Landslide Susceptibility Analysis along Li-shing Mountain Road in Nantou County
力行產業道路 ； 不安定指數 ； 崩塌潛感 ； Li-shing mountain road ； Instability index ； landslide potential
|Volume or Term/Year and Month of Publication||
|Academic Degree Category||
The mountain area in Taiwan is usually high and steep with rugged road. The amount of groundwater seepage in the mountain area will increase due to the improper layout of the drainage system during the typhoon period. This excess seepage will lower the soil stability and easily result in landslide and debris flow damages. Li-shing mountain road in Nantou County has its own weak and frangible rock feature, accompanied by the low vegetation cover ratio due to the overuse on the steep slope-land along the road. Landslide and failure of road’s foundation and pavement will interrupt the transportation, especially for the transport of agricultural products. The target of this reach focuses on the segment with high frequency of landslide, i.e., road mileage from 13K+500 to 37K, and covers the upslope and downslope lands of the road with 700 m in length, respectively. In this study, the orthophoto map before the typhoon Morakot occurred in August 2009 was first collected. Through the artificial identification method, the landslide potential area can be identified and be divided into two groups, namely landslide and non-landslide, by using the landslide ratio with threshold value of 30%. Landslide potential locations along the Li-shing road then can be simulated using the instability index method based on six factors: aspect, slope, elevation, distance to river, distance to road, and topographic wetness index. According to the four ranges of the landslide potential value, four locations can be identified, i.e., stable, low landslide potential, medium landslide potential, and high landslide potential. By comparison of historical landslide locations with the simulated landslide potential map, the simulation accuracy can be evaluated by using the classification error matrix. The simulation accuracies for the landslide, non-landslide, and total groups are 75.93%, 73.02%, and 73.17%, respectively. The simulation accuracy in this study is satisfactory. The information of medium and high landslide potential locations obtained in this study could be useful for the related governmental authorities to plan the landslide warning system and engineering improvement measures.
生物農學 > 生物環境與多樣性