Date of Award

Winter 2025

Project Type

Thesis

Program or Major

Computer Science

Degree Name

Master of Science

First Advisor

Thomas Butkiewicz

Second Advisor

Laura Dietz

Abstract

A variety of elements are used to characterize marine soundscapes, including natural and anthropogenic sound. One anthropogenic activity that potentially contributes to marine soundscapes, but is not sufficiently understood, is the use of different types of ocean-mapping sonar systems. Sonar systems are used abundantly in marine environments for purposes such as ocean floor mapping and military security purposes [Ryu et al., 2023] [Marszal and Salamon, 2013]. The systems used to conduct this research are the Kongsberg TOPAS PS40 sub-bottom profiler and the RESON SeaBat T51-R multibeam echosounder, along with SoundTrap recorders, which were deployed in the Gulf of Maine. Periods of ocean-mapping sonar operation analyzed were pre-exposure, sonar operation exposure, and post-exposure using soundscape metrics, including amplitude, impulsivity, uniformity, and periodicity. Analyses revealed some significant changes in some metrics, most commonly in higher frequency ranges that overlap with the dominant frequencies of the PS40. Machine learning applications, including a convolutional neural network and a transformer, were evaluated on their potential to classify the exposure period, using soundscape metrics, distance from the recorder to the source, and the time of day and the day of week as input. The greatest contributions of the TOPAS PS40 to the soundscape during operation were increased amplitude and periodicity, primarily above 1 kHz. As for the soundscape during operation of the T51-R, there was increased periodicity above 10 kHz and decreased impulsivity across all frequency bands compared to the soundscape before operation. Observed changes in soundscapes during operation of the T51-R can’t be directly linked to the T51-R itself, since its operating frequency is too high to capture on the recorders, whereas direct contributions from the TOPAS PS40 are more easily identifiable than contributions from the SeaBat T51-R, due to the PS40 operating at a frequency at which the SoundTraps could capture. Any sounds that occur during operation of the PS40 or the T51-R could be caused by the boat to which both systems were attached. Limitations of this study mainly include the SoundTrap recorders used being incapable of recording at a frequency high enough to capture the sounds created by the SeaBat T51-R, there was no time at which either ocean-mapping sonar was in operation independent of movement of the ship to which it was attached, and the ocean-mapping sonar systems and the survey ships only operated at around the same time of day on weekdays. This study took advantage of the data and resources available but was limited in scope.

Download

Share

COinS