Landslide Susceptibility Assessment of Forest Roads*

Authors : Remzi Eker, Abdurrahim Aydin

Pages : 54-60

View : 23 | Download : 10

Publication Date : 2016-11-18

Article Type : Research Paper

Abstract :In last few decades, there has been an increasing interest in using Landslide Susceptibility Maps insert ignore into journalissuearticles values(LSMs); especially in planning and decision making stages of landslide prevention and mitigation activities, as well as in landslide related studies. In forested areas, inappropriately located roads potentially cause slope instability problems such as landslides which then result in serious destructions on road networks and deformations on road platforms. Thus, one of the further usages of LSM may involve overlapping analysis with forest roads in order to obtain information about how road networks should be planned and located considering land sliding potential. Statistical approaches such as Logistic Regression insert ignore into journalissuearticles values(LR); method are well integrated with GIS based evaluation of landslide probability of slopes in larger regions. In this study, LSMs of two forest districts insert ignore into journalissuearticles values(Gölyaka and Kardüz); in Gölyaka Forest Directorate insert ignore into journalissuearticles values(Düzce, Turkey); was generated by using LR method based on an inventory of 52 landslides and eight conditioning parameters. These parameters include elevation, slope, land-use, lithology, aspect, distance to faults, distance to streams, and distance to roads. For overlapping analysis, forest road layer was obtained from Bolu Regional Directorate of Forestry insert ignore into journalissuearticles values(RDF); in vector data format. It was found that landslide susceptibilities obtained in study area were between 0 and 0.57 with 0.85 AUC insert ignore into journalissuearticles values(Area Under the Curve); value. The results indicated that all of the selected parameters had positive effects on landslide occurrences. After normalization of generated susceptibility values between 0 and 1, LSM was classified into following five classes: very low insert ignore into journalissuearticles values(0-0.2);, low insert ignore into journalissuearticles values(0.2-0.4);, moderate insert ignore into journalissuearticles values(0.4-0.6);, high insert ignore into journalissuearticles values(0.6-0.8);, and very high insert ignore into journalissuearticles values(0.8-1.0);. Then, classified LSM was overlapped with forest road layer which includes the total of 380.8 km road. According to classified susceptibility map, more than 95% of total area is located in very low and low susceptibility classes, 3% of the area has moderate landslide susceptibility, while remains have high and very high susceptibilities. According to overlapping analysis, 1.3 km of roads is located within very high susceptibility and 5.1 km of roads is located within high susceptibility classes. The rest of the roads insert ignore into journalissuearticles values(i.e. more than 95%); are located in other susceptibility classes.
Keywords : Forest roads, Landslide Susceptibility Maps, Logistic Regression