Search for anomalies in Stromboli's pre-paroxysm activity through an automatic hybrid method of time series analysis

Fabrizio Ambrosino,Carlo Sabbarese,Giovanni Macedonio,Walter De Cesare,Antonietta M. Esposito,Federico Di Traglia,Nicola Casagli,Teresa Nolesini,Salvatore Inguaggiato,Fabio Vita,Sonia Calvari,Giuseppe Salerno,Giuseppe di Grazia,Alessandro Bonaccorso,Carmen López Moreno,Flora Giudicepietro
DOI: https://doi.org/10.1016/j.jvolgeores.2024.108131
IF: 2.986
2024-06-19
Journal of Volcanology and Geothermal Research
Abstract:Stromboli (Italy) is an open-vent volcano with persistent explosive activity producing up to five hundred mild explosions per day. Fluctuations in explosion intensity, varying even by orders of magnitude in terms of emitted volume and their subsequent impact on the surrounding regions, sometimes occur abruptly. Consequently, identifying precursors of larger eruptive activities, particularly for more intense (paroxysmal) explosions, is challenging. In order to search for anomalies in the pre-paroxysm activity related to the summer 2019 eruption, we applied a hybrid method to the automatic analysis of geophysical and geochemical time series. This approach is based on the combination of two methods: 1. the Empirical Mode Decomposition (EMD) and 2. the Support Vector Regression (SVR). The aggregation of these two methods allowed us to identify anomalies in the patterns of the geophysical and geochemical parameters measured on Stromboli in a ten-month period including the July–August 2019 eruption. The results of this study are encouraging for an improvement of the monitoring systems and for volcano early warning applications.
geosciences, multidisciplinary
What problem does this paper attempt to address?