Search

Geophysical Research Abstracts, 20, ISSN:1029-7006, ISSN:1607-7962

The research work described in the current thesis is an effort to join the different scientific domains of soil science, applied geology and landslide modelling over very important landslide phenomena, namely flow-like landslides, which are one of the most dangerous natural hazard. In Campania (southern Italy), which is an Italian region strongly exposed to hydrogeological risk, soil scientists have assessed that a crucial landslide predisposing factor is the presence of Andosols. Such soils are typically very fertile but, unfortunately, at the same time, they are also very fragile and highly susceptible to land degradation processes, such as erosion and landslides. In fact, they have a peculiar set of chemical, physical, hydraulic and mineralogical properties that predispose the soil to the instability. Such previous studies suggested the existence of a pedological control on flow-like phenomena. Nonetheless, these soil studies were mainly developed by applying a descriptive approach, i.e. by studying soil properties. The use of the only descriptive approach is much unfortunate because limits the implementation of soil information towards more practical applications such as the use of dynamic landslide models finalized to future applications to predict these dangerous phenomena. In this framework, the general aims of this work are (i) to implement a dynamic model in order to predict the triggering of flow-like landslides in Swiss and Italian contexts and (ii) to evaluate its potential applications to other Italian study areas where soils showing andic features are settled. We implement a physically-based landslide hydromechanical triggering (LHT) model, which is publically available as software STEP-TRAMM, linking key hydrological processes with threshold-based mechanical interactions. The main difference respect to other models is that STEP-TRAMM incorporates progression of local failures in a chain reaction culminating into hazardous mass release. The model considers the soil cover as an unique and homogeneous layer having assigned mechanical and hydraulic properties. The soil depth is variable within the region of interest and spatially distributed according to a soil depth model which is implemented in the code. To pursue the objectives of this thesis, first of all we implement the selected landslide model to different contexts for both back-analysis and predicting future rainfall-triggered flow-like landslides. In detail, we carry out the following studies: 1) we select two Swiss catchments located at the foothill of the northern Alps and very close to each other, where important flow-like landslides were triggered by well-defined rainfall events. 2) Then, we move our attention to an Italian case study located within the city of Naples (Campania region), i.e. Camaldoli hill, where flow-like landslides involve soils showing andic properties. After these first applications, we make a preliminary study to evaluate the potential application of the same model to more complex areas. More specifically, we analyse soil depth and soil layering variability for different Italian contexts where complex geometry is found (and where soils involved in past landslides show andic features): 3) The first study area is located on the northern slope of Mt. Camposauro (Telesina Valley - Campania region), where flow-like phenomena often occur in the colluvium, involving Andosols, and spread downslope involving the inhabited districts. 4) In the end, we carry out a national scale study to evaluate morphological, chemical, physical and hydraulic properties of 12 Italian soils located in detachment areas of past flow-like landslides. Our results demonstrate the importance of a pedological approach finalized to evaluate key soil properties, intended as predisposing factors of flow-like mass movements. In particular, we found the presence of andic soils where important or minor Italian flow-like landslides occurred. Due to the unique set of properties, andic soils are considered to be highly exposed to landslide phenomena and other land degradation processes. Such findings shed new light on the similarity of the materials involved by flow-like landslides (in Italy), suggesting a pedological control on the flow-like hazard. In fact, soil properties have to be considered a crucial predisposing factor (together with the other commonly recognized hydraulic and geotechnical properties) because they could dramatically increase the susceptibility of a soil to be involved in flow-like landslides. Among dynamic landslide models, we chose a LHT model (namely, STEP-TRAMM) which gave satisfying performances when implemented on homogeneous soils, indicating the location of the first rupture within the soil and the position and the area of each simulated landslide. Despite the limited resolution of the numerical grid that we set, the model was able to properly detect most susceptible landslide areas for the Napf, Pogliaschina and Camaldoli hill catchments. Such results were strongly encouraging even if a bigger effort would be required by considering that, in many Italian environments, landslides usually occur where layered soils occur. In fact, our experience demonstrated that all over Italy there are different environmental and geological contexts where catastrophic or minor flow-like landslides involve layered soils. For such cases, the layering, which is typical of andic soils, is a not negligible landslide predisposing factor. In fact, the contact between different soil horizons, showing different properties (e.g. soil texture, saturated water retention, etc.), represents a physical and hydraulic discontinuity. In this sense, the layering creates weak surfaces within the entire soil profile, where an increase of positive pore pressure can develop and trigger flow-like landslides. Furthermore, the study conducted in Telesina Valley showed that soil depth can importantly vary not only with the depth, but also in the space as a function of the different geomorphological contexts that can be found over the same hillslope.

Debris flows are known to be very dangerous hazards that can produce fatalities and severe damages. In the Campania region they involve Andosols with a unique set of morphological, chemical, physical, and hydrological properties that give them high fertility but, at the same time, high fragility as concerns land degradation processes, such as erosion and landslides. Because of the shallow character of some debris flows (e.g. Campania), they can strongly involve and evolve in and over the soil cover. Therefore, a better understanding of the soil properties is crucial in an integrated approach finalized to evaluate the landslide susceptibility of mountain ecosystems. The aim of this work is to investigate the soil (chemical, physical, and hydrological) properties of the detachment areas of eight debris flows occurred in non-volcanic mountain ecosystems (NVME) of Italy in the last 70 years. In this framework, we will mainly focus on the assessment of the soil andic properties, in order to verify if in the studied landscapes the debris flows involve soils characterised by these properties, as it happens in the Campania region.The analyzed events occurred on moderate to high slope gradient (ranging from 25° and 50°) and at different altitudes. Despite the lithological heterogeneity of the bedrock, the analyzed soils show remarkably similar features. Specifically, soils are generally deep and characterized by well developed topsoils, demonstrating that investigated sites were not subjected to intense erosion, which would have been expected considering the slope gradient. Moreover, these soils are markedly chemically and physically fertile, as indicated by the high organic carbon content and the prevalent loamy texture. As concerns the soil type, the preliminary analyses show that the investigated non-volcanic soils have andic and vitric features (except for Ceriana).These findings shed new light on the similarity of the materials involved by debris flows both in volcanic and non-volcanic mountain ecosystems of Italy, suggesting a pedological control on the debris flow hazard.