GNGTS 2014 - Atti del 33° Convegno Nazionale

during the recent unrests. It would be therefore necessary to develop a seismic investigation that is not depending on the availability of earthquake data. Passive Image Interferometry. Passive Image Interferometry is a technique, which bases its origin in Aki studies on microtremors (Aki, 1957), but it has been mainly developed in very recent times since the availability of long continuous seismic recordings. It affords the seismic analysis from a completely new point of view: the objects of the study are the whole continuous recordings instead of their short cuts around the seismic events. This constitutes a great advantage because we dispose of data to be analyzed independently on the earthquake occurrences. The principle at the base of this technique is that cross-correlating the seismic noise recorded at two different stations we obtain a function that is related to the Green function of the medium between the two station locations (Lobkis and Weaver, 2001). To obtain the exact reconstruction of the Green function we need a homogeneous noise field in space and time (Campillo, 2006, and references therein). In Earth sciences this assumption holds, as a first approximation, if we are dealing with the noise generated by the oceanic waves. Although these noise sources are not uniformly distributed on the surface, and even if they experience very large seasonal variations (Landes et al. , 2010), the Green function�� �������������� ��� ’s reconstruction may be acquired thanks to:i) multiple scattering (scatterers act as secondary sources; Derode et al. , 2003); ii) time-reversal (i.e. taking long time series; Stehly et al. , 2006); iii) reciprocity between sources and receivers (i.e. a dense network may overcome the lack of noise sources; Paul et al. , 2005). These properties of the ambient noise cross-correlations have been used to get the Rayleigh wave (Shapiro et al. , 2004), as well as the P wave arrivals (Poli et al. , 2012), then to acquire tomographic images of the Earth crust (Shapiro et al. , 2005). Anyway, since we apply an interferometric technique, searching for temporal variations of the cross-correlation functions, it is not important the exact reconstruction of the Green function but the only requirement is the presence of quite stable noise sources (Hadziioannou et al. , 2009). Moreover we may overcome also the presence of non-isotropically distributed sources by cutting the central part of the cross-correlations since the source variations would affect the ballistic part of the signal, while the codas would be randomized by the effect of scattering (Froment et al. , 2010). Among the first studies of Passive Image Interferometry to monitor the crustal velocity variations, Brenguier et al. (2008a) could track the co-seismic drop of velocity and post-seismic relaxation of the crust along the San Andreas fault and during a period of time which included the occurrence of two major earthquakes (San Simenon M = 6.5, and Parkfield M = 6.0). Applications to volcanic environments are more rare compared to faults. Sens-Schonfelder and Wegler (2006) firstly analyze the relative velocity changes occurring on Merapi volcano by finding out that the more superficial layers of crust (tens of meters in depth) were seasonally influenced by the rainfall. Brenguier et al. (2008b) firstly associated the relative velocity changes to the eruptive activity by studying the Piton de la Fournaise volcano. This is probably the more studied volcano thanks to its alternation of quiescence phases and eruptions in very short times and the long record of data (Duputel et al. , 2009; Sens-Schonefelder et al. , 2014). On the contrary Colima volcano, which is very active as well, showed velocity variations only weakly associated with eruptive activity, probably reflecting the open state of the volcano during the 15 year period of study, while major changes were associated to large tectonic events (Lesage et al. , 2014). Finally Ueno et al. (2012) analyzed the seismic recordings from Izu peninsula and suggested a relationship between subsurface velocity changes and magma intrusion into the crust. All these previous works depict a great variety of results demonstrating the high potentiality of Passive Image Interferometry technique applied to the volcanic areas, and its capability to discern the crustal changes related to (hopefully future) changes in eruptive activity, large earthquake occurrences, magma injections at depth… thanks also to the nonlinear elastic behavior of the soft and not well compacted material characteristic of the volcanic edifices. 232 GNGTS 2014 S essione 1.3