An extensive study of clustering features of seismicity in Italy from 2005 to 2016

Abstract

To study the characteristics of seismicity in Italy, we have made use of the ISIDE (Italian Seismic Instrumental and parametric Data-basE) catalogue since 2005 April 16, which was compiled by the Istituto Nazionale di Geofisica e Vulcanologia (INGV). This catalogue includes high quality records of the occurrence times, locations, magnitude and other information about the earthquakes that occurred in and near Italy. We made use of the original form and two extended versions of the space–time ETAS model, namely, the 2-D ETAS model, the hypocentral 3-D ETAS model and the finite-source (FS) ETAS model. Our results show that the rupture geometries of large earthquakes, including the L’Aquila (2009 April 6 Mw6.1/ML5.9 ), the Finale-Emilia (2012 May 20, Mw5.8/ML5.9), the Amatrice (2016 August 24 Mw6.0/ML6.0) and the Norcia (2016 October 30 Mw6.5/ML6.1) earthquakes, control the spatial locations of their direct aftershocks. These direct aftershocks are mainly distributed in some areas adjacent to, but seldom at, the parts with the biggest slips along the main shock rupture, implying that aftershocks compensate the rupture of the main shock. The background seismicity rate is not stationary in all these areas, but shows several phases tuned by the major events. Regarding the difference among the three versions of the ETAS model, we found: (i) hypocentral depth plays an important role in triggering; (ii) when classifying background and triggered seismicity, all three models give similar results, but when classifying the family trees in the catalogue, the geometry of an earthquake rupture should be considered. The FS ETAS model classifies most aftershock events as aftershocks of the main shock; (iii) adopting point sources together with isotropic spatial response causes underestimates of the effects triggered by the main shocks. Such biases can be corrected by incorporating the rupture geometries of major events into the model formulation. Compared to the original point-source model, more direct aftershocks from the main shock are estimated by using the FS ETAS model and (iv) The rupture geometry of a major earthquake can be inverted to some extent from small aftershocks following it by fitting to the finite-source ETAS model.