Falanga, Mariarosaria

Collecting massive seismic signals is a high-priority task in seismic risk evaluation, especially in densely populated areas, with cases of strong magnitude earthquake occurrence. At the same time, with the advent of the Internet of Things (IoT) paradigm, distributed and real-time environmental monitoring, supported by device interoperability, enhances the ability to collect data and make decisions especially in critical domains such as the seismic one. A crucial role is played by Semantic Web technologies that, in IoT ecosystems, promote syntactic and semantic interoperability, by enhancing the data quality that becomes ontology-annotated. This article introduces an IoT-oriented framework to collect seismic data, process and store them into a knowledge base. An ontology called Volcano Event Ontology (VEO) modeled for the seismic domain aims at gathering seismic signals collected by sensors for seismic event detection. The ontology is built on the well-known SSN/SOSA ontology, modeled to describe the systems of sensors, actuators, and observations. Seismic data have been collected by monitoring networks at Mt. Vesuvius (Naples, Italy) and Colima volcano (Mexico) and consolidated in the ontology. Moreover, the seismic data are also processed by a classification module to detect different seismic events (Volcano-Tectonic and long-period earthquakes, underwater explosions, and quarry blasts) and then stored in the knowledge base. Prompt detection and classification are, indeed, relevant to track any variation in the volcano dynamics, becoming crucial in cases of explosive crises. Finally, the VEO-driven knowledge base can be queried to get time-based seismic data and detected events, by queries.

The Long-Period (LP) seismicity is common at active volcanoes and is usually modeled as due to pressurized magmatic fluids flowing through rock cavities. These signals are sensitive to the thermodynamic conditions of the magma-gas mixture in the shallow plumbing system and can thus be adopted as “detectors” of an impelling eruption. We found that at Stromboli (Italy) before and/or during recent volcanic crises the LP events can occur in swarms, which show different statistics, higher energy and shallower location than the stationary LP activity. We imputed the LP swarms to a quick depressurization (|ΔP|≥105 Pa) of the shallowest (<0.8 km) part of the conduit. At Shishaldin (Alaska) the 2004 eruption is anticipated by a migration towards the surface of the LP source, which moves from ~8 km to ≾5 km below the crater rim. By simple assumptions, we modeled this source change as produced by an increase of the confining pressure within the plumbing system of ~5x107 Pa, possibly induced by an upward migration of ~108-1010 kg of magma.

A fundamental task in volcano-seismology is the characterization of the source of Volcano- Tectonic (VT) earthquakes; this passes through the discrimination of seismic events from the ambient background noise and the identification of the onset of the main phases, i.e. the body waves. An automatic procedure based on the Independent Component Analysis (ICA) which successfully performs the blind source separation of convolutive mixtures, has been developed to have a prompt discrimination among the different sources in the seismic signals. Specifically, the ICA is adopted to obtain a clear separation among meteo-marine microseism, anthropogenic noise, and volcano-tectonic activity at Campi Flegrei. A coarse- grained variable, i.e. the frequency associated with the maximum amplitude of the power spectral density of the Independent Components (FMPSDA), is introduced. This parameter is sensitive to the variation in the frequency bands of interest (e.g. that corresponding to the corner frequencies of VT events) and can be used as marker of the insurgence of seismic activity. In addition, the ICA also provides the wavefield decomposition of VT earthquakes into basic sources which are naturally polarized into the vertical and horizontal planes, thus allowing the identification and separation of the main seismic phases. On this basis, a novel approach, “ICA-based Polarization” (ICAP), which consists in the estimate of the polarization parameters directly from the Independent Components is introduced. The technique is suitable for directly retrieving polarization features of the seismic signals in a single step, avoiding a priori cumbersome segmentation and filtering procedures. On the basis of the presented results, the application of the ICA to large massive seismic datasets represent an useful tool for fast pre-processing, thus efficiently supporting the volcano monitoring practice

