GNGTS 2018 - 37° Convegno Nazionale

GNGTS 2018 S essione 1.1 99 direction. Then it reaches the climax on October 30 with the Mw 6.5 earthquake, the largest event of the sequence, that hits right in between Norcia and Ussita. The third part of the sequence started January 18, 2017, with the occurrence of four moderate- magnitude earthquakes (5.0 ≤ Mw ≤ 5.5) located in the area south of Amatrice, close to the Campotosto area, elongating the seismogenic volume activated by the sequence in the South-East direction. Ambient noise seismic monitoring. Noise-based seismic monitoring of relative velocity variations is a recently developed technique that has been proved as an efficient way for tracking crustal changes at the time of earthquake occurrences (Brenguier et al. , 2008a), and even before impending eruptions (Brenguier et al. , 2008b). This analysis base on the property of seismic noise cross-correlation function computed between two stations, which corresponds to the impulse response of the medium (the Green’s function) recorded at one station as if the source was placed on the other station location (Weaver and Lobkis, 2001). Thus, any change in travel times measured from noise cross-correlation functions corresponding to different time periods reflects relative velocity variations, i.e. the elastic property modification, in the propagation medium within the stations. We adopt the Multi-Window Cross-Spectrum analysis firstly applied to earthquake coda waves (Poupinet et al. , 1984; Ratdomopurbo and Poupinet, 1995), and then transferred to cross-correlation codas (Brenguier et al. , 2008a). We apply this technique to the 2016-2017 Central Italy seismic sequence in order to get the temporal variability of the crustal modifications Crustal velocity changes. We considered here two years (from January 2016 to December 2017) of continuous seismic recordings from 28 stations of the Italian Seismic Network inside a radius of 90 km around the Amatrice village (see map on Fig. 1). After some preliminary operations on the data (instrument correction, 1-bit normalization, whitening in the [0.1 1] Hz frequency band), we computed the 1-hour cross-correlations for the entire period between the 378 couples of stations. We then stacked the 1-hour cross-correlations in order to obtain a reference function from the sum over the two years of data, and many current functions by stacking 50 days sliding 2 days, for each station couple. Finally, putting together the result of the comparison between currents and reference functions of each station couples, we end up with the temporal evolution of the relative seismic velocity undergone in the whole volume of crust encircled by the seismic stations from the surface to a depth of few km (Fig. 2). After a nearly constant trend during the first months of 2016, the relative velocity variations exhibit three distinct drops, suddenly starting at the time of occurrence of the major events (Mw ≥ 5.5), and a slower increasing trend, starting after the last mainshock occurrence and almost completely recovering to pre-mainshock values towards the end of 2017. These patterns are Fig. 1 - Map view of the stronger aftershocks (Mw ≥ 4.0) of the 2016-2017 seismic sequence (grey circles). The red stars indicate the epicentral locations of the nine Mw ≥ 5.0 earthquakes, black triangles are the seismic stations of the Italian Seismic Network considered in this study.