降雨型崩塌地時機預測及土石流模擬-以卡那岸流域為例

NCUIR > Research Center > Master of Science Program in Remote Sensing Science and Technology > Electronic Thesis & Dissertation > Item 987654321/90300

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/90300

Title:	降雨型崩塌地時機預測及土石流模擬-以卡那岸流域為例
Authors:	陳孝怰 Chen, Hsiao-Hsuan
Contributors:	遙測科技碩士學位學程
Keywords:	淺層崩塌;物理機制模式;土石流路徑;裂縫流;崩塌時間;shallow landslide;physically-based model;debris flow path;fracture flow;landslide timing
Date:	2022-08-18
Issue Date:	2022-10-04 14:18:23 (UTC+8)
Publisher:	國立中央大學
Abstract:	颱風伴隨著短延時強降雨的特性，當降雨快速入滲至不飽和土壤，使得土壤剪力強度降低，土石崩落導致淺層崩塌發生，並有可能演變為土石流，挾帶強大的下切力與大量的土石流體衝擊下游地區，突發及未知的特性常對生命安全及公共設施造成強烈的侵害。本次研究以臺灣東部之太魯閣沿海河系內一子流域「卡那岸流域」為研究區域，2012 年蘇拉颱風狹帶強勁降雨於此登陸，導致大小崩塌與土石流事件發生，更出現土石流改道之現象，大量土石沖刷進入當地部落內。本研究採用物理機制模式(physically based model)，其在淺層崩塌的潛勢分析中是常見的分析方法，並以 SHALSTAB 模式為基礎，修改其「單層」之土層結構假設，並將其提出之臨界降雨Qcr發展為臨界濕度 Wcr，結合動態水文模式以及邊坡穩定模式，發展一耦合水文穩定性模式(hydrologic-stability model)，並對於颱風誘發之淺層崩塌影響之範圍、土石流路徑及崩塌發生時間點進行模擬與預測。本研究假設土層結構為「雙層」，在原土壤與基岩之間加入一裂縫層(fracture layer)，地下水累積之初始位置在裂隙層底部，而破壞面的位置則假設在土壤與裂隙層的交界面上，以此探討土壤與基岩裂縫中流動的裂縫流對於邊坡穩定及崩塌時機所帶來的影響。模式結果之淺層崩塌影響範圍將與崩塌目錄交互比對驗證，並透過正確預測受影響區域比例 Nc、整體預測精度 Ac 兩項標準，與誤授(errors of commission, EAcom)、漏授(errors of omission, EAom)進行量化評估。結果顯示，卡那岸流域全區正確預測的受影響區域的比例 Nc 為 81%，整體之預測精度 Ac 為 94%，而在誤授 EAcom 為 35%、漏授 EAom 則為 44%。整體而言，耦合水文穩定模式在受影響區域預測的表現上，預測精度高且受影響面積被高估的情形並不明顯。而崩塌發生時間方面，本研究以不同破壞情境「單層結構」與「雙層結構」預測之崩塌發生時間與水保局發布之重大土石災害報告內提供之「災害發生時間」作為地真資料進行比較。結果顯示，在崩塌時間預測方面，雙層土壤預測之崩塌時間為 61 小時，與報告紀錄之災害發生時間相符，而單層土壤比紀錄之時間提前 37 小時預測到崩塌發生。另外本研究也發現對預測土壤飽和區域，單層結構相較於雙層結構模擬出更多飽和的區域(+22.7%)。由此可知，裂縫層的加入不僅影響預測的土壤飽和範圍，還影響了預測時間，總體而言，研究結果顯示了在模式中考量裂縫流的特性，確能改善模式表現，具有一定效益。 ;Typhoon events are often accompanied by short-duration intense rainfall. If the slopes are steep and rich in loose soil, rocks, or colluvium, it is prone to induce significant debris flows and landslides, causing heavy casualties and economic losses. The occurrence of debris flows and landslides is influenced by geology, topography, and the hydrological environment. To assess the affected area of the combined process, the landslide and debris flow requires an integrated modeling framework. This study adopts the physical-based model, which is based on the SHALSTAB model proposed by Montgomery and Dietrich (1994). Our study developed a coupled hydrologic-stability model by changing its assumptions about soil structure and modifying the critical rainfall into the critical humidity. This study assumes a double-layer model incorporating a fractured layer under the soil layer. The failure plane is located at the interface between the soil and fracture layer, and the groundwater is also accumulated at the soil-fracture layer interface. The landslide event induced by Typhoon Saola in Heping Village, Xiulin Township, Hualien County on August 2, 2012, is selected as the study case. The results will compare with the landslide inventory digitized by the post-disaster orthoimage of the Aerial Survey Office (ASO) and will be quantified through two criteria, which are the proportion of correctly predicted affected areas(Nc) and the overall prediction accuracy(Ac), and examine the over-predicted or under-predicted of the model through errors of commission(EAcom) and errors of omission(EAom). For the landslide timing validation, we use the Major Disaster Event Reportsissued by the Soil and Water Conservation Bureau (SWCB) as the ground truth data for comparison. In the simulation of the debris flow path, the model results show that Nc is 81%, Ac is 94%, EAcom is 35% and EAom is 44%. Under different failure scenarios, the single-layer model simulates 22.7% more saturated areas than the double-layer. In landslide timing prediction, the predicted landslide timing of the double-layer model is 61h, which is corresponded with the report. In contrast, the single-layer model predicts the failure 37h earlier than the observation. It can be seen that the addition of the fracture layer not only influences the predicted affected area but also affects the prediction landslide timing. Overall, the research results show that considering the characteristics of fracture flow in the model can indeed improve the model performance and has certain benefits.
Appears in Collections:	[Master of Science Program in Remote Sensing Science and Technology ] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	87	View/Open

View Licence

社群 sharing

Loading...