We propose a novel approach to analyze continuous seismic signal and separate the sources from background noise. A specific application to the seismicity recorded at Campi Flegrei Caldera during the 2006 ground uplift is presented. The fundamental objective is to improve the standard procedures of picking the emergent onset arrivals of the seismic signals, often buried in the high-level ambient noise, in order to obtain an appropriate catalogue for monitoring the activity of this densely populated volcanic area. This is particularly useful in order to estimate the release of the seismic energy and to put constraints on the source dynamics. An Independent Component Analysis based approach for the Blind Source Separation of convolutive mixtures is adopted to obtain a clear separation of Long Period events from the ambient noise. The approach presents good performance and it is suitable for real time implementation in seismic monitoring. Its application to the continuous seismic signal recorded at Campi Flegrei has allowed the extraction of high-quality waveforms, considerably improving the detection of low-energy events.

We show an analysis of tiltmetric time series from borehole instruments at Campi Flegrei caldera. We evaluate the crustal response in terms of ground tilting of the entire caldera to external excitations such as long/medium period tidal constituents, by adopting Independent Component Analysis, a nonlinear technique. The main aim to understand an eventual relation between long-period tides and fluid circulating in the hydrothermal shallow system. Indeed, diurnal (solar) and long-period (fortnightly and monthly) components are recognized in the tilting. These tidal constituents cause an oscillatory deformation pattern, superimposed to the normal deformation trend of the area. Moreover, we show that the tilting plane orientations are controlled by the local stress field and the structural features and that the amplitude of the tilt reflects the rheology of the site. These observations indicate the occurrence of structural and thermoelastic site effects. Their knowledge is useful not only for removing the external tidal contribution in the tiltmetric series, but also in delineating the local geology and focusing on the internal sources related to the volcano dynamics. A variation in the fluid circulation may induce a change in the revealed pattern, which can be promptly detected. The promising results we obtain lead us to believe that the same approach can be extended to other tiltmetric networks in volcanic areas, for thorough and detailed analyses of the tiltmetric time series and more accurate studies of the endogenous sources.

In the present work a new approach for the analysis of polarization of seismic signals is proposed. The method is based on Independent Component Analysis and allows the identification and separation of the basic sources, which are naturally polarized into the vertical and horizontal planes. The results from the case study of a swarm of volcano-tectonic earthquakes occurred at Campi Flegrei in October 2015 are impressive: a clear separation of the P- and S-wave seismic phases in the time domain is obtained. In addition, the efficiency of the method in retrieving the polarization parameters is demonstrated by the comparison with other standard techniques. The presented approach provides wavefield decomposition and polarization analysis in a single step, thus avoiding a priori cumbersome filtering procedures and segmentation of the signals. It is useful for discriminating and analysing different seismic phases and can be applied to a variety of volcanic and tectonic signals, therefore it can strongly support all the studies on propagation and source mechanism. Moreover, due to its fastness and robustness this stand-alone tool can be routinely used in the volcano monitoring practice.

We analyze the volcano seismicity recorded during the 2007 eruption of Stromboli. Data-set is composed of the continuous recordings of a three-component broad-band seismometer and of a strainmeter. Starting from the characterization of the standard activity as a stationary phase of equilibrium, we investigate the non-equilibrium phase of the effusive process. A statistical analysis of the explosions reveals that the occurrence is always driven by a Poisson process as for the standard activity, even approaching the effusion phase, with the only difference in shortening the inter-times just during the effusion. A slightly different process can be advocated for the swarms of the explosions, because a maximum in the distribution of inter-times can be evidenced. Regarding the amplitudes of the explosion-quakes, they have a log-normal distribution until the effusion onset as in the standard Strombolian activity. The actual departure from that stationarity seems to be traced by an early deformative response at very long period. It appears as a transient oscillating signal characterized by a period of about three days that modulates the explosion amplitudes. In a conceptual organ pipe-like model it is related to the chocking of the pipe. The successive activity can be interpreted as the response of volcano to restore the equilibrium condition.

