Abstract In this study, the relationship between the sustainable development goals (SDGs) and snow avalanche susceptibility has been analyzed for the first time. Snow avalanche susceptibility of Uzung?l basin, which is a specially-protected area in Trabzon, Türkiye, was generated by a novel proposal of ensemble modeling of hesitant fuzzy sets and decision tree-based Machine Learning (ML) algorithms. The uncertainty effect of the snow avalanche conditioning factors was expressed regarding the Bel values at each class on snow avalanche susceptibility. Hesitant fuzzy ordered weighted averaging operator was used for aggregation of the ML classification of snow avalanche conditioning factors. The predicted avalanche susceptibility maps were validated by a receiver-operating-characteristics curve method and the areal-percentages of avalanche classes, and avalanche percentages at the classes. Area under curve and true skill statistics values for HFS-J48, HFS-RT and HFS-REPTree for the training process were calculated as 0.985, 0.966, 1.000, 0.989, 0.969, and 0.943, respectively. These values for testing process were calculated as 0.975, 0.947, 0.917, 0.840, 0.955, and 0.920, respectively. Although HFS-RT predicted the best for the training process, the HFS-J48 model was found to be performing the best predictions of snow avalanche susceptibility regarding the testing process and predicted classified areal and avalanche percentages. The findings of this study may contribute to further understanding achievement of many goals regarding environmental, ecological, and spatial, and landscape planning. The results of this study may be considered to achieve the goals of some SDGs such as tourism planning, developing economic activities, providing sustainable transportation, and land use control.
展开▼
