The hazards associated with landslides result in complex and changing landscapes whose evolution must be mapped and monitored in detail to forecast future behaviour. Italy is one of the European countries with the highest landslide hazard, and 13% of the territory is prone to landslides (Nadim et al., 2006). The Civil Protection Authorities (CPAs), who are pointed to prevent landslide hazards and to manage emergencies deriving from potentially damaging events, need a monitoring strategy that uses reliable data acquired in a short period of time over wide areas (at times covering entire regions) for use in forecasting, prevention and emergency phases. U-Geohaz project is designed to set up useful products for Civil Protection user community at EU level (operating at different administrative and organizational levels, from local to continental), based on radar data acquired by the Sentinel-1 (S1) satellite constellation (S1A and S1B). Radar interferometric techniques, based on S1 wide areas acquisitions (Interferometric Wide swath), allow a continuous regional to local scale monitoring, with a 6 days temporal sampling of 6 days A list of realistic user need for WP1 products, to make this products suitable to be integrated also in a different national contexts.
In Italy, detailed records of landslides do exist, such as the IFFI (Italian Landslide Inventory, Inventario dei Fenomeni Franosi in Italia), which mapped over 490,000 landslides covering an area of 21,200 km 2 , equal to 7% of the total surface (Trigila et al. 2014).
Find here de related deliverbles: WP2 Deliverables
Sentinel-1 processing: Deformation Activity Map & Active Deformation Areas map
This part of the WP2 is dedicated at the generation of the Deformation Activity Map (DAM) and the derived Active Deformation Areas (ADA) map, based on the Sentinel-1 (S-1) data processing. The ADA map is the main input of the successive analysis focalized at supporting landslides risk management activities, early warnings and alerts. Until now, two versions of the results have been delivered.
The S-1 data processing has been done by using and tuning the software tools developed by CTTC in the framework of the European Project Safety ( http://safety.cttc.cat/).
Valle d'Aosta presents difficult characteristics for what concerns the radar response and processing. The most significant factor is represented by the snow coverage during the colder seasons, which strongly affects the coherence and the spatial sampling of the results, mainly over the higher altitudes. In the pictures below an example of how the coherence of the interferograms strongly variate in space and time: A) the area of study; B) an example of high coherence of a 6-days interferogram during a no-snow period (June), it is also evident a strong atmospheric component that follows the highly variable topographic relief; C) an example of low coherence of a 6-days interferogram during a snow period (December), you can see how the coherence is totally lost on the higher altitudes, while does not change much in the main valleys.
Due to the explained aspects (mainly to the snow coverage), the version zero (V0) of the DAM shows a very low spatial sampling: the measures are densely located at the bottom of the valley end poorly cover the higher altitudes. This aspect has been strongly improved in the second and third processing results, version 1 and 2 (V1, V2) of DAM, thanks to the introduction of a coherence analysis in order to select interferogram networks to be used. We refer to the related documents (listed below) for more details about the processing steps and the results.
From the DAM, a semi-automatic extraction of the most significant detected active areas, based on the methodology developed in the Safety project (Barra et al. 2017), and using the software ADAFinder developped in the framework of the project MOMIT (Navarro et al., 2018), has been performed to generate the ADA map (see the picture below).
Impact assessment: the vulnerable element at risk map (VEAM)
The Vulnerable elements activity maps (VEAM) are one the main outputs of the WP2 “Tools to support the Early Warning for Landslides geohazard”. VEAM are derived starting from interferometric products, specifically from the ADAs (Active Deformation Areas). The ADAs provide the spatial distribution and the magnitude of ground deformation over the Valle d’Aosta Region, updating the state of activity of already known phenomena or mapping new potential slope movements. The aim of the VEAM procedure is to assess the impact of detected geohazard on road networks and built-up areas. The VEAM consists in a simplified colour scale map indicating those structures and infrastructures with a greater probability to suffer for the impact of a geo-hazard and those structures and infrastructures affected by the dynamic of an active geohazard.
Considering the urban distribution and the type of landslides that can be monitored, the VEAM procedure has been based on two main approaches: the direct and the indirect impact estimation. Direct impact, which has been already followed by Solari et al. (2018) for geohazards mapping in Canary Islands (Spain), is based on the overlapping between ADAs (expression of landslide intensity) and elements at risk. Thus, vulnerability is estimated as a value (ranging between 0 and 1) as a function of intensity and building/road typology. The indirect impact is needed when the type of landslide (i.e. debris flow) and the localization of the ADA (i.e. detrital area at high altitude) do not allow the direct estimation of vulnerability. In this case, a further elaboration is needed. In particular, the Gravitational Process Path (GPP - Wichmann, 2017) is applied to model, using as source the ADAs, the landslide path hypothesizing its complete failure. We refer to the related documents (listed below) for more details about the processing steps and the results.
Barra, A., Solari, L., Béjar-Pizarro, M., Monserrat, O., Bianchini, S., Herrera, G., ... & Ligüerzana, S. (2017). A methodology to detect and update active deformation areas based on sentinel-1 SAR images. Remote Sensing, 9(10), 1002.
Navarro, J. A., Cuevas-González, M., Barra, A., & Crosetto, M. (2018). Detection of Active Deformation Areas based on Sentinel-1 imagery: an efficient, fast and flexible implementation. In Proceedings of 18th International Scientific and Technical Conference (RACURS 2018), Crete, Greece.
Solari, L., Barra, A., Herrera, G., Bianchini, S., Monserrat, O., Béjar-Pizarro, M., ... & Moretti, S. (2018). Fast detection of ground motions on vulnerable elements using Sentinel-1 InSAR data. Geomatics, Natural Hazards and Risk, 9(1), 152-174.
Trigila, A., Iadanza, C., & Spizzichino, D. (2010). Quality assessment of the Italian Landslide Inventory using GIS processing. Landslides, 7(4), 455-470.
Wichmann, V. (2017). The Gravitational Process Path (GPP) model (v1.0) - a GIS-based simulation framework for gravitational processes. Geosci. Model Dev., 10, 3309-3327.