Related Experiment Video
Updated: Aug 4, 2025
09:32
Development of New Methods for Quantifying Fish Density Using Underwater Stereo-video Tools
Published on: November 20, 2017
9.3K
Biodiversity assessment using passive acoustic recordings from off-reef location-Unsupervised learning to classify
Vasudev P Mahale1, Kranthikumar Chanda1, Bishwajit Chakraborty1
1Council of Scientific & Industrial Research, National Institute of Oceanography, Dona Paula, Goa 403 004, India.
The Journal of the Acoustical Society of America
|March 31, 2023
Summary
Passive acoustic monitoring in the Zuari estuary revealed diverse fish choruses and invertebrate sounds. Acoustic indices like ACI and AEI effectively indicated high fish biodiversity, aiding in real-time fish stock assessment.
Area of Science:
- Marine Bioacoustics
- Ecosystem Soundscape Analysis
- Quantitative Ecology
Background:
- The acoustic environment of marine ecosystems provides crucial insights into biodiversity and ecosystem health.
- Grande Island in the Zuari estuary is a dynamic habitat requiring detailed acoustic characterization.
Purpose of the Study:
- To quantitatively characterize the off-reef acoustic environment of Grande Island during the pre-monsoon season.
- To assess fish biodiversity and identify sound-producing species using passive acoustic monitoring.
- To develop and evaluate an unsupervised classification method for fish sound types.
Main Methods:
- Deployment of passive acoustic recorders to capture underwater soundscapes.
- Analysis of acoustic data using oscillograms and segmentation to identify fish calls (Sciaenidae, Terapon theraps) and invertebrate sounds (snapping shrimp).
- Calculation of biodiversity indices: Acoustic Evenness Index (AEI) and Acoustic Complexity Index (ACI), and Sound Pressure Level (SPLrms) across different frequency bands.
- Application of unsupervised classification using Principal Component Analysis (PCA) and K-means clustering for fish sound identification.
Main Results:
- Prominent fish choruses and invertebrate sounds were detected, including specific vocal groups.
- ACI and AEI metrics effectively indicated increased fish biodiversity in both low and high-frequency bands.
- SPLrms showed significant variations across the analyzed frequency bands.
- Unsupervised classification achieved high accuracy (89.84% overall) in identifying four dominant fish sound types.
Conclusions:
- Passive acoustic monitoring is a valuable tool for characterizing marine soundscapes and assessing biodiversity.
- Acoustic indices (ACI, AEI) provide robust indicators of fish species richness and distribution.
- The hybrid PCA and K-means clustering approach demonstrates significant potential for real-time fish stock monitoring.
