Potential Analysis of Deep-Seated Landslides Caused by Typhoon Morakot Using Slope Unit

Abstract

Taiwan is prone to landslide hazard due to fragile geological conditions, frequent earthquakes, and heavy rainfall in the monsoon seasons. The severe rainfall often caused deep-seated landslides. This study utilized the slope unit and discriminant analysis to assess the deep-seated landslide potential in four sub-basins in the Namasia area, Kaoshiung City. The deep-seated landslides were mapped using LiDAR data and aerial photos of the study area. Delineation method of the slope unit was developed, and the resulting slope units consisted well with boundaries of deep-seated landslides. Typically, one deep-seated landslide enclosed one to few slope units. The assessment model was constructed using discriminant analysis on two sub-basins, C and D, with slope angle, aspect, height, and area as causal factors. The model was validated and provided good accuracy for the fitting and validation data. The model was then adopted to assess the landslide potential and moderate results were obtained. Differences in property distributions among sub-basins and choice of factors could cause some discrepancy in assessment results. The potential maps of the study were developed accordingly, and good to moderate results were obtained.KeywordsDeep-seated landslidePotential analysisSlope-unitDiscriminant function