GNGTS 2013 - Atti del 32° Convegno Nazionale

sector of the volcano highlights that an increment of the occurrence frequency of the events precedes by some days (sometimes months) the unrest of several flank eruptions and some summit eruptions (Fig. 2). It is indeed quite clear that the increase of seismic activity starts in the deeper seismogenic level and continues in the shallower one. We are keen to interpret these results as a first sign of the deep (z > 12 km) recharging of the volcanic system that is soon after followed by the increment of the seismic rate at shallower levels which seems to mark a sort of magma migration (5 < z ≤ 12 km). When the magma rising process keeps going it triggers shallow seismic swarms (z ≤ 5 km) indicating the opening of eruptive fractures. It has however to be specified that such tendency towards a magma migration is not observed in some instances (i.e. before the 2002 and 2004 eruptions). In such cases it could be hypothesized (Andronico et al. , 2005; Collins et al. , 2013) that magma was already still standing in the shallower part of the volcano consequently to a previous uprising. Temporal Analysis of Seismicity. Trying to describe the time evolution of the seismicity recorded onMt. Etna in the study period, it appears clear (Fig. 1d) a change in the seismic style of the volcano. The cumulative curve shows indeed a significant difference in dip before and after 1987. Moreover it is noticeable that all summit eruptions, except the 1998 one, occur without significant increments of the seismic rate whereas most of the flank eruptions are marked by a sharp increase of the seismic activity. Inspection of Fig. 1d shows that the dataset 1976-2011 can be subdivided into four time intervals characterizing both periods of intense volcanic activity with lateral eruptions and periods with summit activities without flank episodes. The IET distributions was therefore carried out during the following periods: 1976-1987, 1988-2000, 2000-2003, 2003-2011. In Fig. 3 the values of skewness (S) for each period are reported since they better represent the different seismic styles and the IET distribution patterns. During the 1976-1987 time interval several eruptive events, both of lateral and summit kind, occurred. The chiefly active seismogenic level, as defined by the statistical parameters, is the one at depth z ≤ 5 km. It shows high S values (Fig. 3a) pointing out that we deal with several correlated events especially in the summit area. At z > 5 km the skewness assume slightly negative values (S < -0.5) indicating seismicity characterized by independent events with relatively long recurrence times. During the second time interval (1988-2000) only summit eruptions occurred, except the 1991-1993 flank activity. In this case all the seismogenic levels appear equally affected by the seismic activity and the observed statistical parameters are those typical of time stationary activities represented by independent events (S < 0) linked to the background seismicity (Fig. 3b). It is in particular noticeable that, as already observed when the whole dataset has been analyzed (1976-2011), a marked difference between the seismic style of the western and eastern sectors of the volcano is set into evidence particularly at depths z > 5 km, the former showing some correlated events. The values of the statistical parameters calculated for the time interval 2000-2003, for the depth level z ≤ 5 km are quite pronounced (0.5 < S < 2.5, Fig. 3c), pointing out in this way the tight relation existing between seismic and volcanic activities, relationship that is set into evidence by IET distributions with a lot of correlated events (seismic swarms preceding eruptive activities). Although the analyzed interval is particularly short, the small number of data anyway allows us to observe, similarly to the findings obtained when the whole dataset was analyzed, that below 5 km depth a distinct different behaviour of the western and eastern sectors of Mt. Etna is still detectable (Fig. 3c). The dataset concerning the period 2003-2011 is also formed by a few data so that information regarding only small sectors of the volcano can be inferred from the IET distributions. Similarly to the 1988-2000 period, low values of the statistical parameters are obtained for the depth interval z ≤ 5 km, therefore indicating the presence of background seismicity only without earthquake swarms (Fig. 3d). Below 5 km depth, in the NW sector of the volcano high values of skewness (S ~ 1.5) are obtained, suggesting the occurrence of seismic swarms. Such seismic activity occurred in recent time (after 2009) and appears not linked to any eruptive phenomenon. 282 GNGTS 2013 S essione 1.3