Quantitative landslide risk assessment at the river basin scale

Landslides and mass movements in general are very common in Italy, especially along the main mountain chains such as the Alps and the Apennines. The study area is no exception to this rule, as it is strongly subjected to mass movements. The study area is the Arno river basin, which is located in northern Apennines, Italy, and has an extension of 9116 km2. A new landslide inventory of the whole area was carried out using conventional (aerial-photo interpretation and field surveys) and non-conventional methods such as remote sensing techniques including DInSAR and PS-InSAR, (Farina et al. 2006). The great majority of the mapped mass movements are rotational slides (75%), solifluctions and other shallow slow movements (17%) and flows (5%), while rapid flows and falls seem less frequent everywhere within the basin. This research is aimed at assessing landslide hazard and risk at basin scale. The final goal is to create a dynamic tool, managed in a GIS environment, useful for landslide risk pre-disaster planning and management. The assessment of landslide hazard in terms of probability of occurrence in a given time, based for mapped landslides on direct and indirect observations of the state of activity and recurrence time, has been extended to landslide-free areas through the application of statistical methods implemented in an artificial neural network (ANN). On the basis of the more common landslides in the Arno river basin and the results of the univariate statistical analysis five preparatory factors were selected: slope angle, lithology, profile curvature, land cover and upslope contributing area. The definition of position, typology and characteristics of the elements at risk has been carried out with two different methodologies: i) buildings and infrastructures were directly extracted from digital terrain cartography at the 1:10,000 scale, whilst ii) non-urban land use was identified and mapped based on an updated and improved CORINE land cover map at the 1:50,000 scale. The definition of the exposure for each type of element at risk was based on their presumed asset and income values. Landslide vulnerability, defined as the degree of lost of elements at risk due to a landslide of a settled intensity, usually expressed as a value ranging from 0 to 1, was estimated on the basis of the typology and economic and social relevance of the elements at risk. Landslide intensity, usually defined as proportional to kinetic energy, was obtained considering landslide typology as a proxy for expected velocity. The landslide risk was assessed both in a qualitative and quantitative way. In the former case contingency matrices were used to intersect hazard classes with vulnerability and exposure classes, while in the second case quantitative assessment of risk was carried out through the application of the risk equation, therefore applying the product of the numerical values of hazard, vulnerability and exposure (Cruden and Fell 1997).