The definition of threshold values within an early warning system for monitoring of an instability phenomenon in nearly real time represents a key step. Several approaches exist in literature, among which numerical modelling for the simulation of instability scenarios is the least common. The present paper illustrates how discontinuum numerical modelling can be used as a support tool for risk management with reference to a slope instability case study. The aim is to investigate the displacement rate that can anticipate the instability phenomena. To highlight the reliability of the models, back-analysis of a past instability event is carried out and a time-scaling coefficient is defined, and eventually validated, to match numerical results and monitored data. Based on such results, that reproduce the slope displacement trends from triggering to collapse, thresholds and related timespans to be adopted within the existing monitoring system are suggested.
Landslide risk management relies increasingly on the prompt protection against damages of the elements exposed to risk, such as buildings, infrastructures, industrial plants. In this framework, Early Warning Systems (EWS) and adoption of procedures are efficient tools that can be used to mitigate this risk by keeping away people from dangerous areas with a sufficient lead time in case of expectation of an imminent collapse. A reliable forecast of the collapse of landslides is however problematic since large rock slope deformations are very often characterized by a non-linear behavior. Not all rock slope instabilities obviously lead to rapid and catastrophic failures according to the classical evolutionary creep theory of landslides (Saito 1969, Varnes 1982, Cruden & Masoumzadeh 1987) but many of these diverge from the theoretical trend and remain characterized by slow or extremely slow continuous movements while still others can show intermittent behavior over long periods with acceleration and deceleration phases. Data collected by monitoring networks, in terms of displacement, velocity and acceleration can therefore provide useful indications about the short-term prediction of a slope failure (Rose & Hungr 2007) but, to manage and mitigate the risk for the exposed elements, a critical element needed is the definition of specific thresholds for relevant monitored quantities able to anticipate the occurrence of such phenomena.
