GNGTS 2019 - Atti del 38° Convegno Nazionale

98 GNGTS 2019 S essione 1.1 THE 2016-2017 CENTRAL ITALY SEISMIC SEQUENCE: FROM A REAL-TIME ANALYSISTS DATA TO AN AUTOMATIC AND NEARLY REAL TIME CATALOG M. Michele, L. Chiaraluce, R. Di Stefano INGV (National institute of Geophysics and Volcanology), Rome Within few months after the first mainshock (Mw 6.0, Amatrice 2016/08/24) of the Central Italy seismic sequence, a quasi-automatic procedure was developed to locate the aftershocks occurring in the epicentral area, by using the data collected in monitoring environment at INGV (National Institute of Geophysics and Volcanology). The location code implemented in the procedure was NonLinLoc (Lomax et al. , 2001), based on a probabilistic approach in determining the hypocenter location, that allows to retrieve information about the probability density function related to it. For this reason, it gives very exhaustive estimates of the location uncertainty and robustness. We used a 1D velocity model and we included in the relocation process the stations corrections to balance the lack of information about shallow depth layers. With this setting, we produced the first absolute locations catalog for the seismic sequence (Chiaraluce et al. , 2017), covering the period 2016/08/24 – 2016/11/30. We built an automatic procedure to update in near-real time the locations occurring during the sequence, up to January 2018. Considering that we reached more than 90,000 locations, we tried to improve the retrieved locations, by using the double-difference method, implemented in the HypoDD code (Waldhauser et al. , 2001) extending the catalogue in time up to 2018/01/18. Since our goal was to highlight active structures, building a fast procedure, able to produce in a short time refined pictures of structures, we selected the earthquakes locations’ qualities (RMS £ 0.5 s, horizontal and vertical formal errors £ 1.5 km and £ 3.0 km, respectively, at least 10 P- and S- phases, and an azimuthal gap £ 180°) and the earthquakes magnitudes (M w ³ 1.5). The selected dataset, including ~34,000 events, consists of 612,207 P and 465,693 S travel times recorded by the INGV monitoring room personnel and ~4.4 millions of P and ~1.1 millions of S times computed by means of cross-correlation procedure. The obtained relocations allow us to describe the fault system structure, achieving improved locations resolution, that highlight minor anti- and synthetic structures not visible before and that give insights about the basal layer (Michele et al. , submitted to JGR). In particular, we focused on the space-time evolution of the sequence, on the system geometry description, exploring details in map view and in cross-sections. The resulting complex fault system is 80 km long and NW-SE trending, composed of SW-dipping normal faults segments (15-30 km long). The dip of the entire fault system is variable and ranges from 38° to 55° moving from South to the North, while the strike is pretty constant. The updated locations clearly underline the shear-zone, whose depth ranges from 6 to 9 km, presenting diverse undulations and a step probably corresponding to the intersection between the main and the secondary faults of the Norcia earthquake (2016/10/30). Seismicity alignments seem to correspond to geological faults at depth, looking at the comparison between the geological profiles and cross-sections. Another focus regards the comparison of the relocated 2016-2017 sequence with the two previous Central Apennines sequences: Colfiorito 10997 and L’Aquila 2009. Looking at the structures activated during the sequence and before, we see that to the North, the faults described by the seismicity are parallel to the Colfiorito fault, while to the South seismicity aligns on the Campotosto fault, already activated during the 2009 sequence, and extends its lengths at depth. The approach used, able to recover the uncertainties on the travel times recorded by mean of cross-correlation procedure and to achieve accurate scientific results starting from a real- time analysists data, suggested to adopt this work flow in near real time, aiming to determine