GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 1.1 51 SLIP IMAGING OF THE MAY 20, 2012, MW 5.8, PO PLAIN (NORTHERN ITALY) EARTHQUAKE FROM THE INVERSION OF THE SOURCE TIME FUNCTIONS V. Convertito, N.A. Pino Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Napoli - Osservatorio Vesuviano Introduction. The May 2012 seismic sequence occurred in the Po Plain (Northern Italy) and started on May 19, 2012, at 23:13:27 GMT with a Mw 3.8 earthquake. On 20 May a Mw 5.8 event was recorded, followed by a second Mw 5.6 main shock on 29 May (http://cnt.rm.ingv.it/ tdmt) and thousands of aftershocks, six of them with magnitude larger than 5.0 (Govoni et al. , 2014). The sequence caused 27 fatalities and widespread severe damage to dwellings forcing the closure of several factories. This seismic sequence took place on a south dipping blind thrust fault system (Ferrara arc) in the Emilia-Romagna region, covered by the quaternary sediment of the Po Plain. The aim of this study is to analyse the rupture properties of the largest event, occurred on May 20. In particular, we analysed rupture kinematics and image the slip distribution from the analysis of the apparent source time functions by using two different modelling. Method. We applied an Empirical Green’s Function (EGF) approach (see, for instance, Mori (2003) and reference therein) to obtain relative Source Time Functions (STFs) as seen at different azimuths. In order to gain information about the source directivity and the slip distribution we analysed the obtained STFs by using a forward modelling illustrated in Convertito et al. (2016) and a Bayesian inversion modelling described below. In particular, for this latter, the direct problem is solved by assigning the slip at a set of a given points (control points) distributed on the fault plane and then interpolated on a finer grid. The number of control points defines the size of the subfaults and is selected on the basis of the magnitude of the EGF. Each subfault is described by: the slip value, the rise-time and the onset time. Slip distribution is mapped on a finer grid by means of a bilinear interpolation and by using a Gaussian bidimensional filter (e.g., Király-Proag et al. , 2019). The size of the finer grid is selected according to the coherent rupture condition of six source points per wavelength (Archuleta and Hartzell, 1981). The nucleation point is located at fault center and the rupture propagates at a constant rupture velocity. At each source depth we evaluate the v p value (i.e., the propagation velocity of the selected seismic phase) using the crustal model proposed by Govoni et al. (2014) and compute the Mach number a=v r /v p , being v r the rupture velocity. Each sub-fault is allowed to slip only once Fig. 1 - Geographic map showing the location of the May 20, 2012, Mw 5.8, Po Plain (Northern Italy) earthquake along with the stations used in the present study.