Marine Mammal Monitoring Research Articles

Arctic marine mammal passive monitoring and tracking with a single acoustic sensor

The Arctic Ocean is a unique environment in the number of physical mechanisms that may be potentially exploited with much simpler acoustic systems than would be required in other oceans. The Arctic sound speed profile forms a surface duct with favorable cylindrical-spreading for near-continuous detection of marine mammal vocalizations. This ducted waveguide exhibit low seasonal variability, particularly under the ice cap, forcing under-ice sound to heavily interact with this rough elastic stratified boundary. The ice roughness introduces steeper slopes that enhance water-to-ice sound penetration [Arvelo, POMA 2012]. The ice elasticity is responsible for the excitation of a radially polarized longitudinal wave and a transverse-horizontal shear wave with group velocities around 2700-3000 m/s and 1550-1650 m/s, respectively. A third dispersive flexural vertical plate wave propagates at much slower speeds (<1200 m/s) at low frequencies [Stein, Euerle & Parinella, JGR 1998]. Vocalization distance may be estimated from the time delays between the three wave types via blind deconvolution, while an arctangent bearing-estimator may increase the azimuthal localization resolution for high SNR vocalizations [Maranda, Oceans 2003]. Therefore, the unique Arctic environment is well suited for passive marine mammal monitoring and tracking with just a single ice-embedded geophone or under-ice vector sensor.

Investigating the Potential Use of Environmental DNA (eDNA) for Genetic Monitoring of Marine Mammals

The exploitation of non-invasive samples has been widely used in genetic monitoring of terrestrial species. In aquatic ecosystems, non-invasive samples such as feces, shed hair or skin, are less accessible. However, the use of environmental DNA (eDNA) has recently been shown to be an effective tool for genetic monitoring of species presence in freshwater ecosystems. Detecting species in the marine environment using eDNA potentially offers a greater challenge due to the greater dilution, amount of mixing and salinity compared with most freshwater ecosystems. To determine the potential use of eDNA for genetic monitoring we used specific primers that amplify short mitochondrial DNA sequences to detect the presence of a marine mammal, the harbor porpoise, Phocoena phocoena, in a controlled environment and in natural marine locations. The reliability of the genetic detections was investigated by comparing with detections of harbor porpoise echolocation clicks by static acoustic monitoring devices. While we were able to consistently genetically detect the target species under controlled conditions, the results from natural locations were less consistent and detection by eDNA was less successful than acoustic detections. However, at one site we detected long-finned pilot whale, Globicephala melas, a species rarely sighted in the Baltic. Therefore, with optimization aimed towards processing larger volumes of seawater this method has the potential to compliment current visual and acoustic methods of species detection of marine mammals.

Passive acoustic density estimation of sperm whales in the Tongue of the Ocean, Bahamas

AbstractLong‐term passive acoustic monitoring of marine mammals on navy ranges provides the opportunity to better understand the potential impact of sonar on populations. The navy range in Tongue of the Ocean (TOTO), Bahamas contains extensive hydrophone arrays, potentially allowing estimation of the density of deep diving, vocally active species such as the sperm whale (Physeter macrocephalus). Previous visual surveys in TOTO have been of limited spatio–temporal coverage and resulted in only sporadic sightings of sperm whales, whereas passive acoustic observations suggest the species is present year round. However, until now the means of acoustically determining the specific number of individuals in each cluster has been limited. We used recently developed algorithms to identify the number of echolocating whales present during a 42 d study period. We screened a 297 h acoustic data set to determine the proportion of time animals were present; fifty 10 min samples during presence were analyzed to estimate the number of individuals vocalizing during each sample. These counts were combined with an independent estimate of the proportion of 10 min periods when tagged animals vocalize. The estimated average density was 0.16 whales/1,000 km2(CV 27%; 95% CI 0.095–0.264). The method is potentially applicable to other areas containing dense hydrophone arrays.

Effectiveness of Monitoring Marine Mammals during Marine Seismic Surveys off Northeast Brazil

The possible behavioral and physical problems involving marine mammals and marine seismic surveys have attracted the interest and concern of ...

