Morphotectonic analysis of the Lunigiana and Garfagnana grabens (northern Apennines, Italy): Implications for active normal faulting

This work integrates existing structural geology data with new detailed geomorphic analyses of the fluvial network to characterize active and potentially seismogenic faults bordering the Lunigiana and Garfagnana basins in the northern Apennines of Italy. These two basins are NW-SE-oriented asymmetric grabens, bordered by several normal faults with a poorly known, but probable recent slip history. Several strong earthquakes (M 5.0-6.5) have occurred in the area in the last millennium, demonstrating that this is one of the most seismically active areas of the northern Apennines. However, the lack of reliable instrumental data for strong earthquakes, generally low deformation rates, and poor exposures of faulted Quaternary sediments render the characterization of active, seismogenic faults problematic.Here, we quantify the relationships between faults and watershed-scale geomorphology using 10-m digital topography to extract channel and basin metrics, such as steepness, concavity, and stream length-gradient indices of modeled river longitudinal profiles. In particular, convex segments of longitudinal profiles (knickpoints) are investigated in the spatial context of suspected active faults. Several knickpoints arise locally from juxtaposed rock types of different erodibility; however, many others mapped along major normal faults have a clear tectonic origin. In fact, the height of the footwall knickpoints (the closest to the fault trace) varies along-strike the fault, increasing toward the fault center and tapering off toward the fault tips, mimicking the expected displacement profile of a fault. In these cases, we consider the knickpoint height as a proxy of the fault throw accumulated by the youngest fault activity, probably during the late Quaternary. The along-strike distribution of knickpoint heights helps in defining the likely segmentation pattern of the fault system. The identified active normal fault segments have lengths ranging from 9.5 to 28.5 km. The inferred late Quaternary throw rate ranges from 0.3 to 0.8 mm/a; however, the absence of any offset datable material limits our ability to assign precise numeric ages and rates of offset to the faulting.