Insights into Mt. Etna volcano dynamics by seismic and infrasonic signals

Abstract

Active volcanoes are one of the most severe natural hazards in the world. Volcanoes are geologic manifestations of highly dynamic and complexly coupled physical and chemical processes in the interior of the Earth. They are complex dynamical systems that produce distinctive patterns. Volcanic eruptions are the culmination of a complex ensemble of processes that occur on a broad range of time scale, from tens or hundreds of years (e.g. magma rise and differentiation) to fractions of seconds (e.g. fragmentation). About 550 volcanoes have erupted in historical times. Reconstruction of the eruptive history of many volcanoes has shown that inactive periods of thousands of year are not uncommon. Historical data indicate that eruptions are almost always preceded and accompanied by volcanic unrest manifested by physical and/or geochemical changes in the state of the volcano. Detection of precursory phenomena (e.g., seismic, geodetic, gravity signals, gas emission) is the main aim of volcano monitoring which provides parameters for early warning systems. Systematic collection and analysis of huge amount of data recorded on active volcanoes are performed for both research and monitoring purposes. The fact that strongly different precursory patterns can be observed for different eruptions at the same volcano means that there exists no universal sets of empirical parameters relating precursor to eruptions. However, data from volcano monitoring constitute the only scientific basis for short-term forecast of imminent eruption or changes in the volcano behavior. Most active volcanoes are routinely monitored observing the pattern of the seismic activity and ground deformations. During the last years a key role in volcano monitoring is played by time series analysis methods and pattern recognition techniques, both in time and time-frequency domain, in order to detect and analyze different eruptive patterns. The aim of this thesis is the study of seismic and infrasonic signals generated by volcanoes using signal processing techniques and novel approaches based on nonlinear time series analysis and pattern recognition techniques.