Biological, Earth & Environmental Sciences Centre for Ecosystem Science

Narran Lakes acoustic recorder data

Initial data has been retrieved from acoustic recorders installed in August 2024 at Narran Lakes (Dharriwaa) as part of our Lower Balonne Flow-MER waterbird studies.

Personalise
Two people standing in a wetland landscape. Both look down and are inputting into electronic devices.

Sounds of Narran

Listen to a snippet of data recorded as part of the acoustic monitoring.

Some of the bird calls heard during the recording include little grassbirds, black swans, Eurasian coots, pied stilts (a flock flying over) and Australian reed warblers.

Initial data has been retrieved from  as part of the University of New South Wales (UNSW) Centre for Ecosystem Science’s (CES) Lower Balonne (LB) Flow-MER waterbird studies.

Around two weeks of data from when the recorders were installed on 14 August through to 5 September, was retrieved from three acoustic recorders installed at Narran Lakes' Back Lake, Long Arm and South Arm.

The team then used Cornell University’s BirdNET artificial intelligence models with coding and R (a programming language used for data analysis) to run the analyses and provide the preliminary findings.

A total of 190 species were identified with a >50 per cent confidence in accuracy. Of these, the presence of 138 species is supported by observational records from past surveys and historical records.

The diversity at each site was similar. Numbers at each site where:

  • Back Lake - 129 species, 15 unique to this site
  • Long Arm - 131 species, 17 unique to this site
  • South Arm - 145 species, 25 unique to this site.

The team will now undertake further analysis including listening to call records for those species that were identified as questionable.

Over time, with each additional data retrieval, the researchers will be able to study changes in species composition at each site in relation to the season, weather conditions and water abundance.

The amount of data that will be collected during the project is expected to be so vast that UNSW’s High-Performance Computing system will need to be used for the analysis.