EXPERIMENTAL AND NUMERICAL EVALUATION OF COMPLEX SITE EFFECTS IN ARQUATA DEL TRONTO AFTER THE 2016 CENTRAL ITALY EARTHQUAKE

This work deals with the experimental and numerical evaluation of the local seismic response of Arquata del Tronto area (Marche region, Central Italy), severely struck by the Mw 6.0 August 24th 2016 earthquake. The Arquata del Tronto village and the surrounding hamlets of Borgo and San Francesco are located about 9 km NE of the August 24th 2016 mainshock epicentre. In detail, Arquata main village arises above an elongated WNW-ESE-trending ridge of the central Apennines thrust-belt (Central Italy), at elevations about 170 m higher than the underlying alluvial valleys where Borgo and San Francesco are built on. Despite their proximity (less than 500 meters), Arquata del Tronto, Borgo and San Francesco reported a quite different damage distribution after the August 2016 mainshock, suggesting that the seismic response of the area may be controlled by site effects. In order to explore this hypothesis, we evaluated the 2D numerical local seismic response along three representative geological cross-sections crossing Arquata del Tronto, San Francesco and Borgo; additional 1D analyses were carried out at strategic points along the cross-sections in order to explore the 2D physical phenomena governing the local response. Geomechanical properties of lithotypes were deduced by in situ tests. The satisfactory agreement between numerical amplification functions in linear range and experimental amplification functions obtained by the Generalized Inversion Technique (GIT) applied to a large number of aftershocks confirms the substantial reliability of the subsoil models. Numerical analyses representative of the 2016 mainshock were carried out and processed in terms of peak and integral ground motion parameters. A comparison with the observed damage pattern was then undertaken in order to provide general implications about site response and seismic microzonation in similar geological and morphological settings.