Temporal stability and species specificity in bacteria associated with the bottlenose dolphins respiratory system

We demonstrate that the exhaled breath condensate, or 'blow', from marine mammals can be used to examine respiratory associated microbial communities using non-invasive sampling methods. Blow samples from two species of bottlenose dolphin, Tursiops truncatus and T. aduncus, along with hybrid offspring, were examined using molecular microbial ecology methods. A temporal analysis revealed that microbial community structure of each individual remained distinct from other individuals over a two-month period, indicting strong host specificity. The taxonomic composition of samples, based on pyrosequencing of the V1-V3 regions of the 16S rRNA gene, from 24 healthy individuals was dominated by the Cardiobacteraceae lineage of Gammaproteobacteria, comprising on average 52% of sequences in all samples. Sequences in this taxa were associated with novel clades that contain only sequences from dolphin respiratory tracts. Other genera that likely form part of the core biota include the Saccharospirillaceae (Gammaproteobacteria), Arcobacter (Epsilonproteobacteria), Hydrogenimonaceae (Epsilonproteobacteria), Halotalea (Gammaproteobacteria), Aquimarina (Flavobacteria) and Helococcus (Clostridia). Significant differences between samples from different species were observed only at the species/ strain level, driven by the relative contributions of strains from the most common phylogenetic lineages. Analysis of communities associated with hybrid animals provides tentative evidence for a paternal role in community assembly. Clear overlap was observed with data collected by capture and swabbing of bottlenose dolphins blowholes, indicating this method provides a novel non-invasive alternative to monitoring marine mammal population health.

A framework for the automated real-time detection of short tonal sounds from ocean observatories

Acquisition of acoustic data from ocean observatories is expected to play a key role for the long-term monitoring of marine mammals and anthropogenic noise. It typically requires processing of a large volume of acoustic data and it must rely on automated identification of signals. We present an algorithmic framework for the detection of short tonal sounds (e.g. cetacean calls, anthropogenic pings) intended to act as a first stage in a system for the automated real-time detection, classification, and localisation of acoustic sources. The algorithm was validated under a diversity of scenarios expected at ocean observatories. Using simulated signals that emulate a variety of cetacean call-types, perfect identification of signal position was obtained for signal to noise ratios of −15 to −5 dB, depending on the signal-type. Separation of real-world data segments with short tonal sounds (mainly cetacean calls) from segments with other sounds or noise resulted in Area Under the ROC Curve values between 0.96 and 0.98. The algorithm can be used to automatically identify cetacean calls and anthropogenic short tonal sounds much faster than in real-time, thereby reducing the burden put on data transmission, storage, or processing by classification and localisation algorithms.

Nighttime foraging by deep diving echolocating odtonocetes in the Hawaiian Islands

Ecological acoustic recorders (EARs) were deployed in deep waters at five locations around the island of Kauai and one in waters off Ni'ihau in the main Hawaiian island chain. The EARs were moored to the bottom at depths between 400 and 800 m. The data acquisition sampling rate was 80 kHz and acoustic signals were recorded for 30 s every 5 min to conserve battery power and disk space. The acoustic data were analyzed using a suite of software including the M3R (marine mammal monitoring on navy ranges) algorithm, an energy-ratio-mapping algorithm developed at Oregon State University, Triton software developed at Scripps Institute of Oceanography, and custom Matlab programs. A variety of deep diving odontocetes, including pilot whales, Risso's dolphins, sperm whales, spinner and pan-tropical spotted dolphins, and beaked whales were detected at all sites. Foraging activity typically began to increase after dusk, peaked in the middle of the night, and began to decrease toward dawn. Between 75 and 87% of biosonar clicks were detected at night. At present, it is not clear why some of the known deep diving species, such as sperm whales and beaked whales, concentrate their foraging efforts at night.

Passive acoustic monitoring of North Atlantic right whales in Cape Cod Bay

Passive acoustic techniques have been applied extensively to marine mammal monitoring, localization, and tracking. An acoustic fieldwork using hydrophone arrays and free-floating buoys was conducted in Cape Cod Bay in the spring of 2011 for monitoring North Atlantic right whales. Three vertical hydrophone arrays were deployed to form a large triangular network with approximately 12 km on each side, and one horizontal array was mounted on the seafloor at one station. The passive hydrophone arrays operated for 25 days and recorded a vast amount of vocalizations made by humpback, fin, sei, minke, and, most importantly, right whales. These data can be used for localizing these endangered animals and tracking their movements in the bay with a large spatial coverage. Acoustic buoys, the real-time acoustic tracking system, were deployed on two one-day cruises for several hours at a time to triangulate calling whales over small spatial scales, and can be used to ground truth long-range whale localizations derived from the array network. Results from different passive acoustic techniques will be presented and compared. [Work supported by the Ocean Life Institute and the Marine Mammal Center of the Woods Hole Oceanographic Institution.]

