Re-pressurized magma at Mt. Etna, Italy, may feed eruptions for years

Understanding volcanic processes is crucial for hazard assessment. While some volcanoes erupt abruptly, others exhibit prolonged unrest before eruption, with the duration and precursory signals remaining poorly understood. A major challenge is predicting eruption duration and lava volume. Decades of research on Mount Etna, a volcano with intense and varied activity well-monitored by modern systems, have significantly advanced our understanding. Located at the leading edge of the Apennines subduction system, Etna's hybrid origin reflects complex dynamics from the interaction of low-viscosity magma sourced from the asthenosphere with the surrounding crust. While continuously active, Etna experiences episodes of inflation and increased eruptions. Since December 2020, following a period of steady activity since September 2019, volcanic activity has sharply resumed, with paroxysmal events and lava fountains starting in mid-February 2021. This reactivation followed an overall inflation trend observed by GNSS data and increased seismicity since summer 2020. Petrological analysis reveals that the magma erupted since December 2020 is among the most primitive and gas-rich in the last 20 years. This increase in seismicity included shallow (z≤3 km) clusters in the central area and deeper clusters along the flanks. Deep peripheral earthquakes are considered markers of volcano recharging, often preceding eruptive events. This study utilizes near real-time imaging of the magmatic system via time-lapse local earthquake tomography to identify and locate new melt accumulation within crustal reservoirs, enhancing understanding of the ongoing dynamics and providing a quantitative assessment of recently added melt for improved eruption forecasting.

Extensive studies have explored Mt. Etna's magmatic system, revealing a central high-velocity magma mush extending from 3 to 15 km depth. This mush, formed by long-term activity, incorporates gabbroic intrusions and interacts with deep fluids, influencing the re-feeding of shallow reservoirs and flank instability. The volcano's recent history shows periods of intense central activity alternating with phases of shallow intrusions and increased eastern flank sliding. Previous studies using time-lapse tomography identified local variations near shallow intruding dykes, but failed to fully capture the volume of melts involved in larger eruptions. The 2000 explosive summit activity was well-documented, following pressurization of a crustal reservoir, but again lacked quantitative estimates of melt volumes. The current study aimed to address this gap by quantifying magma volumes using time-lapse tomography, leveraging the abundant and diffuse seismicity as a prerequisite for its feasibility.

Time-lapse tomography, which analyzes changes in seismic wave velocity over time, was employed to image the Mount Etna magmatic system. This approach involves computing 4D velocity changes (in space and time) by comparing 3D velocity models from different time intervals. The variations in seismic properties are correlated with changes in stress, deformation, fluid migration, and melt accumulation. Challenges in time-repeated tomography stem from variations in ray sampling between different periods due to non-optimal distribution of earthquake sources and receivers. Careful data selection and inversion are crucial to minimize artefacts. The study compared a static 3D image of the volcano (pre-2019) with data from December 2019 to February 2021. The static model served as the starting model for inversion. Analysis of P- and S-wave residuals provided initial indications of significant velocity changes. By selecting specific events and stations, the study identified positive residuals accumulating beneath the volcano's central portion, indicating the presence of slower anomalies. Tomography revealed several velocity changes, with the most significant being a broad volume of reduced Vp velocity centered at 6 km depth beneath the central area (R-3), extending between 3 and 9 km depth. While Vp reduction was well-resolved, Vs resolution was poor, potentially due to attenuation of S-waves passing through the anomalous body. Two smaller, shallower volumes (R-2 and R-1) with clear Vp reduction were also identified beneath the summit craters. The resolution of the three anomalous volumes was assessed through resolution matrix analysis. To ensure robustness, the inversion was conducted using a conservative approach, first generating a static 3D model using the entire dataset to absorb heterogeneities. The time-lapse tomography was then computed using the 2019-2021 period, with the static model as reference. A synthetic test was conducted where an artificial low-velocity anomaly was introduced to assess the method's reliability. The study used local earthquakes from the INGV-OE database, recorded between January 15, 2019, and February 2021. Around 80% of events had magnitudes above 1.0, with the maximum reaching Ml = 4.1. The study utilized 3140 earthquakes with at least 6 P-wave and 2 S-wave arrival times, applying rigorous selection criteria to ensure data quality. The data was processed using the Simulps code, an iterative damped least-squares algorithm that solves for Vp and Vp/Vs. To assess the reliability of the tomographic results, ray sampling and full resolution matrix were analyzed, with a Spread Function (SF) ≤ 3.0 used as a threshold for well-resolved nodes.

Time-lapse tomography revealed three distinct low-velocity anomalies representing areas of melt accumulation within the Mt. Etna magmatic system. The largest anomaly (R-3) is a deep, elongated volume located at a depth of 4-9 km beneath the central area, suggesting significant vertical distribution of magma accumulation at the northern edge of a high-velocity magma mush. This volume is interpreted as the primary storage for primitive melts ascending from the mantle source. Two smaller anomalies (R-2 and R-1) located at shallower depths (~1 km) beneath the summit craters are linked directly to the summit crater activity. Analysis of velocity residuals over time suggests that the melt accumulation started in mid-September 2019, accelerating in December 2020. The observed changes in P-wave velocity (approximately -6%) are estimated to correspond to a melt fraction of about 4% within the main reservoir (R-3) over the two-year period. However, this may represent a lower limit, as linearized tomographic methods can underestimate the amplitude of the anomalies. The seismicity observed correlates with the re-pressurization, clustering around the high-fluid pressure volumes but not within the reservoirs themselves. The study observes almost simultaneous deep and shallow melt injections. The volume of erupted lava is only a fraction of the accumulated melts. The significant re-pressurization of the main reservoir suggests that eruptions could continue for several years.

The findings provide crucial insights into Mount Etna's dynamics. The observed pattern of deep crustal earthquakes (z≥ 12 km) followed by melt accumulation and re-pressurization in the main reservoir (R-3) supports a model where deep earthquakes signal the ascent of new magma from the mantle. The re-pressurization induces seismicity clustering, consistent with observations from previous events. The rapid on/off of shallow seismicity around R-2 indicates episodic magma uprise from the main reservoir to the shallow reservoirs that directly feed the summit crater activity. The finding that the erupted lava represents a minor fraction of the accumulated melt offers a critical perspective on the potential for future eruptions. The study quantitatively demonstrates that significant melt accumulation persists and could sustain eruptive activity for an extended duration.

This study provides the first tomographic evidence of significant fresh melt accumulation within Mount Etna's central magmatic system since 2019. The substantial volume of accumulated melt, significantly larger than that erupted, implies the potential for prolonged eruptive activity. This research advances our understanding of the volcano's plumbing system and emphasizes the potential for long-term forecasting using time-lapse tomography.

The study's resolution is limited by the spatial sampling of the seismic data, potentially affecting the precise delineation of the melt bodies. The method relies on approximations in the relationship between seismic velocity changes and melt fraction, introducing some uncertainty in the quantitative estimation of melt percentages. The study's focus on a relatively short time window may not fully capture the long-term evolution of the magmatic system.