Landslide is one of the major natural disasters which cause extensive loss of life and property. During the last three decades, different researchers have developed different methodologies to prepare landslide suscept...Landslide is one of the major natural disasters which cause extensive loss of life and property. During the last three decades, different researchers have developed different methodologies to prepare landslide susceptibility mapping and hazard assessment in the world. The main goal of this paper is to apply a fuzzy logic approach to landslide susceptibility mapping in the Garuwa sub-basin area, East Nepal. Eight different causative factors are considered： slope angle, slope aspect, slope shape, relative relief, distance from drainage, land use, geology, and distance from active faults. Likelihood ratios are obtained for each class of causative factors by comparison with past landslide occurrences. Then, the likelihood ratios are normalized between zero and one to obtain fuzzy membership values. Next, different fuzzy operators are applied to generate landslide susceptibility maps. Compar- ison with the landslide inventory map reveals that the fuzzy gamma （γ） operator with a y-value of 0.70 yields the best prediction accuracy which is then used to produce the final landslide susceptibility zonation map.展开更多
文摘Landslide is one of the major natural disasters which cause extensive loss of life and property. During the last three decades, different researchers have developed different methodologies to prepare landslide susceptibility mapping and hazard assessment in the world. The main goal of this paper is to apply a fuzzy logic approach to landslide susceptibility mapping in the Garuwa sub-basin area, East Nepal. Eight different causative factors are considered： slope angle, slope aspect, slope shape, relative relief, distance from drainage, land use, geology, and distance from active faults. Likelihood ratios are obtained for each class of causative factors by comparison with past landslide occurrences. Then, the likelihood ratios are normalized between zero and one to obtain fuzzy membership values. Next, different fuzzy operators are applied to generate landslide susceptibility maps. Compar- ison with the landslide inventory map reveals that the fuzzy gamma （γ） operator with a y-value of 0.70 yields the best prediction accuracy which is then used to produce the final landslide susceptibility zonation map.
