Directional Autonomous Seafloor Acoustic Recorders Research Articles

Multi-year measurements of the underwater noise field directionality in the shallow Beaufort Sea during open-water and drifting ice flow conditions

Between 2007 and 2014, over 35 Directional Autonomous Seafloor Acoustic Recorders (DASARs) were deployed over a 280 km swath of the Beaufort Sea continental shelf (20–55 m depth) during the open-water season, in order to monitor the fall (westward) bowhead whale migration. DASARs have one omnidirectional pressure sensor and two orthogonal particle motion sensors, which permit instantaneous measurements of the azimuths of both transient signals and continuous noise between 20 and 500 Hz. The lack of significant shipping or industrial noise in this region provided a rare opportunity to directly measure the properties of wind-driven noise in expanding ice-free regions. Here, we map the azimuthal directionality of the diffuse Beaufort ambient noise field as a function of frequency and space across seven seasons. The dominant directionality of the diffuse ambient noise field varied strongly with frequency and was highly correlated with the received power spectral density. Certain directional features of the ambient noise field remained stable over seven deployment seasons, suggesting that judicious processing of the ambient noise soundscape could provide underwater navigational information in arctic waters. The influence of local drifting floes and sheets of ice on this directionality is also examined. [Work sponsored by ONR.]

An evaluation of density estimation methods using a multiyear dataset of localized bowhead whale calls

For eight years (2007–2014), Greeneridge Sciences deployed, collected, and analyzed an immense passive acoustic dataset to study the response of bowhead whales (Balaena mysticetus) to oil exploration activities along a 280 km swath of the Alaskan Arctic continental shelf. Each year, up to 40 Directional Autonomous Seafloor Acoustic Recorders (DASARs) detected bowhead calls during the whales’ annual fall migration, resulting in over 2.4 million localized calls over the study period. These enormous sample sizes, combined with the unique localization capability of the DASARs, have enabled the investigation and cross-validation of several acoustic density estimation (DE) methods. Here, we compared results among three density estimation methods: (i) direct census, (ii) distance sampling, and (iii) spatially explicit capture recapture (SECR). In the course of our investigation, we encountered and addressed two challenges: (i) how to apply point transect theory to distributed array systems, and (ii) how to implement SECR in situations where detections between sensors are not statistically independent or where false positives exist on single sensors. These density estimates were then used to estimate relative abundance of the Western Arctic bowhead whale population over multiple years. [Work sponsored by the Joint Industry Programme.]

Long-term measurements of the ice-free underwater noise field directionality in the shallow Beaufort Sea

Between 2007 and 2014, over 35 Directional Autonomous Seafloor Acoustic Recorders (DASARs) were deployed over a 280 km swath of the Beaufort Sea continental shelf (20-55 m depth) during the open-water season, in order to monitor the westward bowhead whale migration. DASARs have one omnidirectional pressure sensor and two orthogonal particle velocity sensors that permit instantaneous measurement of the azimuths of both transient signals and continuous noise between 20 and 500 Hz, including diffuse wind-driven ocean noise. The lack of significant shipping or industrial noise in this region provides a rare opportunity to directly measure the properties of wind-driven noise in expanding ice-free regions. Here, we map the azimuthal directionality of the diffuse Beaufort ambient noise field as a function of frequency and location across all seven seasons. The dominant directionality of the diffuse ambient noise field varies strongly with frequency and is highly correlated with the received power spectral density. Certain directional features of the ambient noise field remained stable over seven deployment seasons, suggesting that judicious processing of the ambient noise soundscape could provide underwater navigational information in arctic waters. [Work sponsored by ONR.]

Source level and calling depth distributions of migrating bowhead whale calls in the shallow Beaufort Sea.

Automated and manual acoustic localizations of migrating bowhead whales were used to estimate source level and calling depth distributions of their frequency-modulated-modulated calls over seven years between 2008 and 2014. Whale positions were initially triangulated using directional autonomous seafloor acoustic recorders, deployed between 25 and 55 m water depth near Kaktovik, Alaska, during the fall westward migration. Calling depths were estimated by minimizing the "discrepancy" between source level estimates from at least three recorders detecting the same call. Applying a detailed waveguide propagation model to the data yielded broadband source levels of 161 ± 9 dB re 1 μPa2 s at 1 m (SEL) for calls received between 20 and 170 Hz. Applying a simpler 15 log10(R) power-law propagation model yielded SEL source levels of 158 ± 10 dB. The most probable calling depths lay between 22 and 30 m: optimal depths for long-range acoustic signal transmission in this particular environment.

