Despite the fact that it isn't compatible with the elastic rebound theory of earthquake faulting, the Poisson model is widely used to describe the temporal occurrence of earthquakes in probabilistic seismic hazard analysis (PSHA) studies for various reasons. The most important of these is that the Poisson model is simple, yet adequate enough to describe a complex phenomenon. The Poisson model prescribes that earthquakes occur randomly in time with no “memory” of the time since any preceding event (Kramer 1996). Ben-Menahem (1960) and also Gardner and Knopoff (1974) demonstrated that sequences of earthquakes are Poissonian only after foreshocks and aftershocks have been removed from the listing, as these secondary events are causally linked to their mainshock by virtue of them triggering or being triggered by the latter. Hence, the identification and removal of foreshocks and aftershocks from earthquake catalogs, a process referred to as declustering, are a prerequisite for any PSHA study based on a Poissonian model for earthquake occurrence. Two main approaches and their variants are widely used in the seismic hazard assessment community for catalog declustering: 1) the standard space-time windowing technique, such as the Gardner and Knopoff (1974) method, and 2) the cluster-link scheme, such as the Reasenberg (1985) algorithm or Davis and Frohlich (1991) method. The declustering schemes are generally implemented as forward-searching algorithms. Hence, they process an earthquake catalog in increasing (decreasing) chronological sequence for aftershocks (foreshocks) removal. The standard windowing technique classifies as aftershock any event with magnitude smaller than that of the mainshock, occurring within a defined space-time window of the mainshock. In this approach the space-time interaction window is a function of the mainshock magnitude. A detailed description of the methodology can be found in Gardner and Knopoff (1974). The main shortcoming of this method, as originally proposed by these authors, is …