For the Italian test sites the technique of the Permanent Scatterers (PS technique) has been used. It has been developed and patented by the Politecnico di Milano (Italy) and improved by Tele-Rilevamento Europa. This methodology allows to estimate the measure and the displacement velocity of some points on the ground, on millimetric scale.
The methodology applied for the Swiss test sites is the multi-interferometry SAR: Interferometric Point Target Analysis (IPTA), developed by Gamma R-S. This methodology allows to evaluate the motion rate of some points within the area of interest.
The integration of PS and IPTA data in a GIS environment allows to make a connection between the rate of displacement in the risk areas and the pertinent structural interventions. 

The analysis of the optical images with high spatial resolution, carried out by SPACEBEL, allows to the automatic extraction of features related to landslide presence. 

The Synthetic Aperture Radar (SAR) are space-borne instruments that emit electromagnetic radiation and then record the strength and time delay of the returning signal to produce images of the ground. The radar images are defined by the amplitude  and the phase of the  signal backscattered from the observed scene. 

The main application of SAR Interferometry is the estimation of topographic height from the differential range measured by two radar antennas looking at the same surface. The phase difference is sensitive to both viewing geometry and the height of the point z above the reference surface. If viewing geometry is controllable or at least knowable to sufficient accuracy, then the topography can be inferred from the phase measurement to a precision of several meters.

The Differential SAR Interferometry is an advanced interferometric technique that can usefully be applied to map surface displacements as well as those associated with landslides. The interferometric phase is sensitive to both surface topography and coherent displacement in between the acquisitions of an image pair. The basic idea of differential interferometric processing is to separate the two effects, allowing, in particular, to retrieve a differential displacement map. This goal is achieved by subtracting the topography related phase.

An important problem for SAR Interferometry is the decorrelation. 
The temporal decorrelation occurs from changes in the scatterer characteristics. In particular for the alpine territory, which is characterized by low vegetation, the SAR images show a relatively high coherence only during the snow-free season. But also for built-up areas at relatively high altitudes (i.e. above 1000 a.s.l.), wet snow cover may be the cause of phase decorrelation. The spatial decorrelation prevents interpretation of interferometric phases for extended targets in pairs with long baselines. Other limitations to the spatial coverage arise from layover and shadowing caused by the very rugged topography and by vegetated areas close to built-up areas. In addition, there is a privileged slope direction, namely that facing away from the SAR look vector, where the technique is better suited for detection and monitoring of displacements. 
All these problems make the SAR-based technique only suitable to limited areas on the ground and to selected periods of the year. In order to limit the effects of spatial and temporal decorrelation the differential pairs with short baseline (about 100m) will be chosen between SAR data acquired in the snow free period. But in this way the number of available differential pairs for the product generation will be reduced.

Another drawback of SAR Interferometry is the Phase Unwrapping of differential interferograms,  this is a necessary step to retrieve a quantitative measure of displacement. In rugged terrain, and for complicated displacement fields it is often a difficult task.

top

 

The PS technique overcomes the main limits of conventional approaches to surface deformation detection, allowing to identify single radar benchmarks (called Permanent Scatterers) where very precise displacement measurements can be carried out. These objects usually correspond to man-made structures such as buildings, dams, penstocks, antennas, pylons as well as stable natural reflectors (i.e. exposed rocks) present on the Earth surface.
The PS technique is a powerful tool to detect and monitor different geophysical phenomena such as: subsidence, landslides, seismic faults etc. and even to verify individual building stability.  

The Standard Permanent Scatterers Analysis (SPSA) is developed to investigate a large portion of territory (minimum 100 km²) in order to identify instable area, as well as the area affected by landslides, that deserve further detailed studies. The standard processing requires a limited amount of time and a limited interaction of an operator with a  medium training level. 
The standard analysis allows to:

• identify radar benchmarks (PS);
• compute their geographical coordinates;
• estimate their average velocity on different time intervals (e.g., the whole period and the last 2 years of acquisitions, in order to highlight possible variations in the velocity values).   

 

The Advanced Permanent Scatterers Analysis (APSA) is suitable to study limited areas of interest (up to 20 km²) in which is needed to maximize the information contents. This technique requires skilled technical staff, able to identify PS characterized by very high coherence values where phase unwrapping errors are very unlikely.
The advanced technique permits to:

• increase the PS density;
• estimate the full time series of all radar benchmarks of interest.

Before applying the APSA, a preliminary geological, geomorphological and geotechnical characterization must be performed, by means of field surveys, analysis of technical reports from local administrations and the study of monitoring data. This step aims to support the interferometric analysis and the use of kinematic models for the selected mass movements.    

top

 

The Interferometric Point Target Analysis (IPTA) is a method to exploit the temporal and spatial characteristics of interferometric signatures collected from point targets to accurately map surface average deformation rates, deformation histories, terrain heights, and relative atmospheric path delays.
The advantage of using point targets is that these do not exhibit geometric decorrelation such as distributed targets, permitting a more complete use of the data as even pairs with very long baselines (longer than the critical baseline) can be interpreted. This in turn  it results in an improved accuracies and temporal coverage. Consequently, more observations are available permitting reduction of errors resulting from the atmospheric path delay and leading to better temporal coverage. An important element of the IPTA is the analysis across the data stack, respectively in the time dimension.

top

 

In a first stage, the images have to be orthorectified using the topographical maps and DEM of the region.
The main interpretation technique applicable to a single optical satellite image is visual interpretation. Different features detectable on optical imageries indicate newly occurred landslides in the region.
Morphological features related to landslides, such as scarps, landslide crowns, terraces, slump blocks, concave surfaces of rupture, drainage lines, the slide body and the tong are to be detected on the image. Anomalous patterns and vegetation differences (disarranged drainage systems, tilted trees, etc.) characterizing damaged terrains are other indicators of landslide occurrence on optical images.
Other spatial aspects known as triggers for slope failure like vegetation, land use and slopes (derived from the DEM) can be added to the image to redirect the visual interpretation. 
When multi temporal series of optical satellite images are available for the interpretation, change detection is very useful to indicate landslide occurrence.
Further on, some enhancement techniques like masking of irrelevant zones (e.g. flat zones) and image sharpening can be applied to improve the interpretation.

The geological analysis consists in an integration in a GIS environment of the data coming from the interferometric techniques with the ground based information. 

For the first type of product an update of the existing landslide map will be realized through a compliance analyisis in a GIS environment of the points target, the existing landslides map and the geologic lineaments. The observation will be supported also by visual interpretation of the SPOT5 images rendered on DEM and by aerial photo interpretation.

For the reduced scale monitoring UNIFI and Geotest utilize in-situ measurements,aerial photos, points target data, high accuracy DEM, technical reports and geomorphologic map as input data. For the Italian case VHR images represent a further important input data. As first step bibliographic information will be collected and reviewed, then a map overlaying of the points target on the geomorphologic map will be performed in order to define the areas affected by movement and define zones with different deformation rates. Discrepancies will be analysed by means of interpretation of VHR images and aerial-photos and field surveys. As final step comparison and integration of in-situ and EO measurements will be fulfilled. The key point of this analysis is the geological modelling: analysis of monitoring data with reference to the geological features of the area, in order to define the landslide typology, the distribution and the possible scenario of evolution. As a result the geological team will produce a monographic report which includes: general information, geological/geomorphological features, monitoring data (integration of traditional and EO-derived) and interpretation of the deformation field.

