3d Spatial Data Integration for Avalanche Risk Management

The southern part of the Alps is interested by frequent avalanche phenomena. It is very important to provide a suitable method to manage avalanche risk. The usual approach is based on the mapping of the avalanches occurred on the "Carta di Localizzazione Probabile delle Valanghe" (C.L.P.V. Possible Avalanche Location Map). Modern methods are based on statistical approaches or on the evaluation of the morphology and vegetation characteristics. We have chosen to follow the latter approach because in this way we can locate not only the events occurred but also the dangerous areas where no past events have been reported. Different kinds of data have been integrated in a Geographic Information System to evaluate land morphology and vegetation types. Val di Pejo, located in the north-western Trentino, an Italian alpine region which shows frequent and sometimes huge avalanche phenomena, has been selected as test area. A digital terrain model has been created with the 3D algorithm of the Geographic Information System GRASS. A map representing the different vegetation types has been obtained using the information of the Trento’s Forest Management Bureau. The boundaries of the vegetation types in the maps used for forest management are generally approximated, so it has been necessary to verify the real extension of the different kinds of vegetation. An orthophoto has been obtained by differential rectification of digitalised aerial photographs with the help of the DTM and some control points. The orthoimages have been used to test the real location of the boundaries and the extension of the parcels. The avalanche risk areas have been recognised applying morphologic criteria (slope between 28° and 55° and minimum surface of about 625 m, upstream slope change greater than 10°). An algorithm which uses these morphologic rules has been developed and applied to obtain a map of the "morphologic risk", i.e. areas showing an avalanche probability based only on their geometric features. The vegetation has been classed in three different coverage types depending on their density, since the latter influences their ability to avoid the creation of a compact and homogeneous snow layer. A map of the vegetation’s protection ability has been obtained. Both maps themselves can be useful to depict the risk situation but a dramatic improvement of the precise location of the risk areas is obtained by combining the two maps. The resulting map is the one we use to assess the avalanche risk. This map has been verified comparing it to the C.L.P.V.. Three different regions have been recognised: · real risk areas where our map locates high avalanche probability and the phenomenon has been reported; · areas where the protection ability of the vegetation coverage balances the morphologic risk; · areas where no avalanches have been reported but the vegetation cannot face the morphologic risk. The real ability of the different vegetation classes to offer protection against avalanches has been evaluated comparing the morphologic avalanche risk area with the extension of the events occurred. The ratio between the real surface covered by avalanches on the C.L.P.V. and the potential surface obtained following the described criteria, if divided in three different vegetation classes, highlights the importance of the vegetation coverage in protecting from avalanche risk. The creation and the use of the 3D model and its integration with digital images and environment data has permitted the elaboration of thematic maps which contain much information suitable for forest and land management.