臺灣離岸風場鳥類撞擊風險模型之敏感性分析

Abstract

離岸風電為臺灣政府大力支持的再生能源開發計畫之一，然而臺灣海峽也是全球重要候鳥遷徙路線，離岸風電將有可能對遷徙候鳥與本地海鳥帶來撞擊風險之威脅。在環境影響評估中常利用Band Model撞擊風險模型推估潛在鳥類撞擊數量。由於許多輸入參數收集不易，資料上的不確定性將影響評估結果之可信度。因此，本研究針對Band Model撞擊風險模型之基礎模型Option2以及擴展模型Option3進行敏感性分析，並以臺灣海峽之過境鳥普通燕鷗（Sterna hirundo）與冬候鳥銀鷗（Larus argentatus）為例，分別以單因子敏感性分析（One At a Time）以及全因子敏感性分析（All At a Time）對提升撞擊風險模型之方向調整Band Model撞擊風險模型之輸入數值10%，並分析參數調整前後相對於未經調整之變化程度。研究結果發現Band Model撞擊風險模型對於鳥類迴避率最為敏感。在基礎模型中，當迴避率僅調降10%，普通燕鷗撞擊風險模型輸出結果相對於未經調整前改變為1024.96%，銀鷗撞擊風險變異程度更高達2545.00%。此外，扇葉掃掠範圍低點高程（非掃掠範圍輪轂高度）、風機運轉時間比例以及鳥類密度亦對於Band Model撞擊風險模型極具影響力。綜上所述，本研究建議未來應盡速對本地鳥種進行詳實之鳥類調查，並針對迴避率、鳥類飛行高度、夜間飛行活動比例等輸入建置本地鳥類行為之資料集。另建議避免使用固定之迴避率輸入數值，使環境生態影響評估能納入資料變異所造成的不確定性，力求反映臺灣海峽鳥類撞擊風險之真實情形。

Offshore wind power is one of the renewable energy development projects strongly supported by the Taiwan government. However, the Taiwan Strait is also a crucial migratory route for birds globally, and offshore wind power may pose a threat to migrating birds and local seabirds. According to environmental impact assessments in Taiwan, the Band Model has often been used to estimate the potential risk of bird collisions. However, the input parameters used in the collision risk model were often difficult to obtain, and uncertainties might affect the accuracy of the assessment results. Therefore, this study conducted a sensitivity analysis of the Band Model's basic model (Option 2) and extended model (Option 3), using Common Tern (Sterna hirundo), a transient species , and Herring Gull (Larus argentatus), a wintering species in the Taiwan Strait, as examples. This study adjusted the input parameters of the Band Model using the One At a Time (OAT) and All At a Time (AAT) methods, and analyzed the relative changes of collision risk before and after parameter adjustments compared to the output of the original model.The results of the study found that the Band Model is most sensitive to avoidance rate. In the basic model, when the avoidance rate was reduced by only 10%, the output results of the Common Tern collision risk changed by 1024.96%, and the Herring Gull collision risk changed even more dramatically by 2545.00%. Additionally, factors such as the Hub Height Add (the height of the lower tip of the turbine blade), turbine operational time, and bird density significantly impacted the Band Model's collision risk estimates. In summary, this study recommends conducting detailed bird surveys for local bird assemblage and establishing datasets on local bird behaviors, especially avoidance rates, flight heights, and nocturnal activities. Furthermore, it is advised to avoid using fixed avoidance rate, so that environmental impact assessments can incorporate the uncertainties caused by data variations, striving to reflect the true collision risks for birds in the Taiwan Strait.