Bioacústica do golfinho-comum-de-bico-curto delphinus delphis no Atlântico Sul Ocidental

Abstract

The short-beaked-common-dolphin Delphinus delphis has a characteristic yellowish lateral color and gray flank and is found in tropical and temperate waters in both coastal and oceanic regions. Individuals are very active and highly communicative. Dolphin repertoires are strongly diversified, comprising whistles and pulsed sounds (clicks and burst sounds). Many studies have been described for D. delphis around the world, although none comprises the slope region of the Western South Atlantic Ocean (WSAO). This study presents the first characterization of the acoustic emissions of short-beaked common dolphin in the WSAO, besides analyze the complexity of whistles and pulsed sounds repertoire as a whole. Intraspecific whistle variations were analyzed considering group size, contours and acoustical parameters. No significant correlation was found between number of whistles and group size (p > 0.322). Upsweep, downsweep and constant were the most frequent contours. A positive association between downsweep contour and group size was detected (p > 0.0403). Whistle frequencies ranged from 1.34 to 35.23 kHz and duration varied from 0.11 to 2.16 seconds. Comparing encounters between years through the Bonferroni method, minimum frequency, duration, and frequency range were significantly different. A C5.0 decision tree indicated that whistles from 2014 and 2015 were correctly classified in 67.0% and 54.0%, respectively. The nMDS grouped the Atlantic and Pacific oceans based on inflection points and duration, further corroborated through PERMANOVA (p = 0.039). Clicks are broadband pulses whose frequencies range from 10 and 150 kHz. The analysis of pulsed sounds took into account echolocation clicks, burst, and short burst sounds. Despite the fact that echolocation clicks are relatively well documented, there is no classification consensus regarding the wide variety of short pulsed signals. K-means analysis resulted in four clusters, whose classification, based on time parameters, exhibited the highest accuracy (87% total) explained by the first two dimensional-axis (Dim1 = 70.2%, and Dim2 = 17.2%). Validation analysis showed that time-frequency dataset was the best classification model for pulsed sounds in D. delphis (Accuracy = 84.6%; confidence interval CI = 65.1%, 95.6%; p < 0.01). This result emphasizes the influence of certain parameters in that classification, possibly reflecting the energy cost optimization for sound production, considering the animal as an acoustical identity. The last topic of this study identified possible whistle contours patterns produced during the Mass- stranding events (MSE) of D. delphis documented on the Southeastern coast of Brazil. Whistle contours were classified into nine categories. The most frequent contours were upsweep (35.2%), prevailing upsweep (29.6%) and concave (9.9%). Of the total, 33.8% contained steps, and 14.1% contained breaks. A Kendall W test showed no concordances among the whistle contours from MSE compared to the results from the WSAO region. These results suggest that a distinct pattern may be related to stress situations, reinforcing the concept that each contour category may carry different information, thus applied in distinctive behavior situations. Dolphins are primarily dependent on sound for living. Studies comprising the complexity of dolphin repertoires and factors that influence their structure are essential in maintaining population diversity and species conservation