Flood-Prone Areas Assessment Using Linear Binary Classifiers based on Morphological Indices

The identification of flood-prone areas is a critical issue becoming everyday more pressing for our society. A preliminary delineation can be carried out by DEM-based procedures that relay on basin geomorphologic features. In the present paper, we investigated the dominant topographic controls for the flood exposure using techniques of pattern classification through linear binary classifiers based on DEM derived morphologic features. With this aim, local features - which are generally used to describe the hydrological characteristics of a basin - and composite morphological indices are taken into account in order to identify the most significant one. The analyses highlight the potential of each morphological descriptor for the identification of the extend of flood-prone areas. Our findings may help the definition of new strategies for the delineation of flood-prone areas with DEM-based procedures.