RECURRENCE QUANTIFICATIONANALYSIS OF ACOUSTIC EMISSION TIME SERIES IN THE PETEROA VOLCANO AREA (ARGENTINA).

Non linear time series techniques are largely being applied in different disciplines, such as Earth Sciences, Astrophysics, Engineering, Economy, Physiology, and Neurosciences. One, in particular, Recurrence Quantification Analysis (RQA), appears quite promising for the analysis of complex systems [1]. Moreover, it has recently been applied to investigate both acoustic emissions from rocky specimens [2] and the dynamics of complex seismic processes [3]. Whenever two objects are put in contact and let them slide with respect to each other, friction induced vibrations may occur. Examples are represented by active faults within seismogenic zones, narrow-banded noise of train wheels running along tight curves, friction in bearings, and micro-scale events in molecular physics. In this framework, the paper describes the application of RQA to the Passive Acoustic Emission (AE) signal released, at ultrasonic frequencies, by stressed rocks in the Earth’s crust above a given threshold (event). The data record is represented by AE time series gathered, with 30 sec. of sampling rate, at the Peteroa Volcano monitoring station (Argentinean Andes). In these site AE data were collected by piezoelectric transducers, working at two ultrasonic frequencies, stuck to a rock [4, 5, 6]. This way, a huge amount of data is available but the AE signal amplitude varies with to the acoustic impedance, related to local rocks stress conditions and particularly sensitive to fracture density and water content. In order to investigate the evolution characteristics of the quiescence and activation status of the crustal system, the application of the RQA method to the AE time-series is focused to pinpoint peculiar recurrence patterns, without taking into account the amplitude. RQA is a quite simple processing technique that introduces few parameters descriptive of the global complexity of a signal, which are computed from the so-called “Recurrence Plot” [7]. In particular, it is possible to monitor quantitative changes in dynamics of temporal distribution [2], loss of synchronization of dynamic mechanism or spatial irregularities occurring in time. The aim of this work is to identify few descriptors that can explain the main characteristics of the AE signals and identify anomalies to be related to crustal stress modifications or, as in the Peteroa case study, paroxysmal volcanic activities or Earth’s tides imprints [5, 6].