Unraveling ocean soundscape dynamics with acoustic scene classification
Abstract
Due to its value in helping scientists understand the distribution and behavior of soniferous marine animals, the use of passive aco,ustics monitoring has grown exponentially over the past decade. The resulting petabytes of data on ocean soundscapes contain informa,tion on how sound source contributors vary over time and space, and how and what animals are using these ecosystems on hourly to sea,sonal to multi-year timescales. Efficiently extracting this critical information and comparing it to other datasets in the context o,f ecosystem-based research management is a big data challenge for the scientific community to overcome. There is also a growing need, for the implementation of standardized processing routines that facilitate the comparison of datasets collected across projects lik,ely with varying time and space. To address these needs and move towards a more holistic understanding of the broader marine sound e,cosystem, this research effort will analyze soundscapes from disparate marine acoustic datasets and integrate with environmental var,iables using innovative analytical approaches that quantify contributions of sound sources over time. This research goal will be ach,ieved through the following objectives: 1) apply freely accessible acoustic processing routines to extract sound level metrics follo,wing international standards from diverse monitoring efforts; 2) implement an acoustic scene classification framework where the soun,dscape is deconstructed and quantified based on contributions from anthropogenic, physical and biological sources; 3) integrate and,examine concurrent environmental data and products available from platforms like NOAA ERDDAP and the IOOS Data Catalog, and 4) visua,lize the results by leveraging existing web-based frameworks to display and explore acoustic scene metrics ambient sound levels buil,t for the NOAA-Navy Sanctuary Soundscape Monitoring Project. The proposed research will be a large step forward to efficient, large-,scale acoustic scene analysis that will leverage and enhance a suite of tools that blend data science approaches with acoustic data,for improved visualization and analysis.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 05, 2022
- Source ID
- N000142212601
Entities
People
- Carrie Bell
Organizations
- Office of Naval Research
- Regents of the University of Colorado
- United States Navy