Preliminary Analysis and Comparison of Acoustic Signals from Delphinids in the N.W. Atlantic Ocean

Defense Date


Document Type


Degree Name

M.S. Marine Environmental Sciences

First Advisor

Edward Keith

Second Advisor

Amy C. Hirons


Cetaceans rely on sound for navigation, communication, migration and foraging. Distinct patterns of these sounds that have been utilized to differentiate between populations within species. Passive acoustic monitoring (PAM) has proven useful in locating and tracking cetaceans in combination with visual monitoring and may improve quantification of cetacean populations. Analysis of whistle characteristics can potentially be used to identify cetaceans when they cannot be visually observed due to weather or behavior such as foraging below the surface. In this study, passive acoustic recordings of cetacean vocalizations were examined statistically to determine if spectrographic parameters could be used to discriminate between nine species of delphinids including: Stenella frontalis, Tursiops truncatus, Delphinus delphis, Stenella attenuata, Globicephala macrorhynchus, Grampus griseus, and Stenella coeruleoalba.

Visual and acoustic data were collected along the east coast of the United States during a research cruise conducted by the National Oceanic and Atmospheric Administration (NOAA) from June to August 2004. The species emitting the sounds were identified in most cases by comparing visual sightings with acoustic detections. Acoustic analysis software, Ishmael, was used to quantitatively analyze delphinid whistles by measuring nine parameters of each acoustic signal including: high frequency, low frequency, central frequency, end frequency, start frequency, bandwidth, duration, number of steps and number of inflection points. Principal Components Analysis (PCA) was used to determine if these nine whistle parameters could be used to reliably discriminate between species and to potentially identify whistles from animals not detected visually.

No significant differences in whistle characteristics were identified and therefore these characteristics could not be used to identify unknown sound emitters. The measurements compiled in this study were then compared to four principal studies. Steiner (1981) showed differences in pure tonal vocalizations in five delphinid species in the Western North Atlantic Ocean and found 99% accuracy rates distinguishing between species. Rendell et al. (1999) investigated five odontocete cetaceans in several locations including the Caribbean, Costa Rica and Newfoundland and found 90% accuracy rates in distinguishing species by location. Oswald et al. (2003) observed nine delphinid species in the Tropical Pacific Ocean and found 41.1% accuracy rates in distinguishing between species. Baronet al. (2008) observed three delphinid species in the North Atlantic Ocean and the Gulf of Mexico with misclassification rates ranging from 5 to 17%. In addition to these principal studies, the results of this study were compared to others in the literature. There were similarities in some of the characteristics to at least one of the four comparison studies for all of the species.

Because this study was a preliminary analysis of acoustic signals of delphinids in the northwest Atlantic, this review further explores additional acoustic methods, management and conservation strategies relying on acoustic studies and possibilities for future studies.

This document is currently not available here.

For NSU Patrons Only.