GNGTS 2014 - Atti del 33° Convegno Nazionale

52 GNGTS 2014 S essione 1.1 been recorded by permanent GPS stations managed by different private and public institutions located around the epicentral area (Figs. 2 and 3). Most of the GPS sites are located in the far field, whereas only few stations were available for the epicentral area. A complete description of the processing details and strategies are given by Serpelloni et al. (2012). We extend the GPS time series to the acquisition interval of the CSK and levelling data set. We modelled site positions as the sum of a coseismic offset and a time-dependent function, which reproduces coseismic displacement and post-seismic response (e.g., Langbein, 2008). The model equation contains parameters that describe the time-dependent signal characterised as follows: where 0 i y is the pre-seismic position, i A is the coseismic offset, t is the time after the earthquake, H is the Heaviside step function, i B is the exponential decay amplitude and τ is the exponential decay time for the i -th component. We calculate an additional coseismic offset associated with the May 29 Mw 5.8 mainshock. Time dependent post-seismic deformation is modelled only for the May 29 mainshock. In this way we assume that the contribution of post- seismic deformation of the May 20 event is minimal, as verified by Pezzo et al. (2013) from a CSK ascending interferogram spanning the period May 22-26. The coseismic displacement field for the May 20 event presents the largest horizontal displacements at Finale Emilia (MO05) which moved to the SSW by ~3 cm, at SERM station that moved southwards by ~1.5 cm and at SERM station that moved NNE-wards by ~2 cm. The other GPS sites, located around the epicentral area, show horizontal displacements < 1 cm. For the May 29 event, all GPS horizontal displacements converge toward the area of the mainshock, with displacements smaller than for the May 20 event (< 1 cm). The preliminary analysis of the post-seismic GPS time series shows that few stations (e.g., SERM, SGIP and SBPO) underwent a progressively decaying motion well-reproduced by a time-dependent function typical of post-seismic evolution. In particular the north components of SERM and SBPO sites move southward, while the north component of SGIP station shows a movement towards north (with horizontal displacements of ~1-2 cm covering ~1 year of post- seismic deformation). Preliminary results of the inversion of geodetic data. In this section we illustrate the results of the inversion for distributed slip on a pair of fault planes corresponding to the Ferrara and Mirandola thrust faults. We combine all geodetic datasets, that is, levelling measurements, InSAR interferograms and GPS data, to simultaneously infer the coseismic and post-seismic slip distributions on the fault planes that ruptured in the two mainshocks. Before modelling, the interferograms were downsampled using a resolution-based resampling technique (Lohman and Simons, 2005), with the highest density of data points close to the source of deformation. We compute the Green’s functions, which relate unit slip on individual fault patches to surface displacements at individual points using dislocations in an elastic, homogeneous and isotropic half-space (Okada, 1985). The Ferrara and Mirandola thrust faults are thus discretized into smaller patches. We explore different possible geometries (e.g., listric versus straight geometry) and dip angle of the rupture planes, defined by moment tensor solutions, aftershock distribution and mapping studies, in order to better define the segments of the Ferrara arc that were activated during the two mainshocks and aftershock sequence. We take as a starting point the fault geometry of the main active thrusts proposed by Boccaletti et al. (2011), which by using morphotectonic, geological-structural, and stratigraphic analysis compared with the current seismicity of the region, suggested that the external part of the northern Apennines is characterised by presence of major systems of Quaternary compressive structures corresponding to the Emilia, Ferrara and Adriatic Fold systems buried below the Po Plain. These structures have a typical arcuate shape in map view (Figs. 2 and 3), giving rise to the Emilia and Ferrara folds (Pieri and Groppi, 1981). In particular, in the aftershock region, the Ferrara arc is structured in two major fold-and-thrust systems, the Ferrara system in the NE