Browsing by Subject "bioacoustics"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Open Access
    Acoustic occurrence patterns of Prince Edward Islands' killer whales on diel, monthly, and seasonal scales
    (2026) Daniels, Robyn Kayla; Shabangu, Fannie; Lamont, Tarron
    Killer whales are highly social marine mammals which utilise their vocalisations for communication, feeding and navigation. They are poorly studied within the Southern Hemisphere, particularly in remote regions such as the sub-Antarctic Prince Edward Islands (PEIs) where fieldwork can be logistically challenging. This region is subject to rapid rise in temperatures and subsequent climate variability potentially causing ecosystem changes. This study therefore aims to investigate killer whale presence using two modalities (acoustic as primary and visual as validation), two methods of acoustic detection and a variety of potential influences, both biotic (primary production and killer whale prey) and abiotic (wind, sea surface height (SSH), sea surface temperature (SST)) factors across multiple temporal scales. A hydrophone was deployed at 46° 46.4'S, 37° 54.7' E and recorded from April 2021 to April 2023. SST, SSH, chlorophyll-a concentration (chl-a) and wind speed data were sourced from satellite and reanalysis data, and prey and sightings data from land-based surveys. Automated call detection was used to investigate these objectives and compared to manual detection. Results indicated that killer whales' calls were detected intermittently throughout the year with a higher occurrence of calls in the spring and early-summer months and no significant year-to-year variability was observed. Their sightings displayed similar patterns, but the environmental variables and prey abundance did not indicate matching year-to-year variability. This was attributed to fine-scale changes in their social structure, long-term resistance in their call behaviours to larger temporal scale environmental anomalies and substantial short-term variability of SSH and wind speed. Random forest modelling identified wind speed and hour of the day as the primary predictors of the occurrence of calls while SST, SSH, chl-a and prey abundance were moderately important. Strong diel call patterns were observed only in winter and summer, reflecting adaptive foraging and social behaviour. Automated call detection was found to be an effective method but is not to be used in isolation as it produced false positive and negative detections. These results provide important insights into PEIs' ecosystem functioning by disentangling the effects of climate variability on keystone species such as killer whales using cost-effective methods such as passive acoustic monitoring, automated detection, and satellite and reanalysis data.