakes in Japan and their influences on the locations of aftershocks

*Sachiko Tanaka¹ (1.National Research Institute for Earth Science and Disaster Resilience)

Keywords:shear strain energy, aftershocks

I investigated shear strain energy changes caused by large crustal earthquakes in Japan and their influences on the locations of aftershocks. The changes of shear strain energy could be connected with those of stress invariant J2 or equivalent stresses, and its increase is considered to correspond to an increase of the averaged amplitudes of shear stresses for randomly-distributed faults (Saito et al., 2018). In the analysis, I used the hypocenters of shallow earthquakes with focal depths of 20 km or shallower and with magnitudes of 2.0 or larger listed in the JMA catalog for the period from 2003 to 2020. I employed an event-clustering scheme based on a single-link cluster analysis (Frohlich and Davis, 1990) and selected the clusters including the earthquakes with magnitudes of 6.0 or larger. I considered the largest event in each cluster as a mainshock and the events following it as aftershocks. I calculated stress changes caused by the mainshocks (Okada, 1992) using the uniform slip fault models of GSI and evaluated shear strain energy changes for the background stress orientations estimated by applying a stress tensor inversion (Michael, 1987) to the NIED Hi-net focal mechanism solutions (Saito et al., 2018). I selected the twelve mainshocks (M6.1-7.3) for the analysis, and calculated shear strain energy changes at the hypocenters of the aftershocks. For the 2016 Kumamoto earthquake (M7.3), I found 77% of the aftershocks occurred at the locations with positive shear strain energy changes. Similar correlations (more than 60% of the aftershocks lie in the areas where the shear strain energy increased by the mainshocks) were observed for six mainshock-aftershock sequences. These correlated sequences were not linked with any specific mainshock magnitudes, fault types, or aftershock b-values in the Gutenberg-Richter law, but tended to have large p-values in the modified Omori formula giving the decay of aftershock activity. Furthermore, these sequences had large aftershock areas compared to the values from the scaling relationship for crustal earthquakes in Japan (Tanaka et al., 2019). This may imply that the locations of aftershocks excluding near-field events can be effectively controlled by an increase of shear strain energy induced by the mainshocks.