Time dependent seismicity in the continental fracture system

Two time-dependent seismicity models are tested by using recent reliable data of earthquakes generated in active regions of ten large areas (West Mediterranean, Aegean, Cyprus, Anatolia, Central Asia, Sumatra-Java, Japan, North Pacific, California, South America) of the continental fracture system. The first one, called TIMAPR (Time and Magnitude Predictable Regional) model is based on interevent times of strong mainshocks (M= 6.3-9.0) generated in circular seismogenic regions (networks of faults). The second, called D-AS (Decelerating-Accelerating Seismicity) model, is based on triggering of a mainshock by its preshocks. Tests of decelerating-accelerating precursory seismicity against synthetic catalogues with spatio-temporal clustering verify the validity of the D-AS model. Backward tests of both models showed that: a) every strong shallow mainshock is preceded by a decelerating and an accelerating preshock sequence within well-defined time, space and magnitude windows, allowing its intermediate-term prediction by the D-AS model, and b) in each circular seismogenic region the mainshocks show quasi-periodic behavior with interevent times following the TIMAPR model which is also applied to predict the mainshock. Forward tests of both models indicate candidate regions for the generation of strong mainshocks during the next decade (2013-2022 or so). Estimated (predicted) values of their basic focal parameters (time, magnitude, epicenter) and their uncertainties are given to objectively define the predicting ability of the joint application of the two models.