Initiation mechanisms of mudflow landslide studied by 2D soil water balance

This work deals with an integrated approach aiming to study the very complex phenomena of the landslide initiation mechanism of mudflow induced by rainfall events. The study has been conducted on the soils interested by the catastrophic landslide events of the 5-6th May 1998 in Sarno and Quindici (Naples district, South Italy). On the base of a preliminary soil survey, five soil profiles have been selected along the detachment crowns of some landslides. These profiles have been characterized on the base of their morphology, genesis and hydrology. All these soils are very fertile and very deep (>200 cm) Andosols rich in easily weatherable minerals; the soils have a complex pedological history with two or more pedogenetic cycles. The main objective of the study was the analysis of the influence produced by soil discontinuities (i.e. roads, cliffs, etc.) over a typical slope in terms of landslide initiation mechanisms. The physical measurements conducted in one of the five profiles have been used in the framework of a 2D water simulation model applied on a soil system of a typical slope of the study area. Soil water retention and hydraulic conductivity functions were measured on soil cores applying the evaporation method. Model sensitivity analysis have shown that a certain degree of uncertainty related with some unavailable input parameters (crop factor, anisotropy in the hydraulic conductivity) can well be tolerated without altering the main results dealing with the water storage. The resulting processes and results produced by a discontinuity of the soil cover over the slope have been simulated using as model output the total weight of a soil column taken in upslope position of the soil discontinuity. In such case, it has been observed, in comparison with the an undisturbed slope (no discontinuities but the same soil features), an increase of 30% of water storage. This produce a soil overload with tangential tension over the slope of the same order of magnitude of the peak resistance measured by direct shear tests. Despite the many simplifications and approximations (i.e. lack of geotechnical analysis), the proposed integrated (pedology, hydrology, modelling) approach demonstrate an important opportunity in terms of quantitification of the consequences of some natural and/or anthropic changes of the continuities of the soil cover over slopes. This finding are especially important in volcanic areas having Andosols as soil cover. The authors are convinced that such approach can much ameliorate the estimation of landslide initiation process involving soils and especially volcanic soils.