Please use this identifier to cite or link to this item:
|標題:||A study on characteristic and hazard assessment of landslides in Shihmen reservoir watershed, northern Taiwan
In July and August 2004, heavy rainfall brought by Typhoon Aere caused over 200 landslides in the Shihmen Reservoir watershed in northern Taiwan. This in turn resulted in heavy sedimentation in the Shihmen Reservoir and interruption of public water supply for more than two weeks in the Taoyuan area. This study firstly investigated the characteristics (i.e. location and landscape attributes of the occurrence sites, .etc.) Secondly, a method for landslide hazard determination was developed with consideration given to both the site factors contributing to potential unstable slopes and the triggering factor - rainfall. This landslide hazard determination method was applied to the Shihmen Reservoir watershed that was partitioned in 40m *40m grids and compared to landslides occurred after the 2004 Aere typhoon.
Very few areas in the world experience landslides of the magnitude and extend such as those occurred in the Shihmen reservoir watershed. A survey in 1968 reported 435 landslides with a total area of 184 ha in area and 2,808,218 m3 in volume. The area of landslides in 1968 represented a 0.24 percent of the total watershed area of 76300 ha. The new landslides induced by extremely high rainfalls in 2004, however, accounted for more than 1.1% of the total watershed, a significant increase over the results of the survey in 1968. Theses landslides in 2004 were mainly caused by the Typhoon Aere rainstorms with high total amounts, ranging from 500 mm to over 1600mm, estimated to be having 200 years return periods. The landslides in 2004 can be related to some landscape and environment attributes of the occurrence sites. Among them, slope gradient appears to be the most apparent factor positively related to landslide occurrence. Accumulated rainfall amount is also found to be important in landslide initiation.
For landslide hazard assessment (determination) this study utilized the theory based on infinite slope method. Essentially the landslide hazard is inversely related to factor of safety. Both site factors such as the slope gradient that contribute to slope instability and the landslide triggering factor of rainfall are included in the landslide hazard determination. These landslide hazard determination methods quantify the influences of infiltrated rainwater on groundwater level and, subsequently on slope stability. Two assumptions are made regarding soil infiltration capacities; one is determined by the NDVI (Normalized Difference Vegetation Index), while the other is related to the land use types of each grid. GIS technique was used to the landslide hazard mapping of the Shihmen Reservoir watershed that was petitioned into 40 * 40 grids.
Comparison of landslide hazard calculated with the method developed by this study and actual landslide occurrence of each grid only give successful prediction rate of slightly more than 1%. This indicates that although the method used by this study may be theoretically sound for the 40m*40m grid landslide hazard determination, it may not represent the real field picture of landslide occurrence in the Shihmen Reservoir watershed. For example, the size of every landslide is a composite of many 40*40m grids, but the real cause may be from only a few grids. In addition, not all infiltrated rainwater contributes to increase groundwater level. Part of infiltrated water move down slope and may drain out the slope land completely. The down slope movement of water may also itself be part of factors influencing the slope instability. Moreover, the up and down/or horizontal slope curvature may influence the movement and concentration of infiltrated water that in turn could be the cause of many landslides. In other words, the landslide hazard evaluation method that includes the influence of infiltrated water may be right theoretically but need many improvements so as to match the real field slope situations before wide application with meaningful and useful results.
本研究利用無限邊坡理論計算安全係數以評估崩塌潛勢，分二種假設決定入滲率與地下水深比關係，第一種假設入滲係數為常態化植生指數(NDVI, Normalized Differential Vegetation Index）之函數，根據荷頓（Horton）理論，入滲會隨時間呈指數逐漸下降，最後達到一穩定值，所以是入滲係數隨時間而變而有不同數值，第二種假設乃依據森林覆蓋率，而給予特定值之入滲率，不隨時間而變，並將石門水庫集水區劃分為472,710個40*40m網格，算出每一網格的崩塌潛勢(危險度)，與衛星變異點判釋之現場實際崩塌地做比對。本研究應用力學、地形及降雨量相關參數，算出每一網格安全係數，此方法配合電腦計算，可快速有效進行大區域面積之安全係數計算，有利於針對尚未發生崩塌地點，進行穩定分析或計算其崩塌潛勢，未來也可以特定集水區或區域範圍進行其崩塌潛勢或土石流潛勢之評估；本研究所採用崩塌潛勢評估方法考慮到3維空間之安全係數條件，但未考慮深層滑動及網格與網格間之摩擦力作用，亦未對同一邊坡剖面之連續網格進行安全係數相加，此外本研究方法針對工程細部設計分析，並未考慮到地形曲率或地表下水流對崩塌之影響，皆值得未來作更深入探討。
|Appears in Collections:||水土保持學系|
Show full item record
TAIR Related Article
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.