The root reinforcement in a distributed slope stability model: effects on regional-scale simulations

The shallow landslides are hazardous mass movements commonly triggered by intense rainfall. The hazardousness of these events is mainly due to their common evolution in rapid mass movements as debris avalanches and flows and to the frequently occurring in the form of clusters of events. Because of their characteristics, the forecasting is a particularly valuable tool to protect people and infrastructures from this kind of landslide events. The presence of vegetation on hillslopes significantly reduces the slopes susceptibility to the shallow landslides, and the stabilising action is mainly due to the reinforcement of the soil by the roots. The spatial variation of the root reinforcement should be therefore considered in distributed slope stability analyses. However, the natural variability of the parameter makes it challenging to insert the root reinforcement into the models. Many approaches to the problem were tested, but nowadays there are still lacking a distributed slope stability model capable of very quick processing in which the root reinforcement is considered and an approach to estimate the root cohesion at the regional scale that it has been tested in very wide areas and for long period-simulations. In this study, we present the effect of the root cohesion on slope stability simulations at the regional scale obtained using a physically-based distributed slope stability model, the HIRESSS (HIgh REsolution Slope Stability Simulator). The HIRESSS model was selected for the purposes, being capable of rapid processing even in wide areas thanks to the parallel structure of its code. The simulator was modified to insert the root reinforcement among the geotechnical parameters considered to computing the factor of safety in probabilistic terms, and for this purpose a commonly adopted model for the root cohesion was chosen. To build a map of the root cohesion for the study areas, the distribution of plant species in the area was obtained from vegetation distribution map and in situ surveys, then a value of root cohesion and a range of variation was defined for each plant species based on the most recent literature in this field, finally, to reproduce the natural variability, the root reinforcement was treated as variable in Monte Carlo simulations, as well as the other geotechnical parameters. The results of the simulations for the study areas were processed and analysed in order to evaluate the effect of the root cohesion on the failure probabilities and the adopted approach to estimate the root cohesion at the regional scale. The comparative analyses carried out on the results of the simulations performed inserting or not the root reinforcement brought out little differences between the two from the point of view the failure probabilities, particularly when the saturated conditions of the soil are reached. 10 Based on the findings of this research, it is considered that a root cohesion model different to the one adopted is preferable in the context of the shallow landslides, in applications in which working with failure probabilities (instead of factor of safety values) is desirable, and in areas similar to the ones of the study.