Consider a statistical model of the seismic process in the subduction zone of the KurilKamchatka island arc according to the catalog of seismic moment tensor [1] over the period 1976–2005, using spatial, temporal and energy criteria of relatedness of seismic events with the test orientation. The question of the presence of non-local effects in time (memory) and space in circuits related seismic events.
In this paper we consider one of the stages predprognoznogo time series analysis, based on the decomposition of the phase portrait on the example quasicycle time-series radon.
In this work the interpolation and regression models of the macroseismic field classical type dependence are offered for the Kurile-Kamchatka region earthquakes. The given models essentially expand the region of the dependence now in use applicability, have the relatively high accuracy and are to a considerable degree free from constant biases in the magnitudes and the hypocentral distances wide range.