Characterizing ambient ocean sound over decadal time scales in the Atlantic Ocean to identify mechan
Abstract
The capability of underwater low-frequency sound to travel for very long distances makes sound the oceans main information career.,Natural (biological and geological) and man-made (anthropogenic) sounds are the main sources of underwater acoustics content, and bo,th can range from few Hz to tens of kHz. The early reporting, from the 1960s to 1990s specifically in the Northeast Pacific Ocean, o,f the low-frequency ambient sound levels showed an increase of 3 dB/decade which was related to human activities. The historical cha,nge of the soundscape is of interest to the Navy and can facilitate the prediction of the long-term sound level patterns and trends,necessary for achieving optimal sonar performance. Specific information content related to unique acoustic signals inherently plays,a key role in identifying the mechanistic drivers of the soundscape. Amplitude, spectral, and source analyses related to North Atlan,tic Ocean soundscapes across decades proposed in this project will provide a comprehensive insight of the acoustic environment. Also,, the analyses will combine information across regions to facilitate the monitoring of the oceanic ambient noise changes at a global, scale given the ocean status and climate conditions changes. Until recently, most studies concerned with understanding changes in t,he soundscape have focused on the NE Pacific Ocean over the past seven decades or specific regions in the past decade. The informati,on acquired from such long-term analyses can help to interpret marine mammal behavior and their habitat choices. In addition, ocean,acoustics affords prompt awareness regarding the impact of human use of the ocean and the global economy. In the proposed study, the, knowledge obtained from the analysis of historical soundscapes in regions outside the NE Pacific Ocean will provide a better unders,tanding of the soundscape changes over the past 5 decades. The goals of the proposed study involve 1) evaluating the long-term patte,rns and trends of ocean sound over multiple decades in two sites in the Atlantic Ocean, and 2) identifying the sound source contribu,tions at each of the two sites to interpret their impact on the soundscape dynamics. To achieve the aforementioned goals, a set of c,lassical signal processing tools along with data-driven techniques such as machine learning techniques will be utilized to facilitat,e the extraction of acoustic features and cluster different sound sources including biology, geology, and anthropogenic activity. Th,e latter sound source plays an increasingly important role when it comes to underwater noise pollution that can be characterized dur,ing this study. The development and assessment of proper statistical models and signal processing algorithms will be determined afte,r the examination of the datasets. However, it is expected that the data will exhibit non-linearity and hence violate the homogeneit,y assumptions where classical linear models would fail. Therefore, data-driven techniques such as mode decomposition and machine lea,rning approaches might play a critical role in addressing the lack of spatial and temporal independence of the data, hence facilitat,ing a robust and flexible framework to address the complex temporal-spatial dynamics that the dataset might exhibit.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 10, 2021
- Source ID
- N000142212032
Entities
People
- Mahdi Al-badrawi
Organizations
- Office of Naval Research
- United States Navy
- University System of New Hampshire