金曜セミナー(11月4日)Natalia Poiata氏

Imaging seismic sources through an automated multi-scale detection and location scheme: application to the analysis of regular and slow earthquake seismicity

Natalia Poiata (National Institute for Earth Physics, Romania)

The ability to outline in fine detail the pattern of seismic energy release at different spatial and temporal scales is a key for understanding the physics of deformation processes. The rapid evolution and densification of seismological monitoring systems, and continuously growing computational resources provide us with the opportunity to explore seismic sources at scales that were previously unknown. This, however, requires advanced, automated methods that can fully exploit complexity of seismic signals recorded across dense, large-scale seismic networks.

In this talk I will present a recently developed method called BackTrackBB, which allows to image the coherent energy radiation from seismic sources associated to different tectonic environments using the continuous recordings from large-scale seismic networks and arrays (Poiata et al. 2016). The method represents a computationally efficient, array-based scheme addressing the issue of simultaneous detection and location of seismic sources by exploiting the multi-scale frequency-selective coherence of signals’ statistical features recorded across the stations’ of seismic network. The advanced signal-processing scheme implemented in BackTrackBB is designed to account for a priori unknown nature and predominant frequency content of the signals, potentially associated to a variety of sources (e.g., earthquakes, tremors).

I will discuss the performance of the method in detecting, locating and characterizing the seismic sources in complex tectonic environments using the following examples: (1) analysis of the preparatory phase of the 2014 Iquique earthquake in northern Chile; (2) detection and location of tectonic tremor sources and low-frequency earthquakes during a selected tremor episode in southwestern Japan. The potential of extending the method to an automated system for seismic activity monitoring will be also discussed. This work is a joint collaboration with IPGP (France) and ERI.