We have analysed the wavefield of the background seismic noise at Ischia Island (Italy), recorded between 2017 and 2018. The Independent Component Analysis highlighted two persistent independent signals, with dominant frequency peak around 1 Hz and 3–4 Hz, respectively. The first signal is the most energetic and persistent. Its polarization shows preferential directions properties, with a shallow propagation that seems to be controlled by the tectonic-volcanic structures and the morphology of the Island. Moreover it is well correlated at all the stations in Casamicciola town. The second signal evidences similar polarization properties, but it is strongly affected by cultural noise and does not show any correlation. We interpret the 1-Hz wave-packets as the signature of the shallow hydrothermal system. Following a conceptual model of self-sustained musical instruments, the shallow hydrothermal system of Ischia can be seen as a solid structure constituted by a network of channels, continuously excited by the circulating hydrothermal fluids. In this framework the 1-Hz seismic signal and the eventual higher modes would be produced by persistent self-sustained oscillations.

We analyze the volcano seismicity recorded during the pre‐ and co‐eruptive regimes of the 2007 effusive crisis at Stromboli volcano (Italy). Data‐set is composed of the continuous recordings of a three‐component broad‐band seismometer and of a Sacks‐ Evertson strainmeter. Starting from the characterization of the non effusive phase as a stationary state of equilibrium, we investigate the effusive phase as a non‐equilibrium state. A statistical analysis reveals that the explosion occurrence is always driven by a nearly Poissonian process, as for the standard activity, even during the effusive phase, with the only difference in shortening the inter‐times. Explosion‐quake amplitudes are lognormally distributed until the effusive phase, becoming then broader. This indicates that many scales are involved. A slightly different process can be advocated for the swarms of the explosions occurring during the effusive phase. This suggests that the dynamics of the exsolution and/or aggregation of the gas slugs should differ from the nucleation mechanism responsible of the standard Strombolian activity. The pre‐eruptive regime is characterized by a very long deformative signal that appears as a transient oscillating signal with a period of about three days that modulates the explosion amplitudes. In a conceptual vibrating cavities model, it is related to a chocking phenomenon induced by magma injection, which in turn leads to the effusion.

Long period volcanic tremors (LPTs), typically termed as VLP, have been widely observed in many volcanic systems around the world. LPTs in different volcanic settings are often repetitive, suggesting a nondestructive source and providing critical insights into the fluid dynamic processes operating inside a volcanic system. Recently, a deep triggering source of LPT has also been discovered, located between the LPT source and the magma chamber. These diverse and rich signals not only help monitor the state of shallow volcanic conduit and its link to upcoming eruptions. it also helps understand causal triggering of LPT and potentially intra-crustal transport of magma. Given diverse and rich signals within the Aso volcanic system, we further explore other possible sources in the tidal period. We analysed continuous data recorded at two borehole tiltmeters installed at Aso volcano, during the period of 2011-2016. We first perform a polarisation analysis of the ground tilt, filtering in the frequency bands matching the medium-long period tidal constituents, Mm, Mf and S1. The results evidence a well-defined direction of the tilting plane at both sites, with minor fluctuations occurring in 2012 and 2014. Moreover, we investigated the tilt time series the 1-8 hour band. There is a clear variation in azimuth orientation in at least one of the two tiltmeters and these azimuthal variations occur roughly from September 2013 at least until November 2014. In the meantime, LPTs were strong and more active, ash eruption occurred at the end of August 2014, followed by a major Strombolian eruption episode starting at the end of November 2014. In this time interval between September 2013 and November 2014, we further discovered a sequence of nearly monochromatic signals, with repetitive waveforms, a typical period of about 2 hours and a duration of about 9 hours. We extracted these signals and will compare their timing against the LPT catalog we constructed previously. Moreover, we will perform waveform correlation in order to quantify the degree of similarity among these monochromatic signals. Finally, we will attempt to reconcile all these observations in terms of the volcano dynamics and of its different eruptive styles during the 2011-2016 eruption cycle.