Characterization of Subduction Seismicity at the Global Scale

In the framework of the EURYI Project 'Convergent margins and seismogenesis: defining the risk of great earthquakes by using statistical data and modelling', a global collection of recent subduction seismicity - partitioned between interplate, intraslab and upper plate events - has been performed. Based on EHB hypocenter and CMT Harvard catalogues, the hypocenters, nodal planes and seismic moments of worldwide subduction-related earthquakes were extracted for the period 1976 - 2007. Data were collected for centroid depths between sea level and 700 km and for magnitude Mw ≥ 5.5. For each subduction zone, a set of trench-normal transects were constructed choosing a 120km width of the cross-section on each side of a vertical plane and a spacing of 1 degree along the trench. For each of the 505 resulting transects, the whole subduction seismogenic zone was mapped as focal mechanisms projected on to a vertical plane after their faulting type classification according to the Aki-Richards convention. Transect by transect, was defined a "subduction box" identifying the spatial limits of the seismicity strictly related to the subduction zone under investigation, then the seismicity considered not related to the subduction process under investigation was removed. For each transect, the events previously identified by Heuret et al (2011) as subduction interface earthquakes were removed from the "subduction box". Upper plate seismicity (i.e. earthquakes generated within the upper plate as a result of the subduction process) was then extracted from the remaining events. We can be reasonably confident that the remaining seismicity can be related to the subducting plate. In this way the subduction seismicity has been properly classified in three categories: interpolate-, intraslab- and upper plate-seismicity. Following Heuret et al. (2011), the 505 transects have been merged into 62 larger segments that were ideally homogeneous in terms of their seismogenic zone characteristics. Several checks were performed, in order both to remove duplicate events, often occurring in these catalogues due to their marginal location relative to the boundary of adjacent transects and to verify if, due to the trench curvature and/or to the huge event clustering typical of some subduction zones, the same event is classified as belonging both to the upper plate and the intraslab category. Comparisons between main seismic parameters (e.g. maximum magnitude, number of events, cumulated seismic moment, recurrence time) with relation to both the different categories selected and the different segments have been performed and the correlation with a wide range of subduction-related parameters taken from the literature (e.g. plates/slab kinematics, thermal parameters) have been evaluated. This allowed to highlight possible cause-effect relationships and to obtain a snapshot on the general behavior of global subduction-related seismicity.