Source Mechanism of Long Period events recorded by a high density seismic network during the 2008 eruption on Mt Etna

129 Long Period (LP) events, divided into two families of similar events, were recorded by the 50 stations deployed on Mount Etna in the second half of June 2008. During this period lava was flowing from a lateral fracture after a summit strombolian eruption. In order to understand the mechanisms of these events, we perform moment tensor inversions. Inversions are initially kept unconstrained to estimate the most likely mechanism. Numerical tests show that unconstrained inversion leads to reliable moment tensor solutions because of the close proximity of numerous stations to the source positions. However, single forces cannot be accurately determined as they are very sensitive to uncertainties in the velocity model. Constrained inversions for a crack, a pipe or an explosion then allow us to accurately determine the structural orientations of the source mechanisms. Both numerical tests and LP event inversions emphasise the importance of using stations located as close as possible to the source. Inversions for both families show mechanisms with a strong volumetric component. These events are most likely generated by cracks striking SW-NE for both families and dipping 70\deg SE (Family 1) and 50\deg NW (Family 2). For Family 1 events, the crack geometry is nearly orthogonal to the dike-like structure along which events are located, while for Family 2 the location gave two pipe-like bodies which belong to the same plane as the crack mechanism. The orientations of the cracks are consistent with local tectonics, which shows a SW-NE weakness direction. The LP events appear to be a response to the lava fountain occurring on the 10th of May, 2008 as opposed to the flank lava flow.

💡 Research Summary

This paper presents a comprehensive moment‑tensor analysis of 129 long‑period (LP) seismic events recorded during the second half of June 2008 on Mt Etna, using a dense network of 50 broadband stations deployed around the summit. The LP events occurred in the aftermath of a summit Strombolian eruption, when lava was being emitted from a lateral fracture on the volcano’s flank. Based on waveform similarity, the events were divided into two distinct families (Family 1 and Family 2), each consisting of a series of nearly repetitive bursts.

The authors first performed unconstrained moment‑tensor inversions to identify the most plausible source mechanisms without imposing any a priori physical model. Because many stations were located within a few hundred meters of the sources, the moment‑tensor components (especially the isotropic or volumetric part) were recovered with high fidelity, whereas the single‑force components proved highly unstable. Numerical experiments demonstrated that the dense, near‑source station coverage mitigates the trade‑off between moment‑tensor and force terms, confirming that reliable moment‑tensor solutions can be obtained even in the presence of moderate velocity‑model uncertainties. However, the single‑force terms remained sensitive to errors in the assumed velocity structure, preventing a robust estimation of any net force acting on the source.

To overcome this limitation and to extract physically meaningful information, the authors then imposed three constrained source models: (i) a tensile crack (double‑couple plus isotropic component), (ii) a pipe‑like conduit (vertical tensile source), and (iii) an explosive point source (pure isotropic). By fitting the observed waveforms under each constraint, they were able to discriminate between the models and to retrieve the orientation (strike, dip, and slip direction) of the dominant crack‑type mechanisms.

Both families exhibited a strong volumetric component, indicating that the LP events were not pure shear failures but involved significant opening or closing of a fracture. For Family 1, the best‑fitting crack strikes roughly SW‑NE and dips 70° toward the SE. For Family 2, the crack also strikes SW‑NE but dips 50° toward the NW. Notably, the crack plane of Family 1 is nearly orthogonal to the dike‑like structure along which the events are spatially aligned, suggesting that the LP bursts may have been triggered by stress release perpendicular to an existing magma pathway. In contrast, the Family 2 events appear to be associated with two pipe‑like bodies that lie within the same plane as the inferred crack, implying a combined crack‑and‑conduit geometry.

The orientations of both crack planes are consistent with the regional tectonic fabric of Etna, which is characterized by a SW‑NE weakness direction. This alignment supports the interpretation that the LP events are a response to the large lava fountain that erupted on 10 May 2008, rather than to the later flank lava flow. The timing of the LP activity, its spatial clustering, and the inferred source geometries all point to a rapid pressure perturbation within the shallow magmatic system caused by the fountain, leading to the opening of pre‑existing fractures and the generation of the observed LP seismicity.

The study emphasizes several methodological lessons. First, the proximity of a dense seismic array to the source dramatically improves the stability of moment‑tensor inversions, especially for low‑frequency LP events that are often dominated by volumetric changes. Second, while unconstrained inversions can reliably recover the isotropic component, incorporating physically motivated constraints (crack, pipe, explosion) is essential for extracting the geometrical parameters of the source. Third, the sensitivity of single‑force estimates to velocity‑model uncertainties cautions against over‑interpreting net force terms without independent constraints.

In conclusion, the authors demonstrate that high‑resolution, near‑source seismic monitoring combined with a systematic inversion strategy can unravel the complex source mechanisms of volcanic LP events. Their findings provide a clear picture of how shallow magmatic pressure fluctuations, likely induced by a vigorous lava fountain, can activate SW‑NE‑oriented fractures and conduits in the Etna edifice. This work not only advances our understanding of LP seismicity at Etna but also offers a robust framework for investigating similar low‑frequency volcanic signals at other active volcanoes worldwide.