Killer whales are apex predators with distributions that vary across the world’s oceans. Their ecology and behaviour remain poorly understood in many regions due to limited research, especially in remote areas. Around the sub-Antarctic Prince Edward Islands (PEIs) in the Southern Ocean, killer whales exhibit seasonal vocalising patterns that align with prey availability and social activities.
The Problem
Manual analysis of large acoustic datasets is labour-intensive and time-consuming. Identifying the seasonal occurrence of killer whales based on their vocalisations around the Prince Edward Islands requires a more efficient approach. Automating this process would significantly reduce the effort needed to detect marine mammal calls.
The Solution
Automated detection of marine mammal calls will streamline the analysis process, reducing the time spent by nearly 75% while maintaining high accuracy. This will allow researchers to process large datasets more effectively and focus on interpreting ecological and behavioural patterns.
Impact
This project, undertaken by the University of Pretoria, will benefit from the time-saving effects of automated detection, despite the absence of dedicated funding. The insights gained from this research will improve the understanding of killer whale ecology in the Southern Ocean and help assess how environmental changes affect their behaviour and distribution.
