揭開熱帶和亞熱帶鳥類獨特的窗殺模式與脆弱性： 來自臺灣十年的公民科學和社群媒體資料分析

Abstract

野鳥撞玻璃（Bird-window collisions），或稱窗殺，意指鳥類撞上人造玻璃建物（包含但不限於窗戶、門、鏡子等）是人為活動間接造成鳥類死亡的主因之一。窗殺是重要的保育議題，保守估計每年全球有十億隻以上的野鳥死於窗殺，然而我們現有的認知卻侷限於溫帶，尤其是北美洲地區。因此，本研究為了彌補此一缺漏，將研究範圍設定在臺灣地區─位於熱帶與亞熱帶，高度都市化的亞洲島嶼，檢視此處的窗殺概況。

為此，我們藉由公民科學建立了一個時間跨度達十年（2012-2022）涵蓋172個物種（約臺灣地區有紀錄鳥種的25%）逾3,500筆案例紀錄的窗殺資料庫。公民科學的資料來自兩個不同平台，分別為社群媒體（臉書，約佔整體的三分之二），及臺灣動物路死觀察網（TaiRON，約佔整體的三分之一）。稀釋曲線分析（Rarefaction and extrapolation analysis）顯示，本研究整併倆公民科學平台資料所形成的窗殺資料庫為可信賴，且有效涵蓋臺灣窗殺樣態的資料庫。

相較於北美洲地區窗殺以春秋遷徙季成群大量遷徙中的小型雀型目候鳥為主要撞擊者，臺灣地區的窗殺在季節分布和種群分類組成頗為不同。其一、儘管臺灣位處東亞澳遷徙線上且為重要的遷徙中繼站，遭遇窗殺的物種中留鳥的物種數多於候鳥物種數。其二、逾半曾被記錄過遭遇窗殺的物種在物種族群豐度對撞擊數量的回歸分析中，落在1:1參考線之上，顯示出牠們面臨窗殺威脅的高風險。其三、所匡列出的「超易撞物種（super-colliders）」包含鬚鴷科（Megalaimidae）、鳩鴿科（Columbidae）、鷹科（Accipitridae）、翠鳥科（Alcedinidae）等非雀形目鳥種。其中，又以五色鳥（Psilopogon nuchalis）、翠翼鳩（Chalcophaps indica）和鳳頭蒼鷹（Lophospiza trivirgata，原Accipiter trivirgatus）在撞擊數量和撞擊風險上均列為前三，為學術紀錄上首度顯示這三個物種具高窗殺風險。

綜觀本研究，我們的成果彰顯出野鳥撞玻璃的季節分布和種群分類組成具有區域獨特性。因此，我們鼓勵過去缺乏相關研究的地區，尤其是亞洲、熱帶和亞熱帶等地區展開相關研究，以因應逐漸嚴峻的國際野生動物保育挑戰。最後，本研究亦顯示公民科學和社群媒體的興盛不僅能輔助資料收集，更能促進群眾參與和落實保育行動，可供未來類似領域的研究借鑑。

Bird-window collisions (i.e. birds crashing into man-made glass structures) are a major contributor to human-induced avian mortality, with global estimates exceeding one billion birds killed annually. While this conservation issue is increasingly recognized, research has predominantly focused on temperate regions, particularly North America. To address this gap, this study examines bird-window collision patterns in Taiwan, a highly urbanized tropical and subtropical Asian island.

We compiled a decade-long database (2012-2022) of over 3,500 bird-window collision cases involving 172 species (ca. 25% of Taiwan’s recorded bird species). Data were collected through citizen science efforts, with approximately two-thirds from social media (Facebook) and one-third from the Taiwan Roadkill Observation Network (TaiRON). Rarefaction and extrapolation analysis suggested that these social media data effectively captured collision events.

Compared to North America, dominated by migratory passerines, bird-window collisions in Taiwan exhibited differences in seasonal distribution and taxonomic composition. First, in contrast to North America, resident birds experienced more collisions than migratory species, despite Taiwan’s role as a key stopover along the East Asian-Australasian Flyway. Second, nearly half of the species involved in collisions fell above the abundance-collision 1:1 reference line, indicating high vulnerability. Prominent “super-colliders” included non-passerine species from families such as Megalaimidae, Columbidae, Accipitridae, and Alcedinidae, with the Taiwan Barbet (Psilopogon nuchalis), Emerald Dove (Chalcophaps indica), and Crested Goshawk (Lophospiza trivirgata, formerly Accipiter trivirgatus) ranking highest in both collision frequency and vulnerability.

Overall, our study suggests that seasonal patterns and taxonomic compositions of bird-window collisions vary across regions. We advocate for expanded research in under-studied areas, particularly in Asian, tropical, and subtropical regions, to develop targeted conservation strategies addressing this pressing global wildlife challenge. Lastly, citizen science and social media not only enhance data collection efforts but also engage the public in conservation, offering a valuable approach for future studies in this field.