A review and inventory of cabled and radio-linked hydrophones for passive acoustic monitoring of marine mammals.

Fixed cabled hydrophone (FCH) and radio-linked hydrophone (RLH) systems are permanently, or semi-permanently, installed acoustic monitoring systems that are located on or moored to the seafloor. These systems have the capability to passively monitor bio-acoustic signals from marine mammals and, therefore, have great potential for monitoring and mitigation of potential impacts caused by anthropogenic activities. As part of a Joint Industries Programme sponsored effort, we reviewed past and present FCH and RLH systems with respect to their capabilities, advantages/disadvantages, and effectiveness for monitoring marine mammals in relation to oil and gas exploration/production activities. Based on this review, we provide examples and applications of these technologies. FCHs are typically powered by an external source and send data continuously to a receiving station that is usually located on shore. RLHs are moored or fixed to the seafloor, and transmit acoustic signals via radio-waves to a receiving station on shore. Both these systems allow acoustic data to be remotely monitored and processed in (or near) real-time. Hybrid systems can offer a good compromise between cost and capability by providing near real-time data transmission/processing with greater flexibility in deployment possibilities, but are usually limited in longevity and bandwidth of monitoring.

A distributed micro sonar network for marine mammal monitoring.

The design of a distributed micro sonar network (DMSN) is presented. The DMSN is a miniaturized, low‐power, and high‐frequency version of the swimmer detection sonar network used to detect and track marine mammals at ranges out to 500 m. The intent is to use it for monitoring during potentially harmful military and commercial activities such as high‐powered Navy sonar, oil and gas exploration, explosive removal of offshore structures, seismic exploration for research and oil and gas, pile driving to install maritime structures including offshore wind farms, and marine hydrokinetic energy devices such as free turbines. The system will consist of multiple single‐beam sonars, each transmitting and receiving a unique signal in a narrow beam. The hardware and software will initially be ported to a Xilinx FPGA, with future versions done within a custom made ASIC. Initial hardware and software research and development will be described, as well as the work done to create computationally efficient signal processi...

Automatic signal detection in noise using entropy.

Automatic detection of signals in noise is a common problem in many areas of acoustics. In the field of passive acoustic monitoring of marine mammals, the signals to be detected are vocalizations. The noise originates from natural (wind, waves, and rain) and man‐made sources (e.g., shipping, construction, and seismic surveys). Signal characteristics vary broadly: frequency ranges from a few Hz to 200 kHz, duration from milliseconds to seconds to hours. Noise characteristics vary by similar orders of magnitude. While specific automatic detectors have been designed to successfully find specific calls in specific environments, the challenge is to find a large variety of calls in a large variety of noise. An exploitable difference between calls and noise is that most noise is a result of stochastic processes (wind, waves, rain, cavitating propellers + seismics generate gas bubbles underwater of varying size + resonance frequency), while many animal signals are a result of deterministic processes (vibrating st...

Autonomous underwater glider based embedded real-time marine mammal detection and classification.

Autonomous marine vehicles offer the potential to provide low-cost data suitable for passive acoustic monitoring applications of marine mammals. Due to their extremely low-power consumption and long range, gliders are an attractive option for long-term deployments. Challenges related to power availability, payload size, and weight have previously restricted the viability of marine mammal monitoring. As an example, the wide bandwidth of odontocete echolocation clicks requires a high sampling rate and poses challenges with respect to limitations in power, size, and weight of the deployed system. Recent developments in commercial off-the-shelf hardware driven by the mobile phone industry’s need for multimedia-rich smart phones have resulted in low-power architectures capable of performing computationally demanding signal processing and stochastic recognition tasks in real time. We describe our work on a small form-factor, light-weight package used to perform real-time passive acoustic detection and classification. The system detects echolocation clicks using Teager energy. Echolocation clicks are then classified using cepstral features processed by a Gaussian mixture model. [Work sponsored by ONR.]

