LONG TEMPORAL VARIATION OF SEISMIC PARAMETERS FOR SEISMIC PATTERNS IDENTIFICATION IN GREECE

Résumé 0

A new approach of detailed spatio-temporal variation analysis of seismic data is proposed by means of FastBEE (Fast estimation of Big Expected Earthquake) aiming at the regional monitoring of seismic activity for premonitory seismic patterns identification. For the investigation of temporal variation, a set of seismic parameters is used, like the logarithm of the number of earthquakes logN, estimates of 6-value obtained by the maximum likelihood estimation model, time clustering of seismic activity AR(t) and of energy released EM, since they can be considered as precursory seismological indicators. Earthquake catalog data, used in this approach, were elaborated in order to construct the time series for each parameter within a time window, large enough, as to guarantee statistical meaningful result. The Hellenic trench-arc region under investigation is chosen in the basis of its seismotectonic characteristics, in relation to the spatial extent of the seismogenic zone. The tools were tested, for long temporal variation features in the Ionian Islands Sea and the North Aegean Sea regions and its successful applicability is presented. The rise of irregularity, along these temporal profiles, was formulated in specific quantitative premonitory seismic pattern. In most of the cases, FastBEE premonitory pattern found shows significant changes from the background values of each parameter. Parameter logN shows a valley form curve, which start to increase before the expected earthquake occurrence, as well as the energy parameter E273, while b-value temporal estimates are forming a mountain shape curve, before the occurrence of a big earthquake. Instead, parameter ÙR(t) present a rapid fluctuation, without any kind of premonitory character