GNGTS 2019 - Atti del 38° Convegno Nazionale

120 GNGTS 2019 S essione 1.1 issue that still awaits to be completely elucidated (e.g. Mallet 1862; Janosi 1907; Kovesligethy 1907; Blake 1941; Sponheuer 1960; Ambraseys 1985; Burton et al. , 1985; Levret et al. ,1996; Musson 1996). To this aim, we analysed the seismicity of an area that includes the NE-verging Northern Apennines fold-and-thrust belt and its foredeep/foreland basin in the southern Po Plain, where earthquakes exhibit largely variable depth, ranging from very shallow (< 10 km) to deep (< 100 km). In the study region, the potential seismogenic sources belongs to one of seven genetically and geodynamically independent fault types (DISS Working Group, 2019; Vannoli et al. , 2015). In the Po Plain, besides the most external Northern Apennines thrust fronts, that are shallow blind thrusts associated for example to the 2012 seismic sequence (e.g. Maesano et al. , 2015), the foreland crust is cut by a set of deeper inherited faults that formed during the Mesozoic extensional phases and are located below the Northern Apennines basal detachment (Scardia et al. , 2015). Moving backward, the Northern Apennines thrusts continue with a ramp-flat geometry and are characterised by both shallow, along the mountain front, and deep seismogenic segments up to several tens of km of depth. At the back of the compressional domain, the ongoing extension is taken up by a regional, NE-dipping, low angle normal fault system (e.g. Boncio et al. , 2000). Further potential seismogenic sources are deep intraslab compressional faults and transverse fault systems related to the differential slab retreat. We used as input web-based macroseismic data from the HSIT database (http://www. haisentitoilterremoto.it/; Sbarra et al. , 2010; Tosi et al. , 2015), relating to instrumental earthquakes that occurred in the study area during the past 12 years and are characterised by a wide range of hypocentral depth. Thanks to this citizen-science system it is possible to gather a large amount of information from the population and to carry out detailed studies that due to the wealth of closely spaced data points may highlight amplification and attenuation areas also at urban scale (Sbarra et al. 2012). The direct correlation between the statistically analyzed web-based observation and the local seismic effects was recently highlighted using as comparison peak ground acceleration and velocity values (PGA and PGV; Molinari et al. , 2015; Sbarra et al. , 2017). Molinari et al. (2015) showed that peaks of ground shaking (PGV) predicted by the 3D MAMBo geological model for two moderate earthquakes occurred in the Po Plain matched the positive anomalies of the corresponding macroseismic intensity pattern as shown by the HSIT maps. At a more regional scale, Sbarra et al. (2017) studying the effects of both regional earthquakes and local shallow and deep earthquakes found that high attenuation zones were highlighted by low PGA values and vice versa . We started our analysis from three instrumental earthquakes that occurred in 2012 in our study area: the 20 May 2012, Mw 5.8 Emilia earthquake, with an hypocentral depth of 6.3 km; the 25 January 2012, Mw 4.9 event, with a depth of 29 km; and the 27 January 2012, Mw 4.9 event, having a depth of 72.4 km (Fig. 1). Fig. 1 - attenuation curves of the three events occurred in 2012 used as starting observation and characterised by different hypocentral depth.