A source separation approach to enhancing marine mammal vocalizations

A common problem in passive acoustic based marine mammal monitoring is the contamination of vocalizations by a noise source, such as a surface vessel. The conventional approach in improving the vocalization signal to noise ratio (SNR) is to suppress the unwanted noise sources by beamforming the measurements made using an array. In this paper, an alternative approach to multi-channel underwater signal enhancement is proposed. Specifically, a blind source separation algorithm that extracts the vocalization signal from two-channel noisy measurements is derived and implemented. The proposed algorithm uses a robust decorrelation criterion to separate the vocalization from background noise, and hence is suitable for low SNR measurements. To overcome the convergence limitations resulting from temporally correlated recordings, the supervised affine projection filter update rule is adapted to the unsupervised source separation framework. The proposed method is evaluated using real West Indian manatee (Trichechus manatus latirostris) vocalizations and watercraft emitted noise measurements made within a typical manatee habitat in Florida. The results suggest that the proposed algorithm can improve the detection range of a passive acoustic detector five times on average (for input SNR between -10 and 5 dB) using only two receivers.

On the consideration of motion effects in the computation of impulse response for underwater acoustics inversion

The estimation of the impulse response (IR) of a propagation channel may be of great interest for a large number of underwater applications: underwater communications, sonar detection and localization, marine mammal monitoring, etc. It quantifies the distortions of the transmitted signal in the underwater channel and enables geoacoustic inversion. The propagating signal is usually subject to additional and undesirable distortions due to the motion of the transmitter-channel-receiver configuration. This paper shows the effects of the motion while estimating the IR by matched filtering between the transmitted and the received signals. A methodology to compare IR estimation with and without motion is presented. Based on this comparison, a method for motion effect compensation is proposed in order to reduce motion-induced distortions. The proposed methodology is applied to real data sets collected in 2007 by the Service Hydrographique et Océanographique de la Marine in a shallow water environment, proving its interest for motion effect analysis. Motion compensated estimation of IRs is computed from sources transmitting broadband linear frequency modulations moving at up to 12 knots in the shallow water environment of the Malta plateau, South of Sicilia.

Blind deconvolution of remote-source signals from acoustic array recordings in multipath environments.

Reconstructing the signal from a remote source in an unknown multipath environment is a task commonly known as blind deconvolution. The existence of generic features of underwater sound fields, propagating modes or ray paths, may be exploited for blind deconvolution via artificial time reversal (ATR), [Sabra and Dowling, J. Acoust. Soc. Am. 116, 262–271 (2004)] from vertical array recordings. ATR uses a weighted sum of array recordings to determine a frequency-dependent phase correction that aligns the various mode or ray-path arrivals at the receiving array to form an estimate of the original source waveform. A description of ATR will be presented along with parameter studies of ATR performance for simulated signals in a 100-m-deep Perkeris waveguide, and for airborne sounds measured in a reverberant laboratory. Here, the correlation coefficient between original and the ATR-reconstructed signals will be presented as a function of the number and spacing of array elements, and signal characteristics (frequency and bandwidth). The limitations imposed by noise will be quantified for the measured signals. Possible applications of ATR will be briefly discussed; these include reducing inter-symbol interference in underwater communication, tracking and identifying remote sources, and monitoring marine mammals that vocalize. [Work supported by ONR Code No. 321OA.]

Warping based lag-Doppler filtering applied to motion effect compensation in acoustical multipath propagation.

The estimation of an impulse response of a propagation channel is necessary for a large number of underwater applications: underwater communication, sonar detection and localization, marine mammal monitoring, etc. Basically it informs us about the distortions of transmitted signal in underwater channels and enables geoacoustic inversion. The propagated signal is usually subject to additional and undesirable distortions due to the motion of the transmitter-channel-receiver configuration. This paper points out the effects of the motion while estimating the IR by matching filtering between the transmitted and the received signals. A methodology to compare the IR estimation in motionless and motion contexts is presented. Based on this comparison, a warping-based lag-Doppler filtering method in the wideband ambiguity plane for motion effect compensation is proposed. It aims at reducing the distortions due to the motion phenomena and accurately estimating the IR of the channel. The proposed methodology is applied to real data sets issued from BASE’07 campaign (NURC, SHOM, and FWG, South of Sicilia, 2007). Motion compensated estimations of IRs are computed from sources transmitting broadband linear frequency modulations moving up to 12 knots in the shallow water environment of the Malta plateau.

