Shallow Landslides Risk Mitigation by Early Warning: The Sarno Case

摘要

Landslide risk mitigation is a societal pressing need in many countries; early warning system supported by monitoring and predictions represent effective risk reduction measures, provided that warning thresholds are properly set. Mapping of landslide risk is extremely complex: slope movements have a wide range of velocity, size and run-out, thus their magnitude and impact on exposed goods can be either very low or very high, depending on site conditions, materials involved and triggering factors. Rainfall is accepted as a major triggering factor in many types of slope movement, including rapid, shallow soil slips and, episodically, deeper landslides. Early warning can be defined as the entirety of actions to take during the lead-time, namely the time interval between the moment of the event prediction and the moment of the landslide impact. In this context, the improvement of analysis methods able to increase this time is essential. The paper presents a joint application of the FLaIR Hydrological Model of rainfallinduced landslide triggering (Sirangelo and Versace Proceedings of XXIII Convegno Nazionale di Idraulica e Costruzioni Idrauliche, Firenze, III, pp D361–D373, (1992)) and an event-based point rainfall stochastic model (Capparelli et al. XXXIII IAHR Congress, Vancouver, Aug 2009, ISBN 978-94-90365-01-1, pp 6812–6819, (2009); Greco et al. Early warning of rainfall-induced landslides based on empirical mobility function predictor. Submitted for publication in Natural Hazards, (2011)). The first one identifies the landslide triggering conditions by defining a mobility function Y(t), obtained through the convolution of infiltrated rainfalls and a transfer function c(t), the second one is a stochastic model of the external structure of point rainfall height series, allowing to predict in real time the residual duration and rainfall height of a partially observed rain storm. The combination of the two models improves the effectiveness of an early warning system, since it allows to gain larger lead times. The potentiality of this approach is shown with regards to the slope of Pizzo d’Alvano (Campania region- South Italy), where mudslides occurred on May 5th 1998, severely hitting the town of Sarno.