Underwater soundscape indicates low anthropogenic influence around two sub-Antarctic islands
Data files
Nov 06, 2025 version files 5.09 MB
-
baleen_processed_data_1hr.csv
1.47 MB
-
killers_processed_data_1hr.csv
794.48 KB
-
noise_processed_data_1hr_FINAL.csv
2.82 MB
-
README.md
6.94 KB
Abstract
Contributions and effects of anthropogenic activities on the underwater soundscape of the sub-Antarctic regions remain poorly studied. Over 21 months (April 2021 to December 2022), we recorded underwater noise levels amid two sub-Antarctic Prince Edward Islands (PEIs) within an offshore marine protected area (MPA) to quantitatively investigate the sources of underwater noise and the impact of such noise on the detectability of marine mammal vocalizations. We measured underwater noise levels within the low (20–120 and 121–800 Hz), medium (801–25000 Hz), and high (25001–48000 Hz) frequency bands. Wind speed was the primary predictor of low and medium-frequency underwater noise levels, whereas iceberg volume was the primary predictor at the high-frequency band. Probabilities of detecting vocalizations of Antarctic blue, fin, humpback, Antarctic minke, and killer whales decreased with increasing noise levels. On the contrary, probabilities of detecting sei and Madagascan pygmy blue whales increased with noise levels. Overall, these novel results indicate that geophonic noise dominates the underwater soundscape of the PEIs in the absence of intense anthropogenic activities such as marine traffic, and that conservation efforts through this MPA might be protecting this remote region from anthropogenic activities.
Dataset DOI: 10.5061/dryad.9s4mw6mtf
Description of the data and file structure
Acoustic data were collected between two sub-Antarctic islands, Marion Island and Prince Edward Island, Southern Ocean using a SoundTrap deployed 5 m off the seafloor. Marine mammal acoustic occurrence and underwater noise levels were determined from these collected data.
Files and variables
These files contain the formatted data on underwater noise levels, baleen and killer whale acoustic occurrence around the Prince Edward Islands, to reproduce the analyses presented in the paper.
killers_processed_data_1hr.csv: Contains data on the hourly acoustic presence of killer whale vocalizations and noise levels at different frequency bands for the period April 2021 to May 2022 used in the random forest model analysis.
- Year: The year when the data was collected.
- Month: The month when the data was collected.
- Day: The day when the data was collected.
- Hour: The hour (UTC) when the data was collected.
- SOCPA: Hourly presence and absence of killer whale social calls indicating acoustic occurrence and communication of dolphins. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- FORPA: Hourly presence and absence of killer whale foraging sounds (i.e. echolocation) indicating acoustic occurrence and foraging of dolphins. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- hz20_120: 20–120 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz121_800: 121–800 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz801_25000: 801–25000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz25001_48000: 25001–48000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
baleen_processed_data_1hr.csv: Contains data on the hourly acoustic presence of six baleen whale vocalizations and noise levels at different frequency bands for the period April 2021 to December 2022 used in the random forest model analysis.
- Year: The year when the data was collected.
- Month: The month when the data was collected.
- Day: The day when the data was collected.
- Hour: The hour (UTC) when the data was collected.
- dpa: Hourly presence and absence of blue whale D-calls indicating acoustic occurrence of foraging activities of blue whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- pa40Hz: Hourly presence and absence of fin whale 40 Hz pulses indicating acoustic occurrence of foraging activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- pa20Hz: Hourly presence and absence of fin whales 20 Hz pulses indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- madpa: Hourly presence and absence of Madagascan pygmy blue whale calls indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- antpa: Hourly presence and absence of Antarctic blue whale Z-calls indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- humpa: Hourly presence and absence of humpback whale songs indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- minpa: Hourly presence and absence of Antarctic minke whale bioduck calls indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- seipa: Hourly presence and absence of sei whale upsweep calls indicating acoustic occurrence of communication activities of whales. The value is given as 0 or 1, where 1 means acoustic presence and 0 means acoustic absence.
- hz20_120: 20–120 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz121_800: 121–800 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz801_25000: 801–25000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz25001_48000: 25001–48000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
noise_processed_data_1hr_FINAL.csv: Contains data on the hourly acoustic noise levels at different frequency bands for the period April 2021 to December 2022 used in the random forest model analysis.
- Year: The year when the data was collected.
- Month: The month when the data was collected.
- Day: The day when the data was collected.
- Hour: The hour (UTC) when the data was collected.
- hz20_120: 20–120 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz121_800: 121–800 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz801_25000: 801–25000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- hz25001_48000: 25001–48000 Hz frequency band noise level measured as equivalent continuous sound pressure level in dB re 1 μPa.
- windsp: Absolute wind speed in m s-1.
- wdn: Absolute wind direction in degrees North.
- adt: Absolute dynamic topography which represents the height of the sea surface relative to the Earth's geoid in m (Please note that this variable is not used in the current study.).
- wave_height: Wave height indicates the at-sea observed total wave field in m.
- wave_period: Wave period in seconds (s) representing the time when majority of the wave energy is concentrated.
- wave_dir: Wave direction - variable is not used in this study.
- tp: Total precipitation (m) consisting of accumulated liquid and frozen water from rain and snow that falls to the ocean's surface.
- current: Ocean current (cm s -1) used as frontal activities proxy around the mooring location.
- total_giga_tons: Iceberg (less than 3 km in length) volume measured in total gigatons.
- vnumber: Weighted number of vessels up to 53 km.
Code/software
All data can be viewed in MS Excel or R statistical software.