Long-term measurements of the directionality and active intensity of the underwater noise field in the shallow Beaufort Sea

In each of the past seven years, at least 35 Directional Autonomous Seafloor Acoustic Recorders (DASARs) have been deployed over a 280 km swath of the Beaufort Sea continental shelf (20–55 m depth) during the open-water season to monitor the westward bowhead whale migration. DASARs have one omnidirectional pressure sensor and two orthogonal particle velocity sensors that permit measurements of the azimuths of both transient and continuous sounds, including diffuse ocean noise. Here, we map the azimuthal directionality of the Beaufort ambient noise field as a function of frequency and location across all seven seasons. Dominant directionalities exist in the diffuse ambient noise field, which change with frequency, time, and location. We examine how localized storms, heavy whale calling activity, seismic exploration, and other industrial activities influence the noise directionality. We also examine how both the active and reactive intensity of the noise evolves with frequency and time, by comparing the phase relationships between pressure and particle velocity. The directionality and active intensity help identify source mechanisms and help determine whether long-term changes in the ambient noise environment are occurring. [Work sponsored by the Shell Exploration and Production Company.]

Long-range tracking of bowhead whale calls using directional autonomous seafloor acoustic recorders

Since 2007 “Directional Autonomous Seafloor Acoustic Recorders” (DASARs) have been deployed at five sites across a 280 km swath of the Beaufort Sea continental shelf to record bowhead whale (Balaena mysticetus) calls during their autumn migration. Composed of an omnidirectional pressure sensor and two horizontal directional sensors measuring particle motion, DASARs provide information for determining the bearing to a sound source. In previous analyses, bearings obtained from multiple DASARs within a single site have been used to localize calls, with a maximum baseline separation of 21 km between instruments. Here, we use data collected from two different sites to exploit a 45 km instrument separation for tracking bowhead whale calls detected under low ambient-noise situations in 2009 and 2011. These data sets have been manually analyzed to extract a series of calls from individual whales swimming from one site toward another. The fact that the tracking method does not require relative arrival time information makes matching calls between these widely separated shallow-water recordings practical. Both empirical and numerical transmission loss models are used to investigate the relationship between animal orientation, source level, and the presence of harmonics.

Bowhead whales and airgun pulses: Detecting a threshold of behavioral reaction

Previous work has shown that bowhead whales cease calling when near (<40 km) seismic exploration activities involving airguns. The aim of this study is to estimate the received level threshold for the onset of this behavior (cessation of calling). The analysis relied on data collected during late summer of 2007-2010 by up to 40 Directional Autonomous Seafloor Acoustic Recorders (DASARs) in the Beaufort Sea. About 98,000 localized calls and hundreds of thousands of airgun pulses were included in the analysis. For each 10-min period of data collected at each recorder, each year, the cumulative sound exposure level (CSEL) from airgun pulses was calculated and paired with the number of calls concurrently localized within ~3.5 km of each DASAR. Poisson regression was then used to estimate the threshold of airgun sound exposure received at the whales when call cessation begins. The CSEL threshold was found to be near 124 dB re 1 μPa2 s (95% confidence intervals = 119-129 dB). For an airgun array firing every 10 sec, this corresponds to a received single pulse SEL at the whale of ~106 dB re 1 μPa2 s. [Work supported by Shell Exploration and Production Company.]

Contributions of airgun activity to the ice-free ambient noise environment of the shallow Beaufort Sea between 2007 and 2011

Every year since 2007 a collection of at least 35 “Directional Autonomous Seafloor Acoustic Recorders” (DASARs) have been deployed across a 280 km swath of the Beaufort Sea continental shelf, in water depths between 15 and 50 m. The ability of these instruments to estimate the arrival azimuth of transient signals has facilitated the development of an automated algorithm for the detection of airgun activity. This algorithm has been applied to five seasons of data, and in this presentation the contributions of this activity to the overall ambient noise background of the ice-free shallow-water Beaufort environment will be quantified with a variety of metrics, in terms of both level (peak-to-peak, rms, sound exposure level), frequency, and time (intervals and fraction of time present). During some years, up to four airgun operations could be detected simultaneously, and a random one-second time sample yielded a 40% chance of containing an airgun signal, but the levels detected are generally within the bounds of natural wind-driven ambient noise levels. This dataset provides useful empirical insight into discussions about the cumulative effects of anthropogenic activity on an environment extensively used by several marine mammal species. [Work sponsored by the Shell Exploration and Production Company.]

Bearing estimation and 2D localization of bowhead whale calls using directional autonomous seafloor acoustic recorders

Passive acoustic monitoring of the summer/fall westerly bowhead whale migration in the Beaufort Sea has been conducted by Greeneridge Sciences, sponsored by SEPCO, every year since 2006. The directional autonomous seafloor acoustic recorder (DASAR) packages used in this effort each contain three acoustic sensors that simultaneously measure the two horizontal components of acoustic particle motion and acoustic pressure. A variety of data-adaptive beamforming methods have been applied to a selected subset of these data to examine direction-of-arrival estimation performance, including quantifying bearing bias and variance. A principal component analysis (PCA)-type eigenanalysis of the sensor data cross spectral matrix is used to decompose the received field into orthogonal components having different particle motion polarization and energy flux properties. Appropriate manipulation of these components provides high resolution directional estimates of the primary arriving energy while maintaining robustness to uncertainties in sensor calibration and acoustic propagation conditions. Representