AUTOMATIC PROCEDURE FOR SUPERFICIAL LANDSLIDE HAZARD MAPPING USING GRASS
In landslide risk management, hazard assessment is an important factor for creating thematic maps of general risk conditions, provided within basin plans, and constitutes an essential support in decision making, both for territorial proper planning and for optimization of preventive interventions.
This work proposes an automatic procedure, implemented in GRASS GIS software, aimed at creating maps of landslide hazard over large areas using multivariate statistical analysis. Instability factors so far considered are slope, lithology and land use, attributes of proven importance in landslides' genesis process and easily available on large areas. The procedure consists in five steps, as follows:
data acquisition and creation of information layers;
definition and delimitation of test areas, in which to calibrate the model;
automatic creation, within test areas, of information layers showing before-landslide conditions. Since the landslide happens, it changes the configuration that gave rise to it: hence the attempt to rebuilt the original information;
model calibration, which consists in intersecting the information levels inside test areas and in extracting areal informations needed to calculate landslide conditional probability given the occurrence of x vector, where x is composed by the analysed instability factors;
model application to the entire study area to produce the landslide hazard map.
This procedure is immediately applicable to any area without the need of calibration phase or inclusion of geotechnical parameters, because study area specificities are implicitly taken into account using as test areas historical landslide areas inside the same region or in areas similar for lithological and vegetation features.
The skeleton of the procedure was already presented at the FOSS4G Italian Meeting on February 2010. Here the completed procedure, the first results and their validation will be presented.
Bianca Federici - University of Genova
Damiano Natali - University of Genova