Date of Award

2024

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

Rycyk, Athena

Area of Concentration

Marine Biology

Abstract

Bottlenose dolphins (Tursiops truncatus) have been the subject of long-term research in Sarasota Bay since 1970, and are extremely vocal animals. A network of passive listening stations in the bay create potential for passive acoustic monitoring of the dolphin community, specifically with individually distinctive whistle types (signature whistles). Attempts to find whistles in large datasets have necessitated huge amounts of time to manually verify detections. An automated whistle detection method was built and tested with Kaleidoscope Pro, an acoustic analysis software. Recordings were used from two stations in Sarasota Bay: Palma Sola Bay (2018) and New College (2022). Three classifiers were tested with a subset of the New College station data, and evaluated for their hit, miss, and false alarm rates. The most accurate (an adapted classifier from the Palma Sola Bay station) was chosen for a complete analysis of the 2022 New College dataset. When the classifier was applied to the full dataset, the highest frequency of detected whistles were found in winter and spring, and in the afternoons.

Download

Share

COinS