Pattern identification in the dynamic fingerprint of seismically active zones

摘要

The aim of this paper is to present a methodology designed to characterize seismically active areas from the point of view of their dynamic fingerprints, which rely on space-time aspects of seismicity patterns. According to this methodology, earthquake events are represented as points in phase space, which leads to an "events thread" connecting the points in their temporal sequence. A multi-scale analysis procedure is applied to the events thread, and scaling exponents are identified. The proposed framework considers projections of the events thread corresponding to different spatial orientations. For each orientation, the events thread is characterized from the point of view of signal persistence. The resulting persistence diagrams portray anisotropy aspects of the dynamic fingerprint. The identification of anisotropic properties of the dynamic fingerprint provides a new perspective on seismically active zones, complementing the information from geological and geophysical studies. The paper describes the proposed methodological framework, and presents applications focusing on two different regions: the Aleutian Arc and Hawaii.