Artificial Intelligence and Machine Learning tools for improving Early Warning systems of volcanic eruptions: the case of Stromboli.

Explosive volcanic blasts can occur suddenly and without any clear precursors. Many volcanoes have erupted in the last years with no evident change in the eruptive parameters and with dramatic consequences for the population living nearby the volcano and the tourists visiting the active areas. In recent years, a big effort has been made to develop Early Warning systems to issue timely alerts to the population. At Stromboli volcano, the development of sensitive instruments to measure the deformation (tilt) of the ground has revealed that the volcano edifice is inflating tens of minutes before the explosion following a recurrent exponential ramp-like pattern. This scale-invariant of ground deformation has allowed the development of a quasi-deterministic Early Warning system which is operative since 2019. In this article we show how Artificial Intelligence and Machine Learning can be successfully applied to improve the efficiency and the sensitivity of Early Warning systems, provided the availability of a comprehensive experimental data set on past explosive events. The approach presented here for the Stromboli case demonstrates promising results also in forecasting the intensity of explosive events, offering valuable insights and new perspectives into the potential risks associated with volcanic activities.

The optical, seismic, and infrasound signature of the March 5 2022, bolide over Central Italy.

On March 5, 2022, a 12 kg meteoroid crossed the sky above Central Italy and was observed by three different observational systems: the PRISMA all-sky camera network (10 stations), the Italian national seismic network (61 stations), and a 4-element infrasound array. The corresponding datasets, each with its own resolution, provided three independent assessments of the trajectory, size and speed of the meteoroid. The bolide traveled across central Italy with an azimuth of 102 degrees, becoming visible at about 91 km above sea level with a velocity of about 15.4 km/s. Its visible trajectory lasted about 15 s. Reasonably, the residual portion of the ablated bolide terminated its path in the Adriatic Sea and could not be recovered. Seismic and infrasound data well match optical observations detecting the bolide Mach cone at 68 km above sea level with a back azimuth of 25 degrees with respect to the array. By comparing results from the three different systems, discrepancies are within the estimated uncertainties, thus confirming the mutual consistency of the adopted methodologies. Therefore, this study shows that different approaches can be integrated to improve the detection capability for bolide crossing the sky in monitored regions.

Atmospheric waves and global seismoacoustic observations of the January 2022 Hunga eruption, Tonga.

The 15 January 2022 climactic eruption of Hunga volcano, Tonga, produced an explosion in the atmosphere of a size that has not been documented in the modern geophysical record. The event generated a broad range of atmospheric waves observed globally by various ground-based and spaceborne instrumentation networks. Most prominent was the surface-guided Lamb wave (≲0.01 hertz), which we observed propagating for four (plus three antipodal) passages around Earth over 6 days. As measured by the Lamb wave amplitudes, the climactic Hunga explosion was comparable in size to that of the 1883 Krakatau eruption. The Hunga eruption produced remarkable globally detected infrasound (0.01 to 20 hertz), long-range (~10,000 kilometers) audible sound, and ionospheric perturbations. Seismometers worldwide recorded pure seismic and air-to-ground coupled waves. Air-to-sea coupling likely contributed to fast-arriving tsunamis. Here, we highlight exceptional observations of the atmospheric waves.

Volcanic CO2 tracks the incubation period of basaltic paroxysms.

The ordinarily benign activity of basaltic volcanoes is periodically interrupted by violent paroxysmal explosions ranging in size from Hawaiian to Plinian in the most extreme examples. These paroxysms often occur suddenly and with limited or no precursors, leaving their causal mechanisms still incompletely understood. Two such events took place in summer 2019 at Stromboli, a volcano otherwise known for its persistent mild open-vent activity, resulting in one fatality and damage to infrastructure. Here, we use a post hoc analysis and reinterpretation of volcanic gas compositions and fluxes acquired at Stromboli to show that the two paroxysms were preceded by detectable escalations in volcanic plume CO2 degassing weeks to months beforehand. Our results demonstrate that volcanic gas CO2 is a key driver of explosions and that the preparatory periods ahead of explosions in basaltic systems can be captured by precursory CO2 leakage from deeply stored mafic magma.

Gas flux cyclic regime at an open vent magmatic column inferred from seismic and acoustic records.

On August 7, 2014, a new effusive vent opened on the northern flank of Stromboli. A characteristic pattern was observed in both seismic and infrasonic signal amplitudes prior to this effusive eruption. The pattern consisted of the repeating cycle: (1) quiet phase, (2) puffing phase, and (3) explosion phase. Correlation between seismic and infrasound signal suggests that pulses in the puffing phase were caused by repetitive bursts of small gas pockets at the central crater, while the explosion phase coincided with an explosion at the central crater. We show that degassing of the magma column occurred in cycles of increasing gas flux, which controlled the transition from a bubbly flow (puffing phase), to a slug flow (explosion phase) gas regime. The quiet phase was characterized by a constant time length of 150 s, indicating that the gas rose in the magma column as well-organized waves of gas layers. These cycles represent cyclic changes of the gas flux regime in the shallow magma column, associated with increases in the magma-gas supply input rate before the effusive eruption.

