New Methods for Analyzing the Nature of Nonstationarity in the Behavior of Seismicity

Three methods of data series nonstationarity analysis, new in geophysics, are used to analyze earthquake catalogs of the Kuril-Kamchatka and Mid-Atlantic Ridge regions. The possibility of identifying the nonstationary component of seismicity and the nature of nonstationarity are discussed. The application of new methods confirmed a number of known (expected) regularities and revealed several nontrivial features. Among these are (1) a trend of the nonstationarity to increase with characteristic time, which may indicate that the seismicity spectrum corresponds to flicker noise; (2) a difference in the pattern of magnitude distribution, probably corresponding to a decrease in the b-values, for the clustering main events; (3) the detection of two trends in the behavior of seismicity with time: clustering at smaller relative distances and repulsion at larger distances. These trends may correspond to the epochs of seismicity growth and subsequent decay during accumulation of tectonic stresses. The results suggest the promising application of these analysis methods, which are new in seismology, to provide a more refined picture of the nature of the nonstationarity of the seismic process.