For the realization of the Landslide Susceptibility/Hazard Mapping the first step consists in the identification of landslide prone areas. Homogeneous terrain units have to be defined by means of intersection within a GIS environment of the main instability factors: lithology, slope, land-use, upslope contributing area and profile curvature.
In fact, landslide hazard evaluation is connected to the geological characteristics, the typology of the slope instability problems interesting the studied area and the observation scale requested for the mapping. For a complete hazard assessment the following main steps are necessary:

• Spatial prediction: the forecasting of where, within a given area, landslides are likely to occur;

• Temporal prediction: the forecasting of when landslides are to occur in specified slopes;

• Type prediction: that is the forecasting of what type of landslide is likely to occur;

• Magnitude prediction: it consists in the forecasting of the velocity, dimension and energy of the landslide;

• Evolution prediction: it consists in the forecasting of how far forward the landslide will reach and how far back ward the landslide may retrogress.

The data necessary for the implementation of these steps could be integrated by using information obtained from satellite imagery, such as SAR data and VHR images. In particular, with reference to the above-described steps, it is possible to integrate EO data within a landslide hazard assessment procedures mainly in relation with the spatial prediction and the temporal prediction.
Spatial prediction: it consists in the assessment of a relative hazard of a slope with respect to the others. It does not signify the probability of occurrence from an absolute point of view or in temporal sense. The relative hazard classes can be provided by using different criteria, such as the analysis of landslide inventory maps, heuristic methods, statistical analysis or neural networks analysis.
Temporal prediction: it consists in the assessment of the probability of landslide occurrence in a given time interval and it provides an absolute hazard that could be expressed in terms of annual probability values, return periods or nominal scales.
The main approaches to obtain temporal prevision are:
-the analyses of landslide temporal series: this is one of the most effective tools to obtain information about the return periods of a landslide. Normally reactivation events can be dated by using documents or population evidences. 
-the analysis of causes temporal series: it consists in the analysis of the triggering factors connected with the landslide occurrence, trying to understand the type of correlation between these and landslides occurrence. In this way, after the definition of a threshold, it is possible to evaluate the return periods of the landslides by using the analysis of the measured factors values. The main triggering factors connected with mass movements are rainfalls, earthquakes, erosion and human activity.
Due to the lack of a general and scientifically proved methodology for landslide hazard assessment, taken as a reference by the national legal mandatory of the landslide risk management within the legislation framework of the countries involved in the SLAM project, for the Italian End Users the P.A.I. Plan Document (i.e. Piano per l’Assetto Idrogeologico) has been chosen as the reference base to implement this type of product. This document, according to the Italian Law 183/1989 and the Law 493/1993, contains the technical information regarding the optimal land use for very high, high, moderate and low risk areas of any given basin. The risk area classification is produced by taken into account three factors: the probability and the intensity of the event, the value of the elements exposed to risk and the vulnerability of the elements under the risk. The first factor represents the hazard. 
The statistical analysis of the homogeneous terrain units previous identified and the spatial distribution of landslides (as per the Landslide Motion Survey product) through the use of different techniques (e.g. multivariate statistics and neural networks analysis) will allow us to define the hazard levels from H0 to H3. In this way only landslide prone areas will be mapped.
As second step the analysis of temporal factor consists in a reclassification of the mapped landslide areas in the 3 hazard levels H2, H3, H4 by taking into account the state of activity of the inventory map (Landslide Motion Survey), the displacement temporal series provided by the monitoring data (Landslide Displacement Monitoring, in-situ data) and the historical information (i. e. AVI, SCAI and IFFI databases). In particular such classification will be performed by evaluating the return periods of each mass movements. The H4 class will be defined with reference to the 2 years return period, the H3 class with reference to the 10 years return period and the H2 to the 100 years return period. Such temporal limits have been identified by analysing the most widespread typologies of landslides in the Arno River Basin.
Finally the landslide hazard mapping is performed overlaying the classification based on the temporal factor on the landslide prone areas zonation. Such consequent classification in 4 hazard classes plus a hazardless area represents the final landslide hazard map to be provided to the End Users. The final hazard map will consist of an ArcMap geodatabase.