Lattice Boltzmann modeling to explain volcano acoustic source.

Acoustic pressure is largely used to monitor explosive activity at volcanoes and has become one of the most promising technique to monitor volcanoes also at large scale. However, no clear relation between the fluid dynamics of explosive eruptions and the associated acoustic signals has yet been defined. Linear acoustic has been applied to derive source parameters in the case of strong explosive eruptions which are well-known to be driven by large overpressure of the magmatic fluids. Asymmetric acoustic waveforms are generally considered as the evidence for supersonic explosive dynamics also for small explosive regimes. We have used Lattice-Boltzmann modeling of the eruptive fluid dynamics to analyse the acoustic wavefield produced by different flow regimes. We demonstrate that acoustic waveform well reproduces the flow dynamics of a subsonic fluid injection related to discrete explosive events. Different volumetric flow rate, at low-Mach regimes, can explain both the observed symmetric and asymmetric waveform. Hence, asymmetric waveforms are not necessarily related to the shock/supersonic fluid dynamics of the source. As a result, we highlight an ambiguity in the general interpretation of volcano acoustic signals for the retrieval of key eruption source parameters, necessary for a reliable volcanic hazard assessment.

Forecasting Effusive Dynamics and Decompression Rates by Magmastatic Model at Open-vent Volcanoes.

Effusive eruptions at open-conduit volcanoes are interpreted as reactions to a disequilibrium induced by the increase in magma supply. By comparing four of the most recent effusive eruptions at Stromboli volcano (Italy), we show how the volumes of lava discharged during each eruption are linearly correlated to the topographic positions of the effusive vents. This correlation cannot be explained by an excess of pressure within a deep magma chamber and raises questions about the actual contributions of deep magma dynamics. We derive a general model based on the discharge of a shallow reservoir and the magmastatic crustal load above the vent, to explain the linear link. In addition, we show how the drastic transition from effusive to violent explosions can be related to different decompression rates. We suggest that a gravity-driven model can shed light on similar cases of lateral effusive eruptions in other volcanic systems and can provide evidence of the roles of slow decompression rates in triggering violent paroxysmal explosive eruptions, which occasionally punctuate the effusive phases at basaltic volcanoes.

Remote monitoring of building oscillation modes by means of real-time Mid Infrared Digital Holography.

Non-destructive measurements of deformations are a quite common application of holography but due to the intrinsic limits in the interferometric technique, those are generally confined only to small targets and in controlled environment. Here we present an advanced technique, based on Mid Infrared Digital Holography (MIR DH), which works in outdoor conditions and provides remote and real-time information on the oscillation modes of large engineering structures. Thanks to the long wavelength of the laser radiation, large areas of buildings can be simultaneously mapped with sub-micrometric resolution in terms of their amplitude and frequency oscillation modes providing all the modal parameters vital for all the correct prevention strategies when the functionality and the health status of the structures have to be evaluated. The existing experimental techniques used to evaluate the fundamental modes of a structure are based either on seismometric sensors or on Ground-based Synthetic Aperture Radar (GbSAR). Such devices have both serious drawbacks, which prevent their application at a large scale or in the short term. We here demonstrate that the MIR DH based technique can fully overcome these limitations and has the potential to represent a breakthrough advance in the field of dynamic characterization of large structures.

Volcano seismicity and ground deformation unveil the gravity-driven magma discharge dynamics of a volcanic eruption.

Effusive eruptions are explained as the mechanism by which volcanoes restore the equilibrium perturbed by magma rising in a chamber deep in the crust. Seismic, ground deformation and topographic measurements are compared with effusion rate during the 2007 Stromboli eruption, drawing an eruptive scenario that shifts our attention from the interior of the crust to the surface. The eruption is modelled as a gravity-driven drainage of magma stored in the volcanic edifice with a minor contribution of magma supplied at a steady rate from a deep reservoir. Here we show that the discharge rate can be predicted by the contraction of the volcano edifice and that the very-long-period seismicity migrates downwards, tracking the residual volume of magma in the shallow reservoir. Gravity-driven magma discharge dynamics explain the initially high discharge rates observed during eruptive crises and greatly influence our ability to predict the evolution of effusive eruptions.

Acoustic waveform of continuous bubbling in a non-Newtonian fluid.

We study experimentally the acoustic signal associated with a continuous bubble bursting at the free surface of a non-Newtonian fluid. Due to the fluid rheological properties, the bubble shape is elongated, and, when bursting at the free surface, acts as a resonator. For a given fluid concentration, at constant flow rate, repetitive bubble bursting occurs at the surface. We report a modulation pattern of the acoustic waveform through time. Moreover, we point out the existence of a precursor acoustic signal, recorded on the microphone array, previous to each bursting. The time delay between this precursor and the bursting signal is well correlated with the bursting signal frequency content. Their joint modulation through time is driven by the fluid rheology, which strongly depends on the presence of small satellite bubbles trapped in the fluid due to the yield stress.