How does the earthquake probability change after the flank eruptions of Etna?

by Andrea Bevilacqua, Raffaele Azzaro, Stefano Branca, Salvatore D’Amico, Franco Flandoli, Augusto Neri.

Click here for Italian version

Historically, flank eruptions of Etna, i.e. those producing lava flows from fissures on the flanks of the volcano, are often accompanied by destructive seismic events in the Etnean region. The summed impact of these two phenomena on the territory can be a severe burden for local communities and a complication in the emergency management.

The occurrence, in a short time interval, of both eruptions and earthquakes represents a research topic debated for decades. Is there a statistical relationship between the two phenomena? In fact, current research poses questions like: “How probable is a strong earthquake after a flank eruption?” And then, “Which systems of faults are the most sensitive to these eruptions?”

This topic is challenging, and still widely unknown; however the probability estimate of seismic events associated with an eruption represents a precious piece of information that may be used to better assess volcanic hazard and risk and plan civil protection tasks. A study aimed at responding to these questions has been recently published by INGV, in collaboration with Scuola Normale Superiore di Pisa, in the ambience of a research project co-funded by the Italian Civil Protection.

With this objective in mind, the behavior of Etna was investigated in the last four centuries, to identify, from historical catalogs of eruptions and earthquakes, the statistical relationships existing between the two phenomena. In quantitative terms, it was possible to describe how evolves the annual average of flank eruptions and most energetic seismic events. In addition, the study quantified the mean probability of earthquakes in the days/weeks after the onset – or, in alternative, the end – of a flank eruption. This probability was then compared to the average based on the entire historical record analized.

In particular, the study considered both the earthquakes capable of light damage and those, less frequent, that produce severe damage. The events of the first typology are those with a maximum intensity greater or equal to the V-VI degree of the European Macroseismic Scale (EMS); the most serious events those with maximum intensity greater or equal to VII-VIII degree of EMS.

The analysis of the flank eruptions sequence since 1600 (see Figure 1) highlighted that the occurrence of a flank eruption does not affect the probability of another occurring in the next months or years. In fact, this is different from what is observed on other volcanoes.

Moreover, it emerged that flank eruptions are not statistically independent from major earthquakes: in the last two centuries, about 30% of flank eruptions was followed in less than 30 days by an earthquake with intensity I ≥ V-VI EMS, i.e. a bit less than one in three eruptions. Nevertheless, about 18% of these earthquakes followed a flank eruption in less than 30 days. Figure 2 shows an example of the damage produced by the earthquake of 26 December 2018, which occurred two days after the onset of a flank eruption of Etna.

The following question rises naturally: “How much greater is the probability of the earthquakes after a flank eruption, if compared to the average?”

The study quantified that the probability of having strong earthquakes increases by 5-10 times after the onset of a flank eruption and this increment lasts for the following 30-45 days; afterwards, the earthquake rates go back to the average. Also after the end of the eruptions it is observed a similar increment. In fact, Figure 3 shows the evolution, week after week, of the occurrence probability of major earthquakes as a function of the number of days passed after the last flank eruption. The plot also highlights that in some cases the strong earthquakes precede of few days the onset of flank eruptions.

In addition, the localization of eruptive fissures in the spatial sectors of Etna influences the statistical relationship between eruptive and seismic phenomena considered here. For example, it was observed that about 50% of the eruptions on the northeastern flank of the volcano preceded a strong earthquake, i.e. one in two eruptions. In this case, the major earthquakes after a flank eruption are up to 10-20 more likely than average for 65-70 days.

Similarly, the spatial groups of the faults responsible for the major earthquakes can affect the statistical relationship with flank eruptions. For example, about 30% of earthquakes of Pernicana fault preceded – instead of followed – a flank eruption, and 40% of them preceded a strong earthquake in the other systems of faults; i.e. a 10-20 increment of the daily probability of both phenomena, for the following 45-70 days.

In conclusion, thanks to the statistical analysis of the historical catalogs, the probability estimates highlighted a strong connection between flank eruptions and earthquakes in the different sectors of the volcano, confirming that the two phenomena are significantly linked in short time intervals.

The results obtained in this study contribute to better constraining the temporal and spatial scales of the physical mechanisms that control the behavior of the volcano. This research also managed to quantify, in statistical terms, the observed correlation between the two phenomena in the most recent centuries, useful information to plan an effective emergency management for future seismo-volcanic crises at Etna.