Nonlinear analysis of magnitude and interevent time interval sequences for earthquakes of the Caucasian region

Abstract. It is well known that lithospheric seismic processes are characterized by self-similarity or scale invariance in terms of earthquake-size, time, space and space-time distributions, although precise details of underlying dynamics are not clear. In this study we apply nonlinear dynamics theory tools, such as a correlation dimension, "surrogate" data analysis and positive Lyapunov exponent calculation, to investigate dynamical characteristics of seismicity in the Caucasian region. Interevent time intervals and magnitude sequences are considered for different area and magnitude windows. We find significant evidence of a low dimensional nonlinear structure of earthquake time distribution, obtained by consideration of time interval sequences between all events encountered, above some threshold magnitude, in the original catalogue. However nonlinear structure is absent in artificially generated sequences of time intervals between independent events as well as time intervals between aftershocks. It seems that this kind of filtration of the original catalogue destroys the existing temporal structure of considered lithospheric processes. Unlike artificial inter-aftershock time interval sequences, obtained by removing independent events from the original series, the time interval sequence between the Racha earthquake aftershocks reveals clear evidence of nonlinear structure. Earthquake magnitude dynamics. for all considered regions and magnitude windows, reveal high dimensional nonlinearity.