PRELIMINARY RESULTS ON THE COMPARISON BETWEEN EMPIRICAL AND PHYSICALLY-BASED RAINFALL THRESHOLDS FOR SHALLOW LANDSLIDES OCCURRENCE

Rainfall induced shallow landslides are hazardous phenomena causing significant damages all over the world. Rainfall thresholds are the most used tool to predict the occurrence of such instabilities over large areas. Most of these thresholds are empirical, basing on past rainfall events triggered landslides. These present several limitations, due to the amount and the uncertainty of available data and neglecting antecedent soil hydrological conditions. Physically-based thresholds could take into account also for soil hydrological responses towards rainfalls. Thus, this work aimed to compare both these typologies of thresholds, reconstructed for a 250 km(2) area of northern Italian Apennines prone to shallow landsliding. Preliminary results of this research show the significant differences between thresholds reconstructed by means of different approaches. Empirical threshold estimates a lower triggering cumulated amount of rainfall than the values for physically-based thresholds, considering the same duration of rainfall event. Antecedent pore water pressure conditions have significant effects on reconstructed threshold, reducing the amount of rainfall leading to instabilities according to conditions approaching soil complete saturation. These preliminary results will be improved in order to create threshold useful for early warning strategies.