考慮不確定性之降雨誘發山崩逆分析

Back analysis strength parameters in uncertainty for rainfall induced shallow landslide

陳麒任(Chi-Jen Chen)

指導教授 ： 董家鈞 ； 劉家男

國立中央大學/地球科學學院/應用地質研究所/碩士(2011年)

摘要

利用物理模型進行山崩潛感分析最大的難題為分析參數不易取得，由逆分析取得參數為可行方法之一，然而，強度參數逆分析常有非唯一性問題，以致於難以客觀決定單一組最佳參數，故如何解決強度參數逆分析非唯一性問題為逆分析重要課題。本研究選定埔里鎮仁愛鄉投71縣道為研究區，並利用民國90年桃芝颱風事件之山崩目錄進行有效凝聚力與有效摩擦角之定值式逆分析。最後再利用貝式定理逆算有效凝聚力平均值與有效摩擦角平均值之機率分布，並以8組實驗值(摩擦角為29.2-42.3度，凝聚力為0-22.67 kPa)驗證定值式與機率式逆分析結果。定值式逆分析參數方法是先挑選模擬-實際崩塌面積比介於0.8至1.2之參數，再以總正確率越高，則參數越佳。機率式逆分析則分兩種情況，其一為考慮山崩目錄更新參數結果，其二為了增加逆分析結果之可信度，故同時考慮實驗值和山崩目錄更新參數。結果顯示，定值式逆分析最佳之參數，其凝聚力平均值為14 kPa，摩擦角平均值為34度，與實際值範圍大致相符，但仍具參數非唯一性。機率式逆分析中，僅考慮山崩目錄更新參數之結果顯示，機率最高之凝聚力平均值為3 kPa，摩擦角平均值為40度，結果雖與實驗值仍有差異，但結果顯示機率法逆分析應用於區域淺層山崩模型之可行性，且不再需要挑選唯一強度參數組，解決了參數非唯一性問題。同時考慮實驗值和山崩目錄更新參數之結果顯示，機率最高之凝聚力平均值為7 kPa，摩擦角平均值為37度。

關鍵字

貝氏定理 ； 土壤強度參數 ； 逆分析 ； TRIGRS ； 淺層山崩

並列摘要

It is difficult to get the soil strength parameters (effective cohesion and effective friction angle) when using the physical model in landslide susceptibility analysis. Back analysis is one possible way to get parameters. However, back analysis result has the non-unique problem so it is hard to determine unique parameter. So, a key problem in the back analysis is how to solve the non-unique problem. In this study we use deterministic back analysis and probabilistic back analysis to estimate average of effective cohesion and effective friction angle. Landslide inventory is caused by Typhoon Toraji event on 30/7/2001. At the same time, we verify the back analysis results by 8 experiment data(effective cohesion is 0-22.67 kPa, effective friction angle is 29.2-42.3 degree). Methodology of deterministic back analysis is that retain predict-real failure area ratio within 0.8 and 1.2 firstly. And then choose the maximum efficiency of parameter. Methodology of probabilistic back analysis has two different situations. One is only using landslide inventory for back analysis, the other one is using the experiment data and landslide inventory for back analysis. The deterministic back analysis results show that the best fit effective cohesion is 14 kPa and effective friction angle is 34 degree. Although back analysis result still have non-unique problem, it is coincide with experiment data. The first situation of probabilistic back analysis results show that the maximum probability effective cohesion is 3 kPa and effective friction angle is 40 degree. Second situation shows that the maximum probability of effective cohesion is 7 kPa and effective friction angle is 37 degree

並列關鍵字

soil strength parameters ； back analysis ； TRIGRS ； shallow landslide ； Bayes’s theorem

參考文獻

44. 林伯勳、許振崑、冀樹勇，「集水區土壤厚度經驗式應用分析」，中興工程季刊，第111期，35-45頁，2011年。

13. 鐘欣翰，「考慮水文模式的地形穩定分析-以匹亞溪集水區為例」，國立中央大學，碩士論文，2008年。

1. Guimaraes, R. F., Montgomery, D. R., Greenberg, H. M.,Fernandes, N. F., Trancoso Gomes, R. A., de Carvalho, O. A., “Parameterization of soil properties for a model of topographic controls on shallow landsliding: application to Rio de Janeiro”, Engineering Geology, Vol. 69, pp. 99-108, 2003.

2. Salciarini, D., Godt, J. W., Savage, W. Z., Conversini, P., Baum, R. L., Michae, J. A., “Modeling regional initiation of rainfall-induced shallow landslides in the eastern Umbria Region of central Italy”, Landslides, Vol. 3, pp. 181-194, 2006.

3. Salciarini, D., Godt, J. W., Savage, W. Z., Baum, R. L., Conversini, P., “Modeling landslide recurrence in Seattle, Washington, USA”, Engineering Geology, Vol. 102(3-4), pp. 227-237, 2008.