Characteristics and Contributions of Noise Generated by Mechanical Cutting During Conductor Removal Operations
Contribution of Sounds Produced by Marine Mammals
This novel exploration of the sound generated by the removal of conductors via mechanical cutting is an effort funded by The Bureau of Ocean Energy Management (BOEM) and The Bureau of Safety and Environmental Enforcement (BSEE). Led by Tetra Tech, Inc., this study involved the deployment of several SoundTrap passive acoustic recorders used to monitor noise surrounding the platform during periods of conductor cutting.
Read the Report
Ocean Science Analytics contributed to the project in several ways. We were primarily responsible for processing the data from a single SoundTrap recorder to determine periods containing vocalizing marine mammals and identify species whenever possible. However, since processing data from a large acoustic dataset can be time consuming and requires experienced acoustician review, we evaluated two additional analytical methods for processing data. We first evaluated three soundscape metrics to determine their utility in detecting periods with either marine mammal vocalizations or periods of conductor cutting noise. Secondarily, we developed a pair of species-specific, scalable deep neural networks for use in future studies within the region. These efforts support the project by providing a first look to better understand how sound generated during conductor removal activities may interact with sounds from marine mammals, and contribute to the overall soundscape within the study area.
This open-source software program is an industry-leader in the passive acoustic monitoring (PAM) of marine mammals. In addition to real-time monitoring, PAMGuard provides a suite of powerful post-processing tools. We use a combination of in-built detectors parameterized to detected a wide range of species, then review the automated detections in PAMGuard's Viewer Mode to annotate acoustic events and assign to species/species group. This tool provides an efficient, semi-automated means of analyzing large acoustic datasets.
We used R programming language and the 'soundecology' and 'seewave' packages to obtain measurements of three acoustic-based biodiversity indices from the raw audio data. We evaluated periods with noise, marine mammal calls, and ambient noise levels to determine if measurements were influenced by these acoustic features.
