圈養瓶鼻海豚 (Tursiops spp.) 在社交與打樁噪音情境變化下的發聲行為反應

Abstract

瓶鼻海豚（\\textit{Tursiops} spp.）仰賴聲音作為在海中溝通的主要方式，因此了解牠們的聲學行為如何隨情境改變，便能以干擾最小的方式評估動物所面對的處境。本研究的目的有二：其一為透過長期追蹤一隻成年雌性印太瓶鼻海豚（\\textit{T. aduncus}）38個月，測量可歸屬於該個體之哨叫聲類型（PW1）的輪廓特徵，以及各情境下海豚的發聲量，來瞭解該個體在八個不同社交情境下的聲學行為變化。其二則為將三隻瓶鼻海豚（\\textit{T. truncatus}）曝露於不同音量與時長的打樁噪音，並測量同樣兩類變數來瞭解海豚群在不同噪音情境下的聲學行為變化。 於社交情境的分析中，於個體與群體隔離之水池的情境，整池的哨叫聲量皆受到壓抑，而此壓抑在回到能夠接觸海豚群體聲音的水池時便會解除。PW1 的單位時長在哺育親生幼豚的時期長於與群體生活的時期，但在與另一母子對共處的時期並未出現相同現象。可歸屬於個體的哨叫聲與整體發聲量因而具有作為圈養動物福利監測指標的潛力，能在目視行為出現變化之前即探知個體狀態的改變。 於打樁噪音情境的分析中，海豚的聲學行為對曝露時長、噪音大小，以及兩者的交互作用出現不同的反應。哨叫聲數量隨曝露時長上升，脈衝聲數量則同時下降，兩者方向相反，使合計的總發聲量掩蓋了各自的變化。V 型哨叫聲的頻率上移，並僅在最大的噪音位準下縮短；正弦波型哨叫聲則產生更多而各自更短的迴圈，總時長維持不變。在平均接收峰值聲壓級（mean received $L_\\mathrm{peak}$）為 130 dB re 1 $\\mu$Pa 的條件下即已出現聲學行為的變化，而同一批動物在同一場次的目視行為在該聲壓級並無改變。此一聲壓級亦低於文獻中對打樁噪音更為敏感的港灣鼠海豚（\\textit{Phocoena phocoena}）出現可見行為改變的聲壓級。 綜合兩項研究，兩個不同的設計指向同一件事：在兩者之中，聲學行為的改變都在動物的可見行為出現任何變化之前就已可被偵測到。由於不同訊號類型與不同參數的變化方向並不一致，監測方案必須同時讀取數個測度，而不能只依循其中任何一個。本研究的部分錄音取自運動相機的影片音檔，顯示此類器材具有推廣至缺乏專業設備的圈養機構與野外場址的潛力。

Bottlenose dolphins (\\textit{Tursiops} spp.) depend on sound as their primary means of communication at sea, so how their acoustic behaviour changes across contexts offers a way to assess what an animal is facing with minimum disturbance. This thesis had two aims. The first was to establish how the acoustic behaviour of one adult female Indo-Pacific bottlenose dolphin (\\textit{T. aduncus}) changed across eight social contexts, by tracking her over 38 months and measuring the contour of a whistle type attributable to her (PW1) together with the number of calls dolphins produced in contexts. The second was to establish how the acoustic behaviour of a group changed across noise contexts, by exposing three bottlenose dolphins (\\textit{T. truncatus}) to pile-driving noise at different levels and for different durations and measuring the same two classes of variable. In the social contexts, whistling was suppressed in every context that placed the focal animal in an isolated pool separated from the rest of the population, and the suppression lifted whenever she returned to a pool within acoustic reach of the others. The duration of a single loop of PW1 was longer while she reared her own calf than during the periods spent with the group, whereas no comparable change appeared while she was housed with another mother-calf pair. Whistles attributable to an individual, together with overall vocal activity, therefore carry potential as indicators for welfare monitoring in managed care, since they can register a change in an animal's state before that change appears in its visible behaviour. Under pile-driving noise, the acoustic behaviour of the dolphins responded differently to the exposure period, to the noise level, and to the interaction between the two. Whistle counts rose with the duration of exposure while pulse counts fell over the same period, so that the combined total concealed both changes. The V-shaped whistles moved to higher frequencies and shortened at the loudest level alone, whereas the sinewave whistles produced more loops that were each shorter while the total call length held. Changes in acoustic behaviour appeared at a mean received peak sound pressure level ($L_\\mathrm{peak}$) of 130 dB re 1 $\\mu$Pa, at which the visible behaviour of the same animals in the same sessions did not change. That level lies below the one at which changes in visible behaviour have been reported in the harbour porpoise (\\textit{Phocoena phocoena}), a species more sensitive to pile-driving noise. Taken together, the two studies converge on one point from different designs: in both, a change in acoustic behaviour was detectable before any change was visible in what the animals did. Because the direction of change differed between signal types and between parameters, a monitoring scheme would have to read several measures together rather than follow any one of them alone. Part of the recordings analysed here were taken from the audio tracks of action-camera footage, so equipment of this kind could be extended to facilities and field sites without specialist instruments.