GNGTS 2013 - Atti del 32° Convegno Nazionale

ascending and descending orbits (Figs. 1C, 1D and 1E). Using a simple dislocation approach we modeled 2-D velocity profiles obtaining a good fit to the observations (Profile A, B and C of Figs. 1C, 1D and 1E).Our model confirms the general left lateral kinematics of the DFS, but an additional important thrust component along the WFZ is necessary to fit the observations (rake angle of 34°±4°). The modeled rake angle results in the left-lateral component being about 2/3 of the slip rate, which is in agreement with the long-term record as reconstructed by structural and geomorphic observations. A steep fault dip to the North (~60°) is well constrained by the observations, and is in agreement with field observations on the Western and Central DFS. Our modeled slip rate of 5±1 mm yr -1 is the first quantitative estimate of strain accumulation for the Western DFS, corresponding to ~4 mm yr -1 of pure horizontal movement. In the long term, Fattahi et al. (2007) estimates ~2.4 mm yr -1 of left lateral slip rate on the central fault zone, by Infrared Stimulated Luminescence Dating (ISLD) of one Holocene alluvial fan. In our model, the anticorrelation between slip and rake indicates that some vertical component (rake > 0) is needed to explain the observations, otherwise the slip rates for pure strike slip fault would be unrealistic compared to the geological estimates. None of the latter are available forthe Quaternary uplift rate along the western fault zone (WFZ). Extrapolating the vertical component of our modeled slip rate (2.4±1 mm yr -1 ) in the geological past does not seem to justify the relatively low relief across the WFZ (500-700 m), but a number of different factors, e.g. observation and/or model uncertainties, spatio-temporal slip rate variations, interaction with close tectonic structures could justify this discrepancy. Based on a geologically determined average slip per event of 4.7 m, the recurrence interval along the Central DFS has been estimated to be ~2000 yr (Fattahi et al. , 2007). The slip rate we estimated for the WFZ would require a slip per event of ~10 m to obtain the same recurrence time. This is unrealistic, especially if we consider the field observations and the segmentation model by Farbod et al. (2011). Accepting a similar slip per event for the WFZ and the CFZ, including a conservative uncertainty estimate of 20%, we obtain a recurrence interval varying between 630 and 1400 yr. The lack of strong seismicity in the last 1500 yr along the WFZ may be due to the incompleteness of the historical seismic catalogues (Ambraseys and Melville 1982), and our results show that recurrence intervals below 1000 yr cannot be excluded. Ground deformation during the Coseismic phase. Regarding the contribution of geological constrains to the interpretation and modeling of coseismic displacement SAR maps, we present two examples in which a priori geological information had a very different significance. The first example concerns two earthquakes occurred in the Emilia region, Northern Italy, respectively on May 20 th 2012, Ml 5.9, and May 29 th , Ml 5.8, inverting COSMO-SkyMed and Radarsat-1 surface displacements and GPS observations (Fig. 2A) (Pezzo et al. , 2013). The Emilia seismic sequence filled a seismic gap existing at least since the year 1000 A.D. (Rovida et al. , 2011), and might therefore give important information on the mechanisms of strain accumulation and release in this area. Thus, some considerations can be made based on our results about the long-term strain accumulation. The location of the coseismic deformation measured by InSAR data corresponds to part of the Mirandola and Ferrara folds, located, under the Po alluvial plain. This evidence supports the long-term geomorphic analyses that attribute to the growth of the same folds the wide northward bend of the Po river course and the deviation of the Secchia and Panaro rivers (Burrato et al. , 2003). Both the 20 and 29 May 2012 sources were found to be well modeled by ~E-W, S-dipping thrust faults with a flat-ramp geometry (Figs. 2C and 2D), corresponding to the Mirandola and Ferrara thrusts. Furthermore, we identified a displacement pattern of ~10 cm towards the satellite sensors, not associated to any of the largest aftershocks (Fig. 2B). The pattern is temporally co-located or following the first event and preceding the second one. Spatially it is located halfway between the displacement fields of the two main events. We investigate some possible interpretations of our observations, favoring the hypothesis of a slip along the fault plane of the May 29 th event, supported by the results of a Coloumb Failure Function analysis, which suggests an increasing 97 GNGTS 2013 S essione 1.1