Modeling debris flows in volcanic terrains for hazard mapping: the case study of Ischia Island (Italy)

Abstract

The Island of Ischia is a densely populated, active volcanic island located in the Tyrrhenian Sea, approximately 30 km WSW from the city of Naples in Southern Italy. The Island is a debris-flow prone area due to its steep slopes covered by loose volcanic lithologies, and the whole territory is vulnerable to such phenomena due to an unregulated urbanization. On April 30th 2006, following several hours of rainfall, four soil slips were triggered on the slopes of Mt. Vezzi (about 400 m a.s.l.) in the SE portion of the island. The soil slips changed quickly into debris flows that reached the inhabited at the foot of the hill. In spite of their limited size, the landslides caused four victims and destroyed several buildings, forcing the evacuation of 250 inhabitants. This paper presents the analysis of the triggering and propagation phase of the phenomena. In particular, to model the triggering conditions, a finite element analysis was used to reconstruct the fluctuations in pore water pressure during the storm in transient conditions. The limit equilibrium (Morgenstern and Price, 1965) slope-stability method was then applied using the temporal pore water pressure distributions derived from the seepage analysis. The dynamic modeling of the propagation phase was carried out by means of two dynamic codes DAN-W and FLO2D, with the aim of evaluating the residual hazard linked to other potential debris flows recognized in the same area. The model calibration was based on the thickness and areal extension of the deposits, on flow velocity and runout. The results have been subsequently compared to adopt a combined approach to the modeling. Once the DAN-W and FLO2D models satisfactorily reproduced the 30th April events, the simulations were extended to a larger area, whose susceptibility to future landslide events has been determined through a detailed geomorphological survey and a following GIS analysis. Several scenarios related to these potential events were used to estimate the inundation areas, flow velocities, and deposit thicknesses.