Introduction
The Apennines of Italy is an outstanding example of syn- and late-orogenic extension spreading over and after the collision between the continental Africa and Eurasia plates (Dercourt et al., 1986). Although the paired extension and compression appear to be common to other Tethys regions (Hymalayas, Molnar et al., 1993 and the Aegean, Jolivet et al., 1998a), this process is still a debated feature of continental deformation. Well documented for the Apennines is a progressive eastward migration of a coupled pair of compression and extension, from the Tyrrhenian to the Adriatic regions (Elter et al., 1975; Bally et al., 1986). Barchi et al. (2006) proposed three main rules for the northern Apennines, for which the compression and extension are always tightly related in space and time.
Figure 1. Instrumental seismicity of northern-central Apennines.
Seismicity of northern-central Apennines in the period 1980-2008 /locations form the CSI catalog. The colour scale is relative to the depth of earthquake. A schematic representation of the thrust fronts is shown. The white dashed line is the limit of the high velocity anomaly in the mantle revealed by Giacomuzzi et al., 2010. The rectangle indicates the region zoomed in Figure 2, the black lines the traces of sections in figures 7 and 8.
The present-day outer compressive front runs broadly along the Adriatic coast of the Italian peninsula (Figure 1) and shows a curved and irregular geometry. Intriguingly, the front is broadly parallel to the present day convergence between the Africa and Eurasia plates and spaced only some tens of kilometres from the extensional region.
Since the late 80's, different models described the Apennines as formed by the subduction and retreat of a mostly continental lithosphere (Malinverno and Ryan, 1986; Dewey et al., 1989; Doglioni, 1991), developed after the Tethys ocean closure and plate collision, the former Alpine history. Geological and geophysical data accumulated in the 90's were used to support geodynamic models whose common base is the W-dipping subduction of Adria (Patacca and Scandone, 1989; Serri et al., 1993; Gueguen et al., 1998; Faccenna et al., 2001; 2003; Peccerillo and Lustrino, 2005; Scrocca et al., 2007; Avanzinelli et al., 2009). The fast slab rollback and retreat is indicated as the cause for the eastward migration of the paired compression-extension pair (Faccenna et al., 2003) and opening of the Tyrrhenian back arc. Such evolution occurred at the back of the former developed Alpine belt, after the Eurasia and Africa collision, which remnants are visible in the Tyrrhenian side and, further south, in the Calabrian arc.
Evidence for the deflection of the Adria lithosphere beneath the belt are numerous and from independent sets of data. Flexural model of subsidence data supported a model of fore-deep formation controlled by deep subsurface load, like a sinking slab (Royden and Karner, 1984; Royden et al., 1987; Royden, 1993, Doglioni, 1994). Modelling is coherent with the deflection of the regional monocline, i.e., the Mesozoic sedimentary sequence of Adria (Mariotti and Doglioni, 2000). Intermediate-depth seismicity occurring along the Apennines (Selvaggi and Amato, 1992; Chiarabba et al., 2005; Chiarabba et al., 2009a) consistently defines the west-dipping Adria lithosphere down to a depth of about 70 km. On the contrary, only tomographic models suggest that the slab is sinking in the upper mantle, revealing the positive velocity anomalies that plunge into the mantle (Lucente et al., 1999; Wortel and Spakman, 2000; Piromallo and Morelli, 2003). Very recently, seismic anisotropy showed that the mantle flow, indicated by the SKS splitting, is not that expected by the retreat of Adria (Salimbeni et al., 2007).
The linked formation of Alps and Apennines is addressed by rotation of individual blocks of the belts revealed by paleomagnetic studies (Mattei et al., 1995; Speranza et al. 1997; 1998; Maffione et al., 2008 and references therein). Those data suggest a common dynamic that drives the Alpine and the Apennine slabs evolution, with the last process being a rollback of the Apenninic slab and related back-arc spreading of the Liguro-Provencal Basin and drift of the Corsica-Sardinia block.
In this scenario, extensional tectonics spread along the Apennines range previously formed by compression. Normal faulting earthquakes, with magnitude up to 6.9, develop on a set of adjacent and fragmented NW-trending faults. For the most recent large earthquakes (Chiaraluce et al., 2004; 2005; Chiarabba et al., 2009b), instrumental data indicate that the dip of the normal fault is of about 40-50 degree, smaller than that expected for Anderson-like faults. Since extension replaced the Mio-Pliocene compression sharing an almost parallel strike of the faults, the re-activation of pre-existing thrusts at depth is a likely scenario. In this paper, we show that a recently developed crustal tomographic model along with a complete catalogue of seismicity yield insights on seismotectonics and strongly support the hypothesis that, in the Apennines, normal faulting earthquakes often re-activate pre-existing thrusts.