Abstracts Submitted: 250
Number of Users: 295
View Abstracts Submitted
Back to home Page
The Indian Ocean humpback dolphin (Sousa plumbea), a common cetacean species in India has a diverse vocal repertoire, which to date has not been described in detail. Our study presents the first description and classification of their whistle repertoire recorded off the Sindhudurg coast, Maharashtra. Acoustic recordings of humpback dolphins were made on 11 separate sampling occasions in the months of May and October 2014, December 2015 and January 2016. A total of 2,260 whistles with high SNR (>6dB) were selected for further analysis. Whistles were manually traced using the Silbido contour tracing algorithm in MATLAB (Roch et al. 2011). A custom written script was used to extract 16 time-frequency based features from the whistle contours. Whistles spanned a wide frequency range between 2.3 kHz and 33.0 kHz, with durations ranging from 0.01 s to 1.60 s. Whistles were first broadly categorized into seven contour classes based on the aural-visual classification method. A classification and regression tree (CART) analysis was then used to quantify variability between the seven contour classes using the measured acoustic features. CART resulted in an overall classification accuracy of 89.5%. Based on the CART analysis, frequency gradient, minimum slope, maximum slope, beginning and ending slope accounted for most variability between whistle types. Comparison between our study and whistles of Sousa spp. from China, Australia, Malaysia, Angola, and Pakistan revealed variation in time-frequency parameters and will be discussed.
© Copyright 2017 All Rights Reserved