![volcano box training volcano box training](https://i.ebayimg.com/images/g/TTwAAOSwOHpg1vGw/s-l600.jpg)
2.1 Seismic Dataįor this study, we leverage 6 years of continuous seismic data recorded across the Observatoire Volcanologique du Piton de la Fournaise (OVPF) network shown in Figure 1.
![volcano box training volcano box training](https://i1.wp.com/ipamusic.co.uk/wp-content/uploads/2019/09/dirt-box-disco-6.jpg)
Eruptions are fed by a magma plumbing system consisting of a succession of reservoirs spreading from depth greater than 30 km below the external western flank of the volcano (Michon et al., 2015) to 2 km below the summit craters (Di Muro et al., 2014 Peltier et al., 2009), where most eruptions initiate. The locations of the fissures generated by these eruptions are shown in Figure 1. Between 2014 and March 2019, 14 eruptions occurred on the flank or at the base of the terminal cone. Following 41 months of rest between the end of 2010 and June 2014, eruptive activity at Piton de la Fournaise renewed on 20 June 2014. Recent activity has been focused in the Enclos Fouqué caldera, formed about 5,000–3,000 years ago (Ort et al., 2016), inside which a terminal cone was formed as a consequence of frequent effusive eruptions characterized by lava fountains and lava flow emissions. It is one of the most active volcanoes in the world, exhibiting 71 eruptions between 19 (Duputel et al., 2019 Roult et al., 2012). Piton de la Fournaise is an active volcano situated on La Réunion, a hot spot basaltic island in the western part of the Indian Ocean located approximately 800 km east of Madagascar. In this work, we describe how statistical features derived from the continuous seismic signal recorded at Piton de la Fournaise volcano can be utilized to build ML models that reveal the characteristic eruptive tremor and eruptive dynamics of volcanic eruptions. With regard to the applications of ML to study and characterization of volcanoes, the primary applications thus far have been in the classification of volcano-seismic signals (Hibert et al., 2017 Malfante et al., 2018 Titos et al., 2019).
![volcano box training volcano box training](https://xdarom.com/wp-content/uploads/2016/11/Volcano-Box.jpg)
The application of machine learning (ML) techniques to the analysis of geophysical signals has become widespread in diverse settings such as the analysis of laboratory experiments (Hulbert et al., 2018 Rouet-Leduc et al., 2017 Rouet-Leduc, Hulbert, Bolton, et al., 2018), tracking slow-slip in real Earth (Rouet-Leduc, Hulbert, & Johnson, 2018), and phase association for the development of earthquake catalogs to name a few (McBrearty et al., 2019 Ross et al., 2019). Several studies have also shown that analyzing the spectral content and location of volcanic tremor can help track the evolution of magma migration (Di Lieto et al., 2007 Jellinek & Bercovici, 2011 Kurokawa et al., 2016) ( 2005) demonstrated that the cumulative amplitude of tremor recorded throughout an eruption could be used to estimate the volume of lava erupted. Battaglia and Aki ( 2003) demonstrated that the eruptive tremor sources at Piton de la Fournaise volcano were good indicators of the locations of eruptive fissures in a subsequent study Battaglia et al. Despite the inherent complexity of volcanic systems, there have been many demonstrations of the utility of volcanic tremor in monitoring eruptions. Attempts to forecast and monitor volcanic eruptions must take into account that the style and magnitudes of eruptions vary drastically, even throughout the duration of a single eruption (Battaglia, Aki, & Staudacher, 2005 Chardot et al., 2015 Kurokawa et al., 2016). Chouet, 1996) as well as degassing from eruptive vents (Battaglia et al., 2005) and are thus of interest in the context of both monitoring and forecasting volcanic activity. These seismic signals are thought to be critical in characterizing the magma migration pathways in the internal plumbing system of volcanoes, as they are typically linked with magma propagation (B. Long-lasting seismic signals known as volcanic tremors are almost ubiquitously present in eruptive episodes at volcanoes (Jellinek & Bercovici, 2011).