在光達數值地形模型上解釋山崩分佈的地形相關因子之最適量測尺度;Optimal Measurement Scale on Lidar Dtm for Each Topography-Related Factor in Interpreting Landslide Distribution

NCU Institutional Repository > 地球科學學院 > 應用地質研究所 > 研究計畫 > Item 987654321/84769

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/84769

題名:	在光達數值地形模型上解釋山崩分佈的地形相關因子之最適量測尺度;Optimal Measurement Scale on Lidar Dtm for Each Topography-Related Factor in Interpreting Landslide Distribution
作者:	李錫堤
貢獻者:	應用地質研究所
關鍵詞:	山崩潛感分析;潛感因子;最適量測尺度;landslide susceptibility analysis;susceptibility factor;optimal measurement scale
日期:	2020-12-08
上傳時間:	2020-12-09 10:54:53 (UTC+8)
出版者:	科技部
摘要:	山崩潛感分析中的許多潛在因子是直接或間接取自地形資料，例如：坡度、粗糙度、曲率等地形因子，或取自與距離尺度有關的各種區位因子，或結合地形資料去做內插的雨量分布或震度分布等促崩因子。用不同尺度量測及運算出來的因子會有不同的輸出值，且並非高解析度地形資料就可以直接得到高效度的因子。對一個特定區域的特定目的而言，每一個因子都會存在一個最適量測尺度。光達數值地形模型是目前可取得的精確度最高且解析度最高的地形資料，其解析度高達1m。但過去的研究及本團隊的經驗上了解，使用高解析度的數值地形模型並不能獲得更好的分析成果，而有須要做每一個山崩潛感因子的最適量測尺度的考量。若能取得最適尺度的因子做分析，再加上光達數值地形模型固有的精確度，則必能建立出更好的山崩潛感模型。本計畫團隊去年在石門水庫上游集水區及曾文水庫上游集水區兩個不同地區及採用各四個代表性颱風誘發山崩資料，分析各山崩潛在因子在不同量測尺度下解釋山崩分佈的效度。研究過程中了解到因子的效度深受數值地形模型精確度及山崩測繪精確度的影響，因此本年度擬將重點放在選用高精度光達數值地形模型及重新在高解析度衛星影像上測繪與仔細檢查每一個降雨誘發山崩的多邊形物件，期能明確找到每場降雨誘發山崩事件中各個因子的最適量測尺度，並以最佳效度因子建立每個降雨事件的山崩潛感模型，比較由最佳效度因子建立的模型相對於原始因子建立的模型其成功率之增長情形暨以交叉驗證了解預測率之增長情形。研究區域仍選擇在曾文水庫上游集水區，但降雨事件則挑選賀伯颱風、桃芝颱風、敏督利颱風、20050615豪雨、20060609豪雨、莫拉克颱風、20110718豪雨、20120610 豪雨、20150523豪雨共九期獨立降雨事件做因子最適量測尺度及山崩潛感模型成功率及預測率增長情形的探討與比較。期藉由高精度數值地形模型及不同大小之降雨誘發山崩資料來精緻化因子最適量測尺度的選取，做出精緻化的事件山崩潛感模型，並將結果分享於世。 ;Many causative factors for landslide susceptibility analysis are directly or indirectly derived from topographic data. These include topographic-related factors, like slope gradient, topographic roughness, curvature, etc., location-related factors, like distance to road which may be counted from different distance scale, and triggering factors, like rainfall value and earthquake intensity which are interpolated by cokriging with topographic data. A factor calculated from topographic data of different measurement scale will output a different value, and high resolution data is not always directly outputting an effective factor for interpreting landslides. Each factor may exist an optimal measurement scale which can best interpret the landslide distribution at a specific area and for a specific purpose. The LiDAR digital terrain model is the highest accuracy and highest resolution terrain data currently available. The resolution is as high as 1m. But we know from past research and the experience of this research team, using high-resolution digital terrain models does not yield better analysis results. It is necessary to consider the optimal measurement scale of each landslide causative factor. If we can get the best scale factor for analysis, and coupled with the inherent accuracy of the LiDAR digital terrain model, it is bound to build a better model of landslide susceptibility.In the last year project proposed by this team, we have selected the Shihmen Reservoir catchment area and the Zengwen Reservoir catchment area as two study areas and select four Typhoon-triggered landslide inventories at each catchment to analyze the scale effect and to determine the optimal measurement scale of each susceptibility factor. During the study, we learned that the validity of the factors is greatly affected by the accuracy of the digital terrain model and the accuracy of landslide mapping. Therefore, this year it is planned to focus on the selection of high-precision LiDAR digital terrain model and re-mapping and careful inspection of each polygonal object induced by rainfall in high-resolution satellite image. Hope we can clearly find the best measurement scale of each factor in each rainfall-induced landslide inventory. And establish the landslide susceptibility model of each rainfall event with the best validity factor. Compare the success rate of the model established by the best validity factor to the success rate of the model established by the original factor. We also use cross-validation to understand the growth of prediction rate.The study area is still selected in the upper catchment area of Zengwen Reservoir, but for the rainfall events, typhoon Hebe, typhoon Toraji, typhoon Mindulle, 20050515 heavy rain, 20060609 heavy rain, typhoon Morakot, 20110718 heavy rain, 20120610 heavy rain, 20150523 heavy rain, totaling 9 independent rain events will be used to study optimal measurement scale for each causative factor, and to compare of the success rate and prediction rate of the landslide susceptibility models. By using high-precision digital terrain model and rainfall-induced landslide data of different sizes to refine the selection of the best measurement scale of factors, a refined event landslide susceptibility model is made. We will share the results with the world.
關聯:	財團法人國家實驗研究院科技政策研究與資訊中心
顯示於類別:	[應用地質研究所] 研究計畫

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	188	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....