| Summary: | This thesis examines the masking effect of underwater noise on beluga whale communication. As ocean water is greatly opaque for light but well conducting for sound, marine mammals rely primarily on their hearing for orientation and communication. Man-made underwater noise has the potential of interfering with sounds used by marine mammals. Masking to the point of incomprehensibility can have fatal results-for the individual, but ultimately for the entire species. As part of our understanding of whether marine mammals can cope with human impact on nature, this thesis is the first to study the interference of real ocean noises with complex animal vocalizations. At the Vancouver Aquarium, a beluga whale was trained for acoustic experiments, during which masked hearing thresholds were measured. Focus lay on noise created by icebreaking ships in the Arctic. As experiments with trained animals are time and cost expensive, various techniques were examined for their ability to model the whale's response. These were human hearing tests, visual spectrogram discrimination, matched filtering, spectrogram cross-correlation, critical band cross-correlation, adaptive filtering and various types of artificial neural networks. The most efficient method with respect to similarity to the whale's data and speed, was a backpropagation neural net. Masked hearing thresholds would be of little use if they could not be related to accessible quantities in the wild. An ocean sound propagation model was applied to determine critical distances between a noise source, a calling whale and a listening whale. Colour diagrams, called maskograms, were invented to illustrate zones of masking in the wild. Results are that bubbler system noise with a source level of 194 dB re 1 μPa at 1 m has a maximum radius of masking of 15 km in a 3-dimensional ocean. Propeller noise with a source level of 203 dB re 1 μPa at 1 m has a maximum radius of masking of 22 km. A naturally occurring icecracking event with a source level of 147 dB re 1 μPa at 1 m only masks if the listening whale is within 8 m of the event. Therefore, in the wild, propeller cavitation noise masks furthest, followed by bubbler system noise, then icecracking noise. Science, Faculty of Earth, Ocean and Atmospheric Sciences, Department of Graduate
|
|---|