Landslide Susceptibility Assessment Under Extreme Rainfall-A Case Study of Chenyulan River Basin

我國於防災業務每年皆投入大量人力及資源，近年來極端氣候頻繁，降雨極端且集中，多處地區歷經眾多重大災害造成多處慘重災情，如暴雨所導致的邊坡崩塌、土石流侵襲等，嚴重重創觀光、農業及人民身家財產安全，若能有效掌握邊坡崩塌潛勢高敏感地區，則可針對不同崩塌風險值給予各種防護及防災應變措施，便可降低災害影響程度。 為瞭解雨量對邊坡之影響，本研究以南投縣陳有蘭溪流域為研究區域，陳有蘭溪為濁水溪的主要支流，全長約42.4公里。本流域歷經多次災害，主要皆為颱風所夾帶大量降雨所導致，大量降雨造成流域類各種災情不斷，故本流域可謂為台灣重點災區。 本研究以不同重現期距年分別代表各極端降雨事件，研究中先蒐集陳有蘭溪流域雨量站之年最大一日降雨資料進行降雨頻率分析，並經由K-S檢定及RMSE指標選出較佳分析模式，將其5、10、25、50、100、200年之重現期距雨量以二維空間推估方式繪出等雨量線圖，空間推估比較距離權重反比法、最小曲率法、克利金法及聯合克利金法等四法，經分析以加入高程因子之聯合克利金法較適於陳有蘭溪流域。研究中並以莫拉克災後SPOT衛星影像進行常態化差異植生指標(NDVI)分析，結合影像分類技術萃取災後現有崩塌地，將崩塌地套入各重現期距雨量圖、地質、水系、坡度、坡向及高程等七大因子計算崩塌網格密度及權重，得出各個因子不安定指數值，建立出各重現期距下之崩塌潛感圖。研究發現當流域達10年重現期距雨量，即驅動邊坡破壞之機制，邊坡破壞之敏感區域大幅增加，判定陳有蘭溪流域以10年重現期距雨量為崩塌臨界雨量值，研究中並針對10年重現期距之崩潛圖進行探討，分析高風險之村里、溪流及易崩塌之條件，以供後續防災及相關研究之參考。In Taiwan, many manpower and resources for the disaster reduction in every years, due to the frequent phenomena of extreme weather and severely concentrated rainfall in recent years, Taiwan has experienced a number of severe disasters, such as heavy rain resulted in slope failure, landslide hazards, seriously damaged tourism, agriculture and resident''s safety and property, in order to reduce the impact of disasters, if could grasp the landslide susceptibility, can give a variety of disaster preparedness and response measures. In this study, watershed of Chenyulan River in Nantou county in central Taiwan, assessing the impact and disasters of rainfall on slope, Chenyulan River is the major branch of Jhuoshuei River, has a mainstream about 42.40 km. After many times disasters in Chenyulan River, mainly seasonal disasters induced by typhoons which caused a serious threat to human life and property, Chenyoulan River as the focus of the key disaster area in Taiwan. The study uses different return periods represent extreme rainfall event, this project collected the annual maximum single day rainfall of many rainfall stations at Chenyulan River, for storm frequency analysis and obtained the best analytical model by K-S test and RMSE, and use spatial interpolation method for to draw 5,10,25,50,100,200 year return period rainfall isohyet chart, used in this study Kriging method, Inverse Distance Weighting method (IDW), minimum curvature methods (Spline), and multivariate geostatistical algorithms(Co-Kriging method) for spatial estimation compared, Co-Kriging method uses the rainfall value and the elevation as auxiliary variable,the results showed that the most suitable is Co-Kriging method. Based on the serial satellite images and Normalized Difference Vegetation Index (NDVI) was calculated to analyze the landslide potential for Chenyulan River, Furthermore, several environmental factors, including elevation, slope, aspect, vegetation index, return period rainfall, and so on were considered to in the landslide potential evaluation model. The study found that when the basin up to 10-year return period rainfall, sensitive areas of slope failure significantly increase, the study judgment that the 10-year return period rainfall is the collapse critical rainfall value at Chenyulan River, and for the 10-year return period of landslide potential evaluation model to explore, the paper concludes by providing guidelines for disaster reduction and related research in this study.第一章 緒論 1.1 前言 1 1.2 研究目的 1 1.3 研究流程與架構 2 1.4 本文組織 3 第二章 文獻回顧 2.1 極端降雨及重現期距定義 5 2.2 崩塌定義及崩塌地判釋相關研究 7 2.3 崩塌潛感分析相關研究 10 2.4 降雨頻率分析相關研究 12 2.5 空間降雨推估相關研究 14 第三章 研究區域概述 3.1 地理位置 17 3.2 流域概況 18 3.3 地形特色 18 3.4 流域高程及坡度概述 18 3.5 地層分佈 19 3.5 流域氣候及水系 21 3.6 歷史災情概述 22 3.7 流域雨量觀測站 25 第四章 研究方法及理論介紹 4.1 衛星影像選取 28 4.2 崩塌地判釋 29 4.2.1 常態化差異植生指標(NDVI) 29 4.2.2 影像分類 30 4.3 降雨頻率分析 32 4.3.1 水文重現期距及基本參數 34 4.3.2 三參數對數常態分佈 36 4.3.3 極端值一型分佈 36 4.3.4 皮爾遜第三型分佈 37 4.3.5 對數皮爾遜第三型分佈 39 4.3.6 頻率分析適合度檢定 39 4.4 二維空間降雨推估 41 4.4.1 距離反比權重法 41 4.4.2 最小曲率法 42 4.4.3 一般克利金法 43 4.4.4 聯合克利金法 49 4.5 崩塌潛勢評估 51 4.5.1 不安定指數法 53 4.5.2 分析因子 55 第五章 成果與討論 5.1 降雨頻率分析 65 5.1.1 降雨頻率分析結果 66 5.1.2 頻率分析模式選定 72 5.2 空間降雨推估 73 5.2.1 颱風事件降雨推估成果 74 5.2.2 二維降雨推估比較 76 5.2.3 重現期距降雨推估 78 5.3 崩塌潛勢評估-不安定指數分析 80 5.3.1 崩塌潛感分析 80 5.3.2 影響因子探討 89 5.3.3 高風險地區界定 93 第六章 結論與建議 6.1 結論 97 6.2 建議 98 參考文獻 10