Landslide Susceptibility Assessment Using Evidence Belief Function and Frequency Ratio Models in Taounate city (North of Morocco)

The exploitation of a suitable cartography of a well-identified geographical territory inevitably passes by reliefs drawn by the movements of ground. The most frequent of these movements are landslides such as the case of the Taounate city in the mountainous area of northern Morocco. The aim of the work is to develop a landslide susceptibility map of Taounate city based on the evidence belief function (EBF) and the frequency ratio (FR) Models. The landslide inventory map was prepared using interpretation of aerial photographs, satellite images and field visits. Twelve predisposition and release factors (slope, lithology, aspect, annual precipitation, elevation, curvature, normalized difference vegetation index (NDVI), land cover, distance from rivers, and distance from roads, topographic wetness index (TWI), and surface roughness) are taken into account in this study to develop susceptibility maps. The validation of the resulting susceptibility maps was obtained by analyzing the area under the curve. The results of the validation showed that the combination of the predictive factors: terrain slope, terrain aspect, lithology, TWI and NDVI proved to be the best combination for both FR and EBF models and give values for the area under the curve equal to 98% (0.9779) and 87% (0.8712) respectively. These coefficients reflect the good degree of reliability of the models used.