Xavier Mouy

Tracking of Pacific walruses in the Chukchi Sea using a single hydrophone

The vocal repertoire of Pacific walruses includes underwater sound pulses referred to as knocks and bell-like calls. An extended acoustic monitoring program was performed in summer 2007 over a large region of the eastern Chukchi Sea using autonomous seabed-mounted acoustic recorders. Walrus knocks were identified in many of the recordings and most of these sounds included multiple bottom and surface reflected signals. This paper investigates the use of a localization technique based on relative multipath arrival times (RMATs) for potential behavior studies. First, knocks are detected using a semi-automated kurtosis-based algorithm. Then RMATs are matched to values predicted by a ray-tracing model. Walrus tracks with vertical and horizontal movements were obtained. The tracks included repeated dives between 4.0 m and 15.5 m depth and a deep dive to the sea bottom (53 m). Depths at which bell-like sounds are produced, average knock production rate and source levels estimates of the knocks were determined. Bell sounds were produced at all depths throughout the dives. Average knock production rates varied from 59 to 75 knocks/min. Average source level of the knocks was estimated to 177.6 ± 7.5 dB re 1 μPa peak @ 1 m.

Autonomous Multichannel Acoustic Recorders on the VENUS Ocean Observatory

The Autonomous Multichannel Acoustic Recorder (AMAR, JASCO Applied Sciences) is a sophisticated precision instrument for passive acoustic monitoring and accurate underwater sound level measurements. It can be integrated with small hydrophone arrays and non-acoustic oceanographic sensors. To date, AMARs have typically been used autonomously and deployed for a few months to a year on oceanographic moorings; however, AMARs are also capable of real-time data streaming when connected to a data telemetry system. This paper describes the capabilities and functionality of the AMAR through the example of its integration within Ocean Networks Canadas VENUS Ocean Observatory deployed off the coast of British Columbia, Canada. The recent deployment of two AMAR-based hydrophone arrays and associated non-acoustic and oceanographic sensors within the VENUS system is presented in detail. The planned research and development within the AMARs on VENUS program, as well as preliminary results on the real-time automatic detection, classification, localization, and tracking of marine mammals, are presented. The two AMARs deployed on the VENUS Ocean Observatory demonstrate that, unlike traditional underwater acoustic recorders, the AMAR can act as a hub for mini ocean observatories, capturing and transmitting both acoustic and non-acoustic sensor data in realtime. It is demonstrated that the AMAR is an effective technology that can be used in near-shore, small-scale, low-cost ocean observatories.

Pacific walrus vocal repertoire in the northeastern Chukchi Sea: Call type description and relative proportion in time and space

Pacific walrus are present in the northeastern Chukchi Sea (NCS) from June to October. The study of their sounds has been largely restricted to the knock and bell sounds produced by males during the breeding season and in-air grunts and barks from mother and pups. A passive acoustic monitoring program conducted in the NCS since July 2006 has brought strong evidence that the underwater vocal repertoire of walrus is more diverse. Nine call types (including knocks and bells) and their variants identified over four years of acoustic monitoring will be described. Spectral measurements along with estimates of variability for high signal-to-noise ratio calls will be provided. The relative proportion of each call type across the study area and throughout the season is currently analyzed based on the identification of all calls in samples recorded multiple times per day in 2009 and 2010. Preliminary results suggest that the vocal repertoire of walrus is dominated by grunt-type calls, which is consistent with the N...

Real-time acoustic monitoring and source level estimates of walruses in the northeastern Chukchi Sea using particle velocity sensors

Particle motion sensors measure the vector component of the sound field. In underwater acoustics, they are used for studying the physics of the sound field, evaluating the potential effects of sound on fish, and defining the direction of arrival (DOA) of sound sources. Measuring the DOA in the vertical and horizontal plane allows two separate receivers to localize an acoustic source in three dimensions. In July 2013, we used two custom-built, real-time particle velocity acoustic recording systems to record and localize vocally active walruses in the water near groups hauled out on ice in the northeastern Chukchi Sea. The system was equipped with a three-axis dipole sensors and a calibrated omni-directional hydrophone. It was deployed at the water surface and transmitted data in real-time to a support skiff. The range between the recorders, support skiff, and calling animals was usually less than 200 m and typically within a few tens of meters, allowing for simultaneous visual observations. Calling walruses were localized using cross-fixes of acoustic bearings. Source levels were estimated by adding modeled frequency-dependent transmission losses to the received levels in each 1/3-octave-band obtained from the calibrated omni-directional hydrophone. Only calls with high signal-to-noise ratio were used in this analysis. The use of the particle velocity sensor allowed for the first source level measurements of walrus grunts and bell calls in the wild.

