關鍵詞：翼手目、棲所、回聲定位叫聲、聲音監測、聲音指數
蝙蝠具有高度的物種和生態多樣性，並提供了包括植物授粉、種子傳播與控制昆蟲族群數量等具高度經濟價值的生態系統服務，再加上蝙蝠對人類活動干擾的快速反應，及對氣候等物理環境的敏感性，使其成為反映陸域生態系環境狀況的生物指數。因此，掌握蝙蝠族群分布與其動態，不管是針對基礎生態研究、蝙蝠保育、生態系統服務評估、或是指數生物監測運用，都有不可或缺的必要性。太魯閣國家公園幅員甚廣，無論是地形地貌、氣候條件與生態系類型皆複雜多變，並孕育了豐富多樣的生物相。過去對園區內翼手目的種類與分布的資訊仍相當缺乏，直到105年度中華自然資源保育協會系統性的方式調查，共記錄到5科16屬25種蝙蝠，顯示園區擁有相當高度的蝙蝠物種多樣性。然而，僅一年的資料難以全面反映園區內蝙蝠群聚與族群的動態，且適合長期蝙蝠監測的工具仍有待開發。本計畫的目標包括(1)提供太魯閣國家公園翼手目之物種名錄及時空間分布資訊、(2)提供各物種外部形值測量結果、(3)回聲定位叫聲資料收錄及初步分析、以及(4)測試聲音監測並評估聲音指數反映蝙蝠活動頻度時空間變異的能力，以作為日後發展長期蝙蝠智慧監測工具。
本研究以中部橫貫公路東段為主軸，由太魯閣閣口往西攀升至合歡山區（低至高海拔）區分成低(1100公尺以下)、中(1100-2200公尺)、高(2200公尺以上)三個海拔帶，於各海拔帶以網具捕捉、蝙蝠偵測器錄音與棲所調查，進行蝙蝠物種共四季的調查。除物種分布資料外，亦針對捕捉個體收集外部形值和回聲定位參考音頻等資料。此外，針對選取的海拔帶及棲所樣點，進行長時間錄音，除嘗試以電腦自動化分析錄音檔中的蝙蝠叫聲數量，亦計算不同聲音指數於蝙蝠叫聲頻度時空間變異的代表性。
本年度於太魯閣國家公園共記錄到5科15屬18種翼手目動物，其中8種為臺灣特有種，1種為臺灣特有亞種。與105年度資料比較，雖無新增物種紀錄，但相較上一年度的物種海拔分布資料，低海拔地區新增了毛翼管鼻蝠和東亞游離尾蝠，中海拔地區新增了臺灣葉鼻蝠的紀錄。其他例如東方寬耳蝠及臺灣長耳蝠等較稀少物種，也分別多次記錄到，顯示太魯閣國家公園在維持臺灣蝙蝠物種多樣性的重要性。棲所調查部分，與105年度相同，仍以寧安橋洞群為園區內已知臺灣葉鼻蝠和臺灣小蹄鼻蝠最重要的生殖育幼的場所，未來可作為長期族群監測樣點。此外本年度多筆蝙蝠棲所的紀錄，包括山家蝠和臺灣管鼻蝠等棲所地點不容易被觀察到的物種，都是經由通報而獲得的資料，顯示未來整合公民調查和園區蝙蝠多樣性監測的可能性。本年度共錄製了13種蝙蝠回聲定位叫聲的參考音頻，分析發現19個選用的聲音特徵可區分我們樣本中不同物種的叫聲，顯示以回聲定位叫聲在辨識太魯閣地區蝙蝠物種的潛力。聲音監測測試結果，根據洞穴監測資料，無論是蝙蝠回聲定位叫聲數量估算和聲音指數(時間熵值差值)大致能反應洞穴棲所內蝙蝠族群量月間的主要變化趨勢。聲音監測不僅能有效地增加物種調查效率，也能長期觀測族群活動量在時空間的變化。此外，與網具捕捉和棲所調查相比，非侵入式的聲音監測技術對蝙蝠的干擾相對較少，且長期監測所需的人力和資源也相對較少。未來仍可選取合適物種發展的相關監測方法，以利未來長期監測太魯閣國家公園園區內豐富的蝙蝠多樣性。
Key words：Chiroptera, roost, echolocation call, acoustic monitoring, acoustic index
Bats are characterized by high species and ecological diversity. They also provide several important ecosystem services, including pollination, seed dispersal, and insect pest suppression. Since bats usually can react to anthropogenic disturbances quickly and are sensitive to environmental changes, they have been suggested as effective bio-indicators to environmental conditions. Therefore, to secure bat diversity and the associated services to human wellbeing, it is necessary to monitor the distribution and population dynamics of bat species. With diverse terrain, climatic conditions and landscapes, the Taroko National Park supports rich mountainous fauna and flora. However, our knowledge of species diversity and the spatio-temporal variations of chiropteran fauna is still limited. To date, 25 species belong to 16 genera of 5 families have been reported by the SWAN in 2016, indicating a rich bat fauna in the park. Therefore, it is essential to develop adequate survey tools to further investigate the dynamics of bat populations and communities. In the present project, we aim to (1) continuously surveying bats for a more comprehensive understanding of chiropteran species diversity and the spatio-temporal patterns of the species distribution, (2) acquiring species trait data, (3) collecting and understanding the echolocation call diversity, and (4) evaluating the potential of acoustic technique and acoustic indices in estimating spationtemporal variation of bat activities.
We surveyed bats along the eastern section of the Central High Way at low (<1100 m asl), medium (1100~2200 m asl), and high (> 2200 m asl) elevation ranges for four seasons. Multiple techniques were used, including mist netting, harp trapping, bat detectors, roost surveys, and interviewing local people, to maximize our understanding of bat diversity. Three external traits and releasing calls were collected from captured individuals. We also set up stationary bat detectors at 3 forest sites along the elevation gradient and 2 roost sites to monitor bat activities for at least one week in each survey season. Bat calls, call features, and 6 acoustic indices were extracted and measured from the sound recordings automatically using R language-coded tools. In this year, we recorded 18 bat species, including 8 species and 1 subspecies endemic to Taiwan. We also got the first occurrence record of Harpiocephulus harpia, Tadarida insignis in the low elevation region, the first record of Hipposideros armiger in the middle elevation region, and records of two rare species, namely Plecotus taivanus and Barbasetella darjelinensis in the park. Among all roosts surveyed, the NingAng Bridge tunnels were the most important reproduction sites for H. armiger and Rhinolophus monoceros in the park. Noteworthy, most our records of roosts of Pipistrellus montanus and Murina puta were reported by national park staff and local people, suggesting the potential of citizen science in gathering wildlife species occurrence information. We found the 19 selected call feathers are effectively in distinguishing calls from 13 species in our collection. For acoustic monitoring, we found both call abundance and acoustic indices varying among sites and seasons for all sampling sites. Based on the roost site data, call abundance and values of the temporal entropy index had significant correlations with the monthly counts of bat individuals inside the roosts. Our results show that acoustic techniques are potentially effective in monitoring the diversity and abundance of bats in Taroko National Park.