Passive acoustical counting of odontocetes using towed and stationed platforms.

Counting the number of animals is essential for monitoring and management of marine mammals. Autonomous stereo acoustic event recorder (A-tag) to identify each sound source enabled to count phonating odontocetes in the wild. Finless porpoises were surveyed by towed A-tags in 1100 km stretch of Yangtze River from Wuhan to Shanghai. The calculated detection probability using acoustic method was approximately twice that for visual detection. The difference in detection probabilities between the two methods was caused by the large number of single individuals that were missed by visual observers. However, acoustics tended to underestimate group size due to the limited resolution of sound source bearing angles. The detection performance regarding stationary acoustic monitoring of this species was also examined. Three stereo acoustic data loggers were placed at different locations near the confluence of Poyang Lake and the Yangtze River. Acoustic monitoring confirmed the low density bottle neck of the population between the lake and the river that was consistent with visual observations. The passive pulse event monitoring was proved to be a practical and useful alternative to visual observations, especially for long-term and range-wide monitoring.

Integration of automated detection methods into NOAA Southwest Fisheries Science Center (SWFSC) acoustic marine mammal monitoring protocol.

Southwest Fisheries Science Center (SWFSC) has used combined visual and acoustic techniques to monitor marine mammal populations for the past 8 yrs. Currently, SWFSC passive acoustic surveys of cetaceans require specially trained personnel to monitor hydrophone signals in real-time. While effective, this method is time-consuming and costly. Automated detection of cetacean vocalizations would be a valuable tool during SWFSC surveys, allowing for detection when experienced technicians are unavailable. This technique is advantageous because it significantly reduces effort and removes sources of human error and bias in detection ability. PAMGUARD 1.0 CORE software was evaluated for use in automated detection of cetacean acoustic signals. Three different detector configurations of PAMGUARD were evaluated. This work shows that the majority of whistle and click events can be detected using PAMGUARD software. All of the PAMGUARD trials were capable of detecting whistles and clicks of cetacean species with varying success. These techniques were field-tested at sea during a recent SWFSC marine mammal survey. Automated detection of beaked whales and Dall’s porpoise during this survey will also be discussed. It is our goal to integrate automated detection methods into SWFSC’s acoustic marine mammal monitoring protocol and this work is an important step in doing so.

From mysticete song to odontocete echolocation: Monitoring cetacean sounds with high-frequency acoustic recording packages (HARPs).

Marine mammals produce a wide range of sounds from 10s of Hz to 100s of kHz. To remotely monitor these various sounds with autonomous systems, instruments with wide bandwidths are required. A general purpose, autonomous, high-frequency acoustic recording package (HARP) is described that can record sounds ranging from low-frequency blue whale (Balaenoptera musculus) song to mid-frequency dolphin whistles and beaked whale sweeps to high-frequency dolphin clicks. In addition to monitoring marine mammals, sounds from fish, contributions from wind and rain, and anthropogenic sources such as ships and sonar also are recorded potentially allowing for the study of cetacean behavioral response to these sounds. HARPs have been deployed for periods of up to one year, in various configurations such as standard moorings, seafloor packages, and arrays, and in remote locations including the Arctic, Bering Sea, Gulf of California, around Hawaii, and offshore of Washington State and southern California. The various configurations of these deployments and the wide range of marine mammal monitoring data they have provided are discussed in addition to lessons learned from these studies. [Work supported by the U.S. Navy CNO-N45.]

On the consideration of motion effects in underwater geoacoustic inversion

The estimation of an impulse response (IR) of a propagation channel is necessary for a large number of underwater acoustic applications: underwater communication, sonar detection and localization, marine mammal monitoring, etc. Basically, it informs us about the distortions of a transmitted signal in one underwater channel. This operation is usually subject to additional distortions due to the motion of the transmitter‐channel‐receiver configuration. This paper points on the effects of the motion while estimating the IR for shallow water environments in the very low frequencies bandwidth with matching filtering between the transmitted and the received signals. We propose a methodology to compare between the IR estimation in motionless and motion contexts, respectively. Using this methodology an objective criterion for motion effect analysis is proposed in order to measure the distortions due to the motion phenomena. The proposed methodology is applied to real data sets issued from PASSTIME campaign (SHOM, Bay of Biscaie, 2005) proving also its interest for motion effect analysis.