基於微地型特徵之溪流型土石流潛勢評估模式-以板岩地區為例

本研究以集水區微地形特徵觀點，探討溪流型土石流之潛勢評估方法。研究採用高屏溪流域2009年莫拉克颱風期間發生坡面型土石流事件，坡面型土石流發生與未發生土石流之案例各35處、驗證案例各16處進行坡面型土石流與崩塌之判釋模式研發，而後再進一步應用判釋成果，發展溪流型土石流之潛勢分析模式。本文首先以費雪區別分析方法，考慮土石流案例集水區之微地形特性，重新建立坡面型土石流判釋模式，並驗證其判釋成效。研究成果顯示坡面型土石流模式判釋成果之母體與驗證區樣本正判率分別為91.4%、88.1%，具有不錯之預測效力。因子以集水區面積、發生區形狀係數、發生區凹陷程度、溪流線長度、崩塌潛感面積、形狀係數比、發生區坡面平均坡度比等7項因子為之組合時，正判率最佳且各因子均為獨立變數，而溪流型土石流潛勢評估模式之研究方面，本研究選取高屏溪溪流型土石流潛勢溪流32處、驗證區10處，研擬出溪流型土石流潛勢之評估指標；成果顯示土石流溪流能以集水區內坡面型土石流單元總面積、發生區分析單元總面積、平均隘口高程比三個指標因子，較符合溪流型土石流潛勢指標，應用此三項指標因子進行潛勢分級可獲得良好正判之成效，適用於溪流型土石流實際發生之案例，以溪流集水區內坡面型土石流單元總面積為指標者，可分成三個等級:84公頃以下為低潛勢土石流溪流，在263公頃以上者為高潛勢，兩者之間者為中潛勢溪流；以坡面型土石流發生區總面積者，79公頃以下者為低潛勢土石流溪流，189公頃以上者為高潛勢，兩者之間者為中潛勢溪流。其次，以平均隘口高程比為指標者，平均隘口高程比0.71以上者為低潛勢土石流溪流，0.62以下者為高潛勢，兩者之間者為中潛勢溪流。綜整溪流型土石流潛勢分級研究結果顯示，以微地形觀點，考慮溪流集水區內潛在坡面型土石流之特性，可有效作為溪流型土石流發生潛勢之評估指標。

關鍵字

坡面型土石流；溪流型土石流；判釋模式；潛勢分析模式

並列摘要

This study discusses the potential evaluation method of potential debris flow torrents in central and south Taiwan based on the prone ability of hillslope type debris flow (HDF). HDFs recognition model was developed by Fisher’s discriminant analysis considering 35 HDFs and 35 landslides occurred during Typhoon Morakot in 2009. 16 HDFs and 16 landslides event were selected to be verification for the recognition model. The research shows the classification rate of samples and verification reached 91.4% and 88% of the HDFs recognition model composed by 7 factors of Watershed area, Form factor of the initiation region, Depression ratio of the initiation segment, Length of the transport segment, Ratio of landslide susceptibility area , Form factor ratio, and Gradient ratio of the initiation region. 7 factor are not only independent, but also presented a physical meaning of HDF. Furthmore, the study develops a new potential evaluation model of debris flow torrents based on the content of HDF prone units in the catchment. The HDF units were discriminated by the HDF recognition model which has developed in previously. 32 sample of potential debris flow torrents and 10 verification torrents were studied. Each catchment is divided into dozens of analysis units, and each unit is interpreted by HDF recognition analysis one by one. Results show that 3 indexs, the accumulation area of HDF units, the accumulation of initial area of HDF units, and average elevation difference of pass ratio of HDF units, are effective to be the potential evaluation index of debris flow torrents. The potential of debris flow torrent is divided into low, medium, and high potential grades. For the index of accumulation area of HDF units in the catchment, the accumulation area below 84 hectares is classified into low potential debris flow torrent , the area large than 263 hectares classified as high potential torrent, the area between both is the median potential torrent; Next for the index of accumulation of initial area of HDF units in the catchment, the accumulation area below 79 hectares is classified into low potential debris flow torrent, the area large than 189 hectares classified as high potential torrent, the area between both is the median potential torrent; Last, For the index of average elevation difference of pass ratio of HDF units in the catchment, the rate large 0.71 hectares is classified into low potential debris flow torrent, the rate low than 0.62 classified as high potential torrent, the rate between both is the median potential torrent. Summying results shows that 3 indexs, the accumulation area of HDF units, the accumulation of initial area of HDF units, and average elevation difference of pass ratio of HDF units, poss good representablity of potential evaluation on debris flow torrents comparing to actual occurrent of debris-flow events.