Sub-project 2: Landslide zonation, hazard and vulnerability assessment

An essential step in stabilization and land use planning in areas prone to landslides is mapping of landslide hazard. This work package investigates the elements of landslide zonation, including collection of data to produce inventory maps, improving the maps with the use of high-resolution images, embedding the maps in a geo-database containing other pertinent data, the probabilistic aspects of triggering mechanism (rainfall and earthquake), and uncertainties in the soil strength parameters.

The objectives of this Sub-Project are thus to (i) apply advanced landslide zonation methods to focus areas in selected landslide prone regions in Europe and develop and implement new models of landslide mapping, (ii) evaluate the technical and economic efficiency of conventional and innovative stabilization methods by numerical modelling of well-documented slides and (iii) develop tools for disaster planning and mitigation policies. The Sub-Project is coordinated by NGI and features also the participation of VCE, SGI-SW, UNEW and UNIMIB.

Disseminated material related to this Sub-Project can be found by clicking each of the links below:

Deliverables

Presentations 

Reports

Publications

Events

Meetings

A summary of each of the tasks involved in this Sub-Project is provided below:

Task 1.2.1: Application of mapping inventory to landslide hazard zonation 

Slope movements are caused by changes of external conditions, such as earthquake and/or heavy rainfall. When the magnitude of the anticipated slide displacement under these conditions is excessive, mitigation/stabilisation measures are required. One approach for developing landslide hazard maps is to utilise and analyse the inventory of past landslides in the region of interest.

The geodatabase for the 600 km² region in the Central Italian Alps developed in Task 1.1.2 of Sub-Project 1 will be analysed. Spatial frequency, typology and size of the mapped landslides will be analysed through multivariate statistical techniques to produce landslide hazard maps. This will involve application of a probabilistic approach based on Bayes Theorem to calculate the probability of occurrence of slope instability in the investigation area, as well as examination of the suitability of the method to assess the influence of rainfall, land use and vegetation cover on the landslide occurrence. Different future scenarios could be simulated by changing the spatial distribution of parameters such as land use and vegetation cover.

The comparison of the produced hazard maps with the landslide inventory will allow for evaluating the possible increase in the level of information and especially the reduction of mapping and subjective errors. The final inventory map will be embedded in a geodatabase to be integrated with existing in-situ monitoring data to create a tool that can suggest the best planning and mitigation approaches to local and regional administrators, with and without technical background.

Task 1.2.2: Probabilistic landslide hazard zonation techniques and vulnerability assessment

Having developed and implemented the methods described in Task 1.2.1, this work package aims at extending the techniques that are well developed for special types of landslides (e.g. quick-clay slides in Scandinavia), other types of geological setting (e.g. deep sea environment), or other types of hazards (e.g. snow avalanche), and apply them to landslide hazard mapping. The probability of a given location being affected by a landslide may be considered as a product of two probabilities, namely the probability of slide initiation times the probability of the run-out distance of the slide going past the point of interest. The former is closely related to the occurrence probability of the triggering mechanism and uncertainties in the soil strength parameters. The latter is related to dynamics of the sliding mass/debris flow, which because of the complexity of the physical phenomenon, is often estimated from statistical models.

The deliverable for this work task is a quantitative, probability-based landslide hazard zonation technique.  The demarcation line between the 'safe', 'potentially unsafe' and 'unsafe' areas will be based on the annual probability of being exposed to a landslide. The preceding probabilistic zonation technique will be tested and validated in two regions in Scandinavia, adjusted and re-calibrated if necessary. In addition, at one of the sites, the results will be compared with those of a deterministic hazard zonation method.  The hazard zonation procedure will also form the basis for developing landslide risk maps by combining the hazard maps with vulnerability of exposed population/facilities/infrastructure, and with the associated costs.

The task will also include review and critical assessment of GIS-based methodologies for landslide susceptibility mapping (statistical/probabilistic approaches, heuristic approaches, etc.), and literature survey on assessment of vulnerability to landslides.

Task 1.2.3: Tool for detailed hazard zonation of areas affected by debris flow

The major objective of this task in the context of landslide hazard zonation is to enhance our capability for predicting the impact of debris flows at the scale of a river basin, as a function of climate, land use and rainfall return period. Until recently, techniques for hazard assessment tended to be approximate (because of lack of data) and gave only qualitative or relative estimates of hazard. Also, they could not be used predicatively to indicate the effects of possible future changes in land use and climate. Recently, though, through various research initiatives a range of modelling techniques, applicable at scales from the hill-slope to the region, have been developed and applied.

The work will consist of collecting datasets and climate scenarios for the focus areas in the central Spanish Pyrenees and the Italian preAlps. The climate scenarios will be refined and rainfall events of different return periods (allowing for duration and intensity) will be generated for both current and possible future climatic conditions. The model will be run to give debris flow incidence and sediment yield for the different events. The relationship between rainfall return period and the debris flow incidence, spatial distribution and sediment yield will then be quantified, for example in rainfall duration-intensity thresholds for different levels of debris flow. The above landslide model can be used as a predictive tool to examine the occurrence of spatially distributed shallow landslides and debris flows at the scale of a river basin (up to 500 km² as opposed to most other landslide models, which typically are limited to scales of a few square kilometres).