Kinematic model of active extension across the Umbria-Marche Apennines from GPS measurements: fault slip-rates and interseismic coupling of the Alto Tiberina low-angle normal fault

Abstract

The Umbria-Marche Apennines are characterized mainly by SW-NE oriented extensional deformation and most of major historical and instrumental earthquakes occurred mainly on the western side of chain, bounded by west-dipping buried high-angle normal faults. Recent studies about the northernmost part of Umbria-Marche region show seismic and tectonic activity on correspondence of the east-dipping Alto Tiberina (AT) low-angle normal fault (LANF), which is widely documented by geological data and deep seismic reflection profiles. In this area which of the known fault systems play a major role in accommodating the extension, and which are the modes (seismic VS aseismic deformation) this extension is taken up, is still a debated topic. During last years on Umbria-Marche Apennines close to Gubbio fault (GuF) a dense network of continuous GPS stations, belonging to the RING-INGV network, has been installed, improving significantly the spatial resolution of the detectable geodetic gradients. We used a self-consistent kinematic block modeling to study this sector of the Umbria-Marche Apennines, in order to understand which fault system is accommodating the tectonic extension. We found that both fault systems, i.e. the Alto Tiberina LANF and the antithetic high-angle normal faults, are needed to better reproduce the nearfield GPS velocities, obtaining kinematic agreement with geological slip-rates. Moreover we parameterized the ATF fault as a, more realistic, curved surface to infer the distribution of interseismic coupling (IC), which is validated by numerous resolution tests. The obtained IC distribution shows a correlation between relocated microseismicity and uncoupled patches attributed to aseismic creeping behavior, which could be explained by the presence of fluid overpressure. Otherwise this correlation has been verified with a very small quantity of events (almost 400) and it might be of interest to evaluate this correlation with future available data.