降雨入滲誘發淺層崩滑與土石流之模式改善

Abstract

台灣山地常見強降雨事件導致土體崩滑後大量土砂和雨水摻混後形成土石流的災害案例。近二十年，土石流災害影響範圍物理計算的技術已有許多模式與相關案例應用驗證的研究發表，儘管此類物理模式用於評估土石流災害範圍解已被認為是決方案之一，然而，對於災害發生前的預測應用，最迫切需要解決的問題仍是如何識別降雨發生時可能破壞的地點、破壞深度分布與坡壞料源的多寡。 評估降雨導致土體崩滑主要包括三個程序，一為降雨入滲分析，即採用適合的降雨入滲模式，針對降雨入滲導致土體含水量或地下水位的變化與其水力作用進行分析，其次再以降雨入滲分析結果為基礎，計算土體受力特徵並進行穩定性分析，再判斷滑坡可能發生的地點與滑坡發生當下的破壞深度與體積量。最後是結合土砂傳輸運移模式，進一步評估土砂運移過程的致災規模與影響範圍。在此原則下，本研究依序進行斜坡滲流擴散現象解析、斜坡穩定分析，並透過守恆方程式擾動分析建立起動條件，藉由引入極限平衡狀態的流深、滲流擴散與密度變化關係與降伏應力項，耦合起動條件滲流擴散關係。土砂運移模式的建立以弱剪層深度積分近似後再進行強剪層質量修正與動量修正，利用動量積分獲得強剪層厚度關係式，由弱剪層近似控制方程式先推求流速與流深的基礎下，代入強剪層厚度關係獲得強剪層厚度後，進一步進行質量傳輸與動量傳輸修正，重覆迭代至穩定收斂後可獲得修正之流深與平面流速。 本研究最後以莫拉克大鳥聚落土石流模擬的案例展示現階段成果，透過降雨促崩與邊坡穩定分析，確定大鳥集水區不穩定區域與不穩定料源的分佈，並據此提供土石流模擬初始料源深度與位置的輸入資料，後續再進一步進行土石流致災規模與致災範圍的評估。研究呈現滲流擴散時山坡地孔隙水壓力隨時間變化以及安全係數的變化，滿足平衡條件當下萃取深度做為土石流初始料源，再以土石流模擬呈現流速與流深的變化與最後的堆積範圍，提供防災資料有用的確認。

The shallow landslide and subsequent debris flow from rainfall-induced was a frequent disaster in Taiwan's mountainous areas. In the past two decades, the physical model about the hazards zone mapping of debris flow has been published in many research. Although such physical models are used to assess the scale of debris flow disaster has been considered as one of the solutions. However, the most urgent problem to be solved is still how to identify the locations that may be damaged, destroy the depth distribution and the source of the collapse source. The assessment of rainfall induces collapse mainly consists of three procedures. The first is analysis the change of soil water content and water table under hydraulic effect caused by rainfall infiltration. Therefore, based on the results of rainfall infiltration analysis, the soil stress characteristics are calculated and the slope stability analysis is carried out, and then the location of the landslide and the depth and volume of the landslide are determined. Afterward, combined with the sediment transport model to assess the hazards zone and disaster scale. Under such a principle, the combining model of this study includes a seepage diffusion analysis, slope stability analysis, and a debris flow model developed. The conservation equations applied depth integral in strongly shearing layer and a weakly shearing layer stratified. We solve a flow depth and flow velocity component through a scheme of a weakly shearing layer approximation and strongly shearing layer modification. The combining model will use a simulation of a real case of a rainfall infiltration induced shallow landslide and debris flow during the Typhoon Morakot on the Daniao tribe’s hill in eastern Taiwan. The detail data of investigation after the disaster of Daniao tribe will be used for the inputs provided and for validation of the simulation. The study revealed the safety factor changed under a seepage diffusion. The hillside collapse occurred when the safety factor reduced to Fs = 1, and the collapse zone and collapse depth were identified by the locations that would satisfy the condition Fs = 1. Therefore, these data provide to simulation of the subsequent motion of debris flow and hazard zone assessment. The result of collapse zone and collapse depth identified within 8.2% and 20.5% errors from a comparison of the DTM variation analysis of before and after a landslide. Further, the debris flow hazard zone of simulation was within 25% errors from the comparison of a disaster aerial photo. The combining model applied to the simulation of a rainfall infiltration induced landslide and, subsequently, the model was applied to the debris flow of Daniao tribe disaster case, a real disaster range with a nearly 75% match. In spite of this match, the hazard zone from the simulation still included a real disaster range, the simulated results that were well-matched. Therefore, we believe that the combining method of the study would provide a better solution for disaster assessment. Two contributions of our work are: (1) to identify the collapsed zone and map the mapping of the hazard zone’s subsequent motion. (2) These results have potential application in large range sediment disaster prediction and would be of great help in the management of slope disaster prevention.