The structural and geomorphic characteristics before and after the Chiufenerhshan landslide and possible mechanisms of the slope failure

一九九九年九月二十一日台灣發生芮氏規模7.3集集大地震，集集地震誘發數個規模龐大的山崩，九份二山即為其中之一，此山崩造成39人死亡，財產損失亦十分可觀。九份二山山崩區位於南投縣國姓鄉，約在埔里西方10 km處。此山崩事件搬運了大約36百萬立方公尺的物質到其下方的澀子坑溪溪谷中，並造成三個堰塞湖。崩塌區位於大岸山向斜的西翼，是典型的順向坡地形，主要岩性組成為砂頁岩互層。此地震誘發之山崩非常獨特，本研究希望透過分析地震前後之地形及構造特徵，以瞭解原始地形及構造特徵是否顯示出此區域之不穩定，並探討九份二山山崩可能的坡體破壞機制。 為了要瞭解地形與構造特徵，選擇以數值高程模型為研究工具，因其能將地表形貌紀錄下來，且利於作空間上的分析與進一步的地質判釋。另外為了研究山崩之地形與構造變遷，需要跨越事件前後的地形資料，然而震前唯一的地形資料為1986年40 m的數值高程模型，實在無法滿足關鍵的構造及地形分析，因此依照航空攝影測量法，重新製作震前1998年九份二山地區2 m數值高程模型。震後則有1999年與2002年解析度分別為9 m與1 m的數值高程模型，本研究使用這三筆資料分析本區之構造及地形。 透過DEM的分析發現，九份二山地區震前的不連續面特徵有裂隙、節理及三個斷層弱帶。震前本區域順向坡上之裂隙與水系的發展，與節理的關係密切；震後之崩塌區邊界，與大岸山向斜及震前的節理、裂隙分布相當一致，堆積區則被澀子坑溪周圍的山脈阻擋，暗示著九份二山地區的地形發育深受構造型態影響。坡腳在山崩前後的DEM地形特徵相當類似，山崩事件可能只稍微破壞坡腳地形，或是被崩塌物質覆蓋，因此坡體真正的破壞面應在坡腳上方。 九份二山崩塌區因坡面上之節理、裂隙及水系的發展、侵蝕，削弱坡體的側向支撐。集集地震造成的強地動使得坡體沿層間破裂，加上由裂隙及節理滲入層間的地表水削弱層間的摩擦力，地層的重力位能幾乎由下方的岩層所支撐。坡面上既存的斷層弱帶無法承受坡體的重力，而發生剪切（shearing off）或拱曲（buckling）破壞。庫侖—莫耳破壞準則的評估顯示，九份二山山崩為剪切破壞機制的情況下，坡體必須存在斷層、節理或裂隙等不連續面，則剪切面很可能沿坡腳上方既存的斷層帶發展。 野外調查的證據則較支持拱曲破壞機制，拱曲軸很可能沿崩塌面上的斷層弱帶發育，依此弱帶至坡頂的有效坡長反求坡體的整體楊氏係數為16.89∼22.68 GPa。因此有斷層弱帶之岩體，其楊氏係數應小於所求之整體楊氏係數；而無斷層弱帶之完整岩體，其楊氏係數應大於所求之整體楊氏係數。本研究提出九份二山山崩應為坡體沿層面破裂滑動，並沿著震前已存在的斷層弱帶拱曲或剪切破壞，而非坡腳的破壞造成。The 21st September 1999 Chi-Chi earthquake (ML=7.3) triggered several large landslides in central Taiwan. Among them, the catastrophic Chiufenerhshan landslide resulted in 39 deaths and enormous capital loss. The landslide region is located in the Nantou area about 10 km west from the town of Puli. About 36 million m3 rock mass moved downward slope to the Sezikeng river valley and formed three barrier lakes. The sliding area distributed at the west limb of the Taanshan syncline is a typical dip-slope terrain, and its main rock formation consists of inter-bed sandstone and shale. The goal of this study is to characterize the geomorphic and structural features before and after this landslide. Subsequently, the influence of the observed structural and geomorphic features is discussed and possible sliding mechanisms are proposed for the catastrophic Chiufenerhshan landslide. In order to better characterize the geomorphic and structural changes of this landslide, detailed digital morphological data before and after the event are necessary. However, only one DEM (40 m resolution, 1986) before the Chi-Chi earthquake is available, which cannot satisfy the intended geomorphic and structural analysis. Thus, a higher resolution DEM (2 m resolution, 1998) of the Chiufenerhshan area is reconstructed using the method of aerial photogrammetry. The reconstructed DEM and two other available DEMs (9 m and 1 m resolution, 1999 and 2002, respectively) are used to analyze the characteristics of the geomorphology and structure of the Chiufenerhshan area before and after the Chi-Chi earthquake. The prominent structural discontinuities of the Chiufenerhshan area, including cracks, joint sets, and three fault zones are mapped and discovered through the analysis of the DEMs. The cracks and the stream gullies on the dip-slope apparently formed along the observed joint sets and may effectively reduce the lateral support of the sliding mass. After the Chi-Chi earthquake, the boundaries of sliding area are confined at the west limb of the Taanshan syncline, the major joint sets, and the cracks. The deposit area is restricted by the ridges around the Sezikeng valley. These phenomena imply that the eventual landslide is strongly influenced by the original slope morphology and structure. Several geomorphic observations indicate that the foot of the slope is still preserved and the initial failure location should be above it. It is possible, however, that the foot of the slope is partly destroyed and covered by sliding materials. If the sliding bedding plane was detached and infiltrated by water during the earthquake, it may dramatically reduce the frictional force on the sliding plane. Consequently, most of the gravitational force was supported only by the strength of the slope strata leading to the slope failure either by shearing-off or buckling mechanisms. Based on the Coulomb-Mohr failure criteria, structural discontinuities such as faults, joints, or cracks must exist in preferable orientations in the sliding mass in order for the shearing-off mechanism to occur. Field investigations on structural discontinuities suggest that the buckling failure mechanism is most probable for the landslide rather than the shearing-off mechanism. The buckling hinge may initiate along the observed weak fault zones which are situated on the sliding plane, and a range of young’s modulus (16.89 ~ 22.68 GPa) for the slope body can thus calculated assuming such failure. It is inferred that the Chiufenerhshan landslide slid along the bedding plane, and the dip-slope strata initially failed along the fault concentrated zones instead of destruction at the lowest portion of the slope.口試委員會審定書 I 誌謝 II 摘要 III Abstract V 第一章、緒論 1 1.1 研究動機與目的 1 1.2 文獻回顧 3 1.2.1 地形因子與山崩的關係 3 1.2.2 地質因子與山崩的關係 4 1.2.3 侵蝕與山崩的關係 5 1.3 論文大綱 5 第二章、九份二山山崩及研究區域概況 7 2.1 九份二山山崩事件概述 7 2.2 研究區域地質概述 10 2.2.1 地層概述 10 2.2.2 構造概述 11 2.3 九份二山山崩之文獻回顧 14 第三章、研究方法 17 3.1 攝影測量原理 17 3.2 數值高程模型介紹 18 3.3 數值高程模型製作方法 19 3.3.1 LPS製作DEM流程 19 3.3.2 航空像片選擇原則 22 3.3.3控制點選取原則 23 3.3.4 誤差報告說明 26 3.4 空載光達萃取地面三維資料 27 3.5 地面光達萃取地表三維資訊 29 3.6 空間分析方法 29 第四章、資料分析與結果 34 4.1 本研究所使用之數值高程模型 34 4.2 數值高程模型之空間分析 38 4.3地形分析 42 4.3.1 山崩前之地形分析 42 4.3.2 山崩後之地形分析 45 4.3.3 山崩前後之地形比較 49 4.3.4 九份二山地區山崩前後之地形特徵 55 4.4 構造分析 55 4.4.1 山崩前之構造分析 55 4.4.2 山崩後之構造分析 56 4.4.3 九份二山地區山崩前後之構造特徵 57 第五章、九份二山山崩可能的破壞機制 59 5.1 山崩前後之斷層弱帶 59 5.2 剪切式破壞 60 5.2.1 最大主應力與坡體長度之關係式 60 5.2.2 岩體不連續面與坡體破壞最小長度之關係 62 5.2.3 可能之剪切面位置 63 5.3 拱曲式破壞 64 5.4 可能的山崩破壞機制 66 第六章、討論 69 6.1 造成順向坡不穩定之因子 69 6.2 影響山崩邊界之因子 69 6.3 坡腳的地形變遷 71 6.4 LPS ATE與Inpho Match-T之比較 71 6.5 地面光達與空載光達之適用情況 72 第七章、結論 74 參考文獻 76 附錄A：攝影測量原理 80 A.1 航空照片簡介 80 A.2視差原理 81 A.3 像片之基本幾何性質 83 A.4透視投影與正交投影 85 附錄B：臺灣的大地基準與投影法 89 附錄C：航照相機參數 91 附錄D：框標點誤差 93 附錄E：外方位參數 94 附錄F：點選地面控制點 95 附錄G：空中三角測量誤差報告 96 附錄H：LPS ATE萃取DTM之誤差報告 100 附錄I：Match-T萃取DEM之參數及誤差報告 102 附錄J：地面光達萃取地表三維資訊 107 J.1 地面光達簡介 107 J.2 地面光達測量 108 J.3 地面光達點雲資料特性與分析 110 J.4 規劃地面光達測量計畫之建議 11