Automatic detection and annotation of eastern Caribbean sperm whale codas.

Area of Science:

• Marine Biology
• Bioacoustics
• Animal Communication

Background:

• Sperm whale (Physeter macrocephalus) monitoring relies on identifying their unique communication signals, known as codas.
• Existing methods for coda identification are limited, hindering comprehensive analysis of sperm whale social structures and communication.

Purpose of the Study:

• To develop and present the first automatic detector and annotator for sperm whale codas.
• To utilize this tool to gain new insights into the complexity of sperm whale communication.

Main Methods:

• Development of an automatic coda detector and annotator employing graph-based clustering.
• The clustering algorithm leverages the inherent similarity between clicks within a coda.
• Validation of the system's performance in detecting and annotating codas at low signal-to-noise ratios.

Main Results:

• Successful detection and accurate annotation of sperm whale codas, even in noisy environments.
• Effective separation of codas from echolocation clicks.
• Discrimination between codas from multiple, simultaneously vocalizing whales.
• Identification of novel coda types and analysis of their distribution across different whales and time periods.
• Evidence of synchronized communication, indicated by shared coda types and transmission times among whales.

Conclusions:

• The developed automatic system significantly advances sperm whale bioacoustic monitoring capabilities.
• Sperm whale communication exhibits a high degree of complexity, including novel signal types and synchronized behaviors.
• The findings provide a foundation for future research into the intricacies of cetacean social interactions and communication systems.