The susceptibility analysis of landslide using bivariate and multivariate modeling techniques in western Algeria: case of Fergoug watershed (Beni-Chougrane Mountains)

Landslides are considered to be the most important and frequent natural events causing considerable damage. Many areas in Algeria and elsewhere in the world are affected by this type of phenomenon. Located in the Tellian Chain of the Northwest of Algeria, Fergoug watershed is largely affected by several landslide events that led to (siltation of dams and roads, bridges, and buildings destruction). The damage cost amounted to hundreds of thousands of dollars and sometimes exceeds the reaction of the authorities concerned. In order to help local authorities in their prevention approach, a landslide sensitivity map has been realized using different models: (i) the frequency ratio (FR), (ii) the linear multiple regressions (MLR), and (iii) the information value model (IVM). The landslide inventory was established and includes 142 landslides and 10 conditioning factors (slope angle, slope aspect, profile curvature, distance to rivers, roads, faults, earthquakes, land use, lithology, and precipitation). These factors were prepared from several multisource data sources. The results were validated using the operating characteristic of the receiver and the areas under the curves obtained using the methods FR, IVM, and MLR are respectively 0.87, 0.83, and 0.81. It is proposed that the landslide susceptibility map produced from the FR model be more useful for the study area. The results demonstrate that for the frequency ratio model, the very high, high, moderate, low, and very low susceptibility classes are 58%, 24%, 5%, 2%, and 9.5%, respectively. Almost 73 landslides events are situated along the Fergoug River. These results could reveal the relative importance of different factors in explaining landslides and help engineers plan land use planning.