Integrating temporal probability in landslide hazard evaluation towards the assessment of the economic risk at regional scale

Among the others, the number of disasters caused by natural hazards and the consequent economic losses have continuously increased during last decades. The risk management posed by the occurrence of natural hazards is challenging. Nowadays, the consequences of natural hazards are expected to worsen. The effect of climate change, as well as the interaction between natural hazards, environmental and human drivers make the consequent risks more complex to manage. The growing awareness of international community about disaster risk reduction is remarked by the ratification of international agreements, guidelines and global targets from local to global scales. However, hazardous phenomena, in particular recurring small-scale and slow-onset ones, have continued to significantly impact communities and their assets undermining the efforts to achieve a sustainable development. The present research work deals with the landslide hazard, which is one of the most frequent natural phenomena and a major threat to human safety and the environment. In future, since most of landslides are rainfall-triggered, an increase of their recurrence is expected in some areas because of the variation in the hydrological cycles and the increased frequency of extreme weather events. Moreover, the urbanization of steep and unstable slopes exposes more people and more assets to the negative impacts of landslide events. Landslide hazard and risk assessment, which are fundamental tools in the management of natural hazards, are a main subject of research since a long time. Several factors influence the choice of the methodological approach, such as: the purpose of the assessment, the landslide triggering factors, the types of elements at risk, the kind of involved mass movements, the available data and those to be collected, the scale of the study (i.e. individual, local, regional, national or global), and the available time. According to these factors, literature provides us many approaches to assess landslide hazard and risk. Whatever the method is, the key basis to start the analysis is a reliable landslide inventory, i.e. a detailed record of spatial and time characteristics of past landslides within a territory. Collecting data about past landslide occurrences may constitute a tedious procedure, inasmuch mass movements are generally isolated and localised events, which need to be mapped and described individually because of their diverse attributes. Most of the analyses aiming at assessing the landslide hazard and risk are limited to the spatial probability of occurrence. As the matter of fact, they do not include the assessment of the temporal probability of landsliding. The difficulty in determining the frequency of landslide events is mainly due to the lack of information about the date of occurrence within landslide databases. Thus, adding the temporal dimension to the hazard module results challenging. In the light of the above, this work aims at improving the current procedures concerning hazard and risk analysis by dealing with the landslide risk assessment of a hazard-prone area located in southern Italy (Daunia area, Apulia region). What clearly emerges from the overview about the current procedure to assess landslide hazard and risk within the study area is that there are many procedures for the assessment of landslide hazard and risk in many areas of the study area, inasmuch different Basin Authorities were competent. Nowadays, although they have been grouped from the administrative point of view in the District Basin Authority of the Southern Apennines, the landslide hazard and risk procedures have not been standardised yet. Therefore, an attempt has been made to develop a methodological approach to hazard and risk assessment, which was univocal for the same territory and took into account the spatial and temporal probability of landslide phenomena. The spatial probability assessment derives from a susceptibility assessment. It consists in a procedure that involves predisposing and triggering factors, which cause landslides. Regarding the temporal aspect, after counting the number of landslide events occurred within slope units, it has been quantified by applying a Poisson probability model. Since the available databases have scarce information about the temporal occurrence of landslide events, a new multi-temporal landslide inventory was necessary. The new inventory has been obtained by the analysis of paper documents collected by the Apulian administrative Difesa del Suolo office (namely, Soil Defense), which is in charge of planning structural mitigation interventions in the field of soil protection. From the analysis of the paper documents associated to the study area (the southern part of the Daunia area), 493 landslide events have been counted in the period 1998-2018. These events have complete information about spatial and temporal aspects, and thus they were useful to carry out the temporal probability assessment. The following landslide risk analysis consists in the combination of the spatial and temporal probability of landsliding, the areal extent of the elements at risk and their economic values. The outputs of landslide risk analysis result in economic values associated to each municipality located in the study area. Moreover, the risk value in monetary terms has been assessed for each slope unit within the municipality in the study area. The comparison between the economic values permits to rank the areas most-at-risk from an economic point of view. Moreover, the estimated economic risk per each municipality has been normalised by the corresponding areal extent, in order to avoid its influence on the risk assessment. After that, the results of risk assessment have been compared with the funds that concern mitigation measures. The results of such comparison could be used as a tool in the management of landslide risk at regional scale, guiding the choices of decision makers involved in the financing of mitigation measures.