Mahalanobis distance-based recognition of changes in the dynamics of a seismic process

Abstract

In the present work, we aim to analyse the regularity of a seismic process based on its spatial, temporal, and energetic characteristics. Increments of cumulative times, increments of cumulative distances, and increments of cumulative seismic energies are calculated from an earthquake catalogue for southern California from 1975 to 2017. As the method of analysis, we use the multivariate Mahalanobis distance calculation, combined with a surrogate data testing procedure that is often used for the testing of non-linear structures in complex data sets. Before analysing the dynamical features of the seismic process, we tested the used approach for two different 3-D models in which the dynamical features were changed from more regular to more randomised conditions by adding a certain degree of noise. An analysis of the variability in the extent of regularity of the seismic process was carried out for different completeness magnitude thresholds. The results of our analysis show that in about a third of all the 50-data windows the original seismic process was indistinguishable from a random process based on its features of temporal, spatial, and energetic variability. It was shown that prior to the occurrence of strong earthquakes, mostly in periods of generation of relatively small earthquakes, the percentage of windows in which the seismic process is indistinguishable from a random process increases (to 60 %–80 %). During periods of aftershock activity, the process of small earthquake generation became regular in all of the windows considered, and thus was markedly different from the randomised catalogues. In some periods within the catalogue, the seismic process appeared to be closer to randomness, while in other cases it became closer to a regular behaviour. More specifically, in periods of relatively decreased earthquake generation activity (with low energy release), the seismic process appears to be random, while during periods of occurrence of strong events, followed by series of aftershocks, significant deviation from randomness is shown, i.e. the extent of regularity markedly increases. The period for which such deviation from random behaviour lasts depends on the amount of seismic energy released by the strong earthquake.