Debris flow hazard and risk analysis at medium and local scale

Increased population, its wealth and activities lead to increased risk from landslide hazards in the mountain areas. The presented study deal with the analysis of debris flow hazard and risk at medium (regional) and local (site-specific) scale. Study took place in part of Valtellina Valley in Italian Central Alps. The first part of the study presents information about landslide hazard, risk and debris flow processes. It also shows historical information about natural disasters in the study area of Mountain Consortium of Municipalities of Valtellina di Tirano together with the results from a geo-database built for the study area. The database covers a period from 1600 till 2008 and it represents an important source of information for civil protection purposes. In the second part of the study, the debris flow susceptibility, hazard and risk analysis is performed at medium scale for the whole studied territory. Firstly, a new debris flow database was prepared and used as an input for susceptibility modelling together with the official sources. Susceptibility modelling was performed, using a bi-variate statistical technique Weights-of-Evidence and then the best performing susceptibility map was selected. Afterwards, spatial comparison of different susceptibility models was made, using advanced statistical techniques (Kappa Statistic, Principal Component Analysis, and Distance Weighted Entropy). Even though the modelling results show similar performance when assessed by standard evaluation techniques, their spatial pattern is very different. This may have serious outcomes in choosing the model which is most correctly delimiting the susceptible areas. Consequently, the results from the susceptibility modelling were coupled with temporal occurrence of debris flows in order to acquire initiation probabilities. These were subsequently used as inputs for runout modelling to obtain hazard maps at medium scale. Finally, medium scale risk maps were prepared by overlaying hazard maps with the elements at risk. Then, three risk maps were obtained. Two of the maps quantify economic risk from debris flows and the third qualitatively delimits total risk areas. Besides limitations and uncertainties within all steps of the analysis, an economic estimation of prospective consequences was done, being very important for the local stakeholders as well for the general public. Third part of the work presents two case studies related to quantitative risk assessment at local scale. Selvetta debris flow event, which happened in July 2008, was studied in the field obtaining information about the debris flow and building damage. The event was back-calculated and used for estimating synthetic vulnerability functions for buildings. In the second case study, located in Tresenda, hypothetical debris flow hazard was modelled for three return periods. Hazard scenarios were prepared and risk was quantified in economic terms using the vulnerability curves from the first case study of Selvetta. The results and the method applied for the quantification of risk reveal as very promising. As a consequence, it may be very interesting to apply this approach in different socio-economic and environmental settings in order to test its robustness. Presented thesis proposes a debris flow hazard and risk analyses approach at medium and local scale. Besides its feedbacks arising from available data, models, changing natural and socio-economic conditions and other intrinsic limitations, the study shows a possible approach applicable within integrated debris flow risk management framework.