Understanding when, where, and how often animals vocalize is essential for understanding animal communication and informing passive acoustic monitoring. Biologging tags are often used to study the vocal behavior of cetaceans, but efficient underwater sound transmission means that tags record sounds of both the tagged animal and the nearby conspecifics. In low-frequency baleen whales, vocalizations from the tagged animal have been linked to signals recorded on the tag accelerometers. Here we evaluate whether higher sample rate accelerometers in newer biologging tags can identify calls from tagged humpback whales. For a singing humpback whale in Western Australia, accelerometer signals were detected in 60% of calls, with tonal calls having higher detection rates than pulsed call types. Calls with higher output level and lower peak frequency had stronger accelerometer signals. In contrast, feeding humpback whales on Stellwagen Bank primarily use lower amplitude social calls. Only 3% of calls were detected on accelerometers, and these were consistently calls with a high output level. Higher activity levels during foraging increased accelerometer noise by more than 30 dB, increasing the probability that calls were masked. Our results show that call type, amplitude, frequency, and behavioral state influence whether accelerometers will detect vibrations associated with sound production.
Shearer et al. (Wed,) studied this question.