Undergraduate Research and Creativity

URECA

2008-2009

Analysis of vocalizations from a captive bottlenose dolphin (Tursiops truncatus)

Melissa Mazzocco and Joseph D. Warren
School of Marine and Atmospheric Sciences, Stony Brook University
&
Robert DiGiovanni
Riverhead Foundation for Marine Research & Preservation, Riverhead, NY

A mother and calf bottlenose dolphins (Tursiops truncatus) were rescued by the Riverhead Foundation in late September 2006 from Sag Harbor, NY. Unfortunately the mother did not survive transport to the Riverhead Foundation's rehabilitation facility. The calf, named Ariel, did survive and was placed in a large tank for rehabilitation. After arrival at Riverhead, Ariel was kept alone in the tank with periodic medical examinations and interaction sessions with Riverhead staff. Roughly one month later an underwater microphone (hydrophone) was placed within her enclosure and recordings were made periodically throughout the day while members of the Riverhead staff were present. The recordings varied in length from 3 to 600 seconds. Initial recordings had a sampling rate of 22.5 kHz but this was later changed to 48.0 kHz. In addition to the recordings, the Riverhead Foundation staff recorded the occurrence of stimulus and other interactions that occurred in the tank. We examined several hundred recordings to measure the characteristics of Ariel's vocalizations which were primarily made up of long series of whistles. Using Raven Lite 1.0 software (Cornell Lab of Ornithology), we measured several different vocalization parameters including: the beginning frequency, end frequency, peak frequency, maximum frequency, minimum frequency, and start time and duration of each complete whistle. We also noted other non-whistle vocalizations and non-dolphin sounds present within the recordings, the number of inflection points in each whistle, the presence of harmonics, and the time between whistles. We used this information to determine the whistle classification types, which was done using the identification numbers outlined in "Pitfalls in the categorization of behavior: a compilation of dolphin whistle classification methods." (V. Janik, Animal Behavior, 57, 133-143). A total number of 4,833 whistles have been identified in the data set, but characterization of them is still ongoing. By far the most common whistle type identified thus far has been type 90. The frequency of whistles varied greatly amongst our data set and we are investigating whether external stimuli may be related to an increase in whistle occurrence.