臺灣海域鳥類船隻穿越線調查之密度估算方法探討

Estimating density of birds at sea with vessel line-transect in Taiwan waters

要評估離岸風場對鳥類的影響，標準化而定量的調查海上鳥類是基本的一步。以船隻行駛穿越線進行目視偵測是海上鳥類調查最常用的方法之一；不過分析穿越線調查的資料有兩項挑戰，一是鳥類被偵測的機率隗其至穿越線的距離而異，二是海上有許多鳥類持續在飛行中，且速度相對於船隻快許多。為了因應這兩個問題，歐洲海鳥調查準則採用固定寬度的帶狀穿越線，假設飛行中的鳥類在此距離內（通常採300m）可以完全被偵測，並對飛行中的鳥類採用快照式記錄法以避免速度的效應。而由於臺灣海域的海鳥密度較低，國內學者在「臺灣保育類海鳥開發衝擊丙應措施評估計畫」中指出，在臺灣海域宜採用不定寬度、連續記錄的穿越線調查法，以得到較完整的資料。對於不定寬度穿越線的資料，鳥類偵測率隨距離的變化可用距離採樣法分析有效帶寬（effective strip width, ESW）來處理；不過連續記錄將無法避免速度效應，會導致鳥類密度估值偏高。本研究分析在臺灣海峽進行不定寬度、連續記錄之船隻穿越線調查的結果，發現若只以300m進行固定寬度穿越線時，會漏失約22%的資料，因此支持使用不定寬度穿越線法；我們並依距離採樣法導出主要鳥種或類群的偵測函數與ESW值，發現在300m內各類群被偵測到的比例介於25%-75%間，顯示假設300m內偵測率為並不合適；另以模擬的方式，顯示鳥類相對於船隻的飛行速度可導致密度估值偏高2-5倍，並據此推估校正係數，作為計算海上鳥類密度的參考。

Vessel line-transect survey is a common method for estimating the abundance of birds at sea. In the European Seabirds at Sea (ESAS) guideline, flying birds are recorded with a snapshot method within a fixed-width strip, usually up to a distance of 300 m. However, bird density in the Taiwan Strait is much lower compared to European seas, such a method may result in considerable loss of data. Previous researchers have thus suggested that a conventional line-transect method with continuous observation would be more suitable for Taiwan. The decreasing detectability of birds with distance can be corrected by distance sampling. The continuous observation is not without disadvantage as the movement of birds can cause substantial bias. In this study, with data collected using conventional line-transect, we showed that a 300 m strip-transect would collect 22% fewer data compared to a line-transect. We analyzed the detection function of major bird species or species groups. Within a distance of 300 m, the overall detection probability was between 25% and 75%, indicating that assuming a complete detection within 300 m would underestimate the bird density. We also used simulation to quantify the bias caused by the movement of birds, and suggested correction factors for estimating bird density.