Open AccessMeasuring the Marine Soundscape of the Indian Ocean with Southern Elephant Seals Used as Acoustic Gliders of Opportunity
Author(s) -
Dorian Cazau,
Julien Bonnel,
Joffrey Jouma’a,
Yves Le Bras,
Christophe Guinet
Publication year - 2017
Publication title -
journal of atmospheric and oceanic technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.774
H-Index - 124
eISSN - 1520-0426
pISSN - 0739-0572
DOI - 10.1175/jtech-d-16-0124.1
Subject(s) - soundscape , underwater , acoustics , ambient noise level , sound pressure , bioacoustics , sound (geography) , environmental science , noise (video) , wind speed , signal (programming language) , multivariate statistics , computer science , geology , oceanography , physics , artificial intelligence , image (mathematics) , machine learning , programming language
International audienceThe underwater ambient sound field contains quantifiable information about the physical and biologicalmarine environment. The development of operational systems for monitoring in an autonomous way theunderwater acoustic signal is necessary for many applications, such as meteorology and biodiversity protection.This paper develops a proof-of-concept study on performing marine soundscape analysis fromacoustic passive recordings of free-ranging biologged southern elephant seals (SES). A multivariate multiplelinear regression (MMLR) framework is used to predict the measured ambient noise, modeled as a multivariateacoustic response, from SES (depth, speed, and acceleration) and environmental (wind) variables.Results show that the acoustic contributions of SES variables affect mainly low-frequency sound pressurelevels (SPLs), while frequency bands above 3 kHz are less corrupted by SES displacement and allow a goodmeasure of the Indian Ocean soundscape. Also, preliminary results toward the development of a mobileembedded weather sensor are presented. In particular, wind speed estimation can be performed from thepassive acoustic recordings with an accuracy of 2ms21, using a rather simple multiple linear model