地層下陷區域之淹水潛勢與脆弱度評估

Abstract

台灣西南沿海區域地層下陷情況嚴重，其中雲林地區下陷面積為全國最大。近年來地層下陷之中心漸漸往內陸移動。在全球氣候變遷下，地層下陷不僅增加災害脆弱度，亦使更多區域暴露在颱洪災害之威脅下，為能合理應用資源於改善措施上，降低脆弱度，減少颱洪災害損失，因此須瞭解受災區域之脆弱程度。 本研究應用區域為雲林北部，首先蒐集降雨與潮位資料，雨量分析採用雲林北部沿海之雨量站資料，潮位分析採用麥寮潮位站資料。得到降雨與潮位之機率分佈後，利用拉丁超立方取樣法對降雨與潮位之累積機率分佈取樣，將取樣資料做為一維SOBEK模式之邊界條件，模擬渠道最高水位，探討潮位對於渠道水位之影響。 為了瞭解地層下陷與淹水潛勢間之關係，將蒐集到之地層下陷觀測數據利用克利金法作空間內插，製作出地層下陷量之等高線圖，套疊至原雲林北部數值高程模型，得出新的數值高程模型。使用前面水理分析方法得出之25年渠道水位作為設計堤高，將不同重現期洪水帶入二維SOBEK淹水模式，模擬在不同數值高程模型各重現期洪水之淹水面積，將模擬結果加入脆弱度指標，分析研究區域內各鄉鎮市之脆弱程度。

Landsubsidence has become worse in southwest coastal areas of Taiwan; in which, Yunlin is one of the most severe subsidence county in Taiwan. Recently, the center of subsidence cone has gradually migrated inland. Under the global climate change, the subsidence not only increases the vulnerability of disaster, but also threatens more area with flood disasters. For allocating reasonable resources on improvement work, reducing the vulnerability, and decreasing flood disaster losses, therefore investigating the level of vulnerability in disaster areas is necessary. The study area is located in the Northern Yunlin Country. First, the observation data of rainfall and sea levels were collected. Data from Northern Yunlin coastal rainfall stations and Mailiau station are adopted for frequency analysis. Latin Hypercube sampling (LHS) method is later applied to produce samples of rainfall and sea level, which are treated as the external forcing and boundary condition of SOBEK 1D model. The simulation results are used to evaluate the distributions of channel peak water level, and to investigate the effect of rainfall and sea levels to channel water levels. To understand the relationship between the subsidence and flooding potential, we use the subsidence data by kriging method for spatial interpolation, creating contour maps of the subsidence. Combining the created contour map with the Northern Yunlin digital elevation model, we may produce a new digital elevation model. Levee protection levels based on 25 years channel water level probabilistic distribution are taken into account for evaluating the inundation area by SOBEK 1D2D model, taking account several flood return periods with different digital elevation model. The vulnerability index is added to simulation results, and used to evaluate the vulnerability for each town within the study area.