Seasonal patterns in marine mammal vocalizations in the western Canadian Arctic

The Arctic marine environment is changing rapidly through a combination of sea ice loss and increased anthropogenic activity. Given that these changes can affect marine animals in a variety of ways, understanding the spatial and temporal distributions of Arctic marine animals is imperative. Here, we use passive acoustic monitoring to examine the presence of marine mammals in the western Canadian Arctic, where we have had recorders deployed near the communities of Sachs Harbour and Ulukhaktok, Northwest Territories, Canada. At both sites, we documented bowhead and beluga whales during the ice-free season, bearded seals throughout the ice-covered season and during their mating season, and ringed seals throughout the year. The sites also had different patterns in marine mammal presence, where we found whales later into the year at Ulukhaktok than at Sachs Harbour, and ringed seals vocalized much more at Ulukhaktok than at Sachs Harbour. These patterns in vocal activity at both sites help to document the presence of each species in the western Canadian Arctic, and serve as a baseline for future monitoring in the region. Next steps in this project involve deploying at more sites throughout the region to more comprehensively monitor marine mammals in the region.

Gray Whale (Eschrictius robustus) Call Types Recorded During Migration off the West Coast of Vancouver Island

Gray whale (Eschrichtius robustus) vocalizing behaviour varies over its range, both in call type and frequency. This study adds to the growing body of passive acoustic research outside of the calving/breeding lagoons, and moves toward a description of calling behaviours throughout the whales’ range. Data are presented here from acoustic surveys for two northward and one southward migration. We found gray whales to be highly sonorous, with extensive calling during the study periods of February to May for northward migrations and September to the end of January for southward travel. Low frequency moans were the most prevalent call type, with others, including knocks, up- and down sweeps, and rumbles, recorded in varying numbers. We hypothesize that calling is an aid for navigation and orientation of the herd along the migration route, in addition to holding more social functions usually assigned to baleen whale calling.

Passive acoustic monitoring of haddock in the Gulf of Maine

We have conducted several studies of haddock sounds in the Gulf of Maine (GOM) with mixed results. An analysis of an archival recording from captive haddock brood stock made in 1970 found that the “spawning rumble” sound occurred variously at the end of short thump trains, in the middle of thump trains, or in isolation. Interestingly, haddock were silent while spawning when we attempted to record sounds in the same facility in March 2000, suggesting that sound production may be negatively affected by chronic noise. Haddock sounds were absent in ROV and tethered instrument surveys in the summer and fall of 2001-2002. During 2006-2007, we deployed bottom mounted recorders while conducting long-line surveys of haddock spawning condition. Surprisingly few haddock sounds were detected and there was no correlation with spawning activity despite recording in highly active spawning areas. Haddock sounds consisted of isolated knocks, which were difficult to distinguish from thumps of unknown origin. We are now app...

Passive acoustic localization of fish using a compact hydrophone array

Passive acoustic monitoring of fish in their natural environment is a research field of growing interest and importance. Although many fish species are soniferous, the characterization and biological understanding of their sounds are largely unknown. Many underwater acoustic recordings contain sounds likely produced by fish, but little information can be extracted from them due to the lack of fundamental knowledge about the behaviors they represent. Deploying small hydrophone arrays can help fill some of these knowledge gaps. Passive acoustic localization using fish calls received on multiple hydrophones can be used to estimate swimming speed, calling rate of individual fish, and source level of their calls. This paper focuses on the three-dimensional localization of fish using a compact array of 6 hydrophones using both simulated and measured data. Fish sounds were detected manually on one of the hydrophones. Time difference of arrivals (TDOAs) were then defined by cross correlating the detected signal w...

Automated classification of Pacific white-sided dolphin (Lagenorhynchus obliquidens) pulsed calls for diel pattern assessment

Diel patterns in marine mammal activity can be difficult to assess visually due to the cost of ship time and limitations of daylight and weather during expeditions that can bias analysis based on sightings. Acoustic data can be a cost-effective approach to evaluate activity patterns in marine mammals. However, manual analysis of acoustic data is time consuming, and impractical for large data sets. This study seeks to evaluate diel patterns in Pacific white-sided dolphin communication through automated analysis of one year of continuous acoustic data collected from the Barkley Canyon node of Ocean Networks Canada’s NEPTUNE observatory, offshore of Vancouver Island, British Columbia, Canada. In this study, marine mammal acoustic signals are manually annotated in a sub-set of the data, and used to train a random forest classifier targeting Pacific white-sided dolphin pulsed calls. Binary and multiclass classifiers are compared, and the effects of different data-balancing methods are evaluated. The results fr...

Effects of anthropogenic noise on fishes at the SGaan Kinghlas-Bowie Seamount Marine Protected Area

Underwater noise from anthropogenic sources has been increasing dramatically for the past few decades and little is known about its effects on fishes. The objective of this study is to describe the occurrence and characteristics of fish sounds in the SGaan Kinghlas-Bowie Seamount Marine Protected Area (SK-B MPA, British Columbia, Canada) and to correlate them with the corresponding anthropogenic soundscape. Here we present preliminary results of the detection of fish sounds at SK-B MPA between July 2011 and July 2013. An automatic detector was used on nearly 40,000 acoustic samples (4,754.5 hours in total) to search for fish sounds. About 1.2% of the data were highlighted as containing fish-like signals. Manual verification of these detections revealed that 95.5% were false positives and the remaining sounds were of unknown origin. Eighty detections were highly stereotyped and are suspected to be produced by fish, but no identification has been confirmed yet. Systematic manual inspection of sub-sampled ac...