Underwater sound propagation makes it challenging to determine which individual produces specific vocalizations in baleen whale groups. This creates obstacles for behavioral studies and passive acoustic monitoring for population assessments, particularly in cue-counting methods where individual cue rates are crucial for abundance estimation. This study evaluates the potential of high-resolution accelerometers in DTAGs to detect body vibrations associated with sound production in baleen whales, offering a pathway toward individual caller identification. Using data from a singing adult male humpback whale tagged in Western Australia, we analyzed 8,648 song units. Calls were classified into different sound types and grouped into clusters based on acoustic features. For each call, acoustic parameters were extracted from hydrophone and tri-axial accelerometer data filtered up to 500 Hz. Accelerometer-detectable body vibrations were present in 64% of calls, with tonal types showing the highest detection rates. Signal presence correlated positively with received level and signal-to-noise ratio, and calls with lower peak frequencies exhibited stronger body-borne vibrations. Vibrations were primarily recorded on the X and Y axes, possibly reflecting tag placement. These findings suggest that while accelerometers alone may not yet provide a fully reliable method for caller identification, they capture meaningful vibrational patterns that can enhance multi-sensor approaches in acoustic studies and inform conservation strategies.
Keywords: Baleen whales; humpback whales; accelerometer-based call detection; body-borne vibrations; individual caller identification; passive acoustic monitoring; biologging; marine bioacoustics.
promotor/supervisor feedback
nothing yet
