Author(s)

F. Mazza, L. Longoni, M. Papini & D. Brambilla

Abstract

The high number of complex processes involved in soil erosion and sediment delivery make their understanding and reproduction a difficult task. Alpine valleys, characterized by high slopes, are particularly susceptible to severe soil erosion. Two of the main consequences are silting of water reservoirs and fostering of shallow landslides. In the last decades several models for the evaluation of sediment production and delivery have been proposed. Different approaches can be split into two main categories: bottom-up and top-down models. Bottom-up models are designed to reproduce the main physical processes involved in soil erosion; these methods are really complicated from a computational point of view. Instead top-down models, like the Gavrilovic one, reproduce the phenomenon at the basin scale with a low number of parameters. In this paper the authors present a hybrid Monte Carlo and possibilistic approach to the Gavrilovic method, designed to take into account uncertainties on input data. An example of application on a test basin situated in the Italian Alps is used to show the potential of the proposed method. The basin was split into sub-areas to reduce the subjectivity of the choice of empirical coefficients. A quantitative comparison between measures of average sediment yield and results obtained with the proposed method was performed. Keywords: soil erosion, silting, Monte Carlo methods, epistemic uncertainty. 1 Introduction Soil erosion and transport are two complex phenomena, acting on a wide range of scales, that are far from be fully understood and modeled, mainly because of the lack of knowledge about the physical mechanisms governing them. Many

Keywords

soil erosion, silting, Monte Carlo methods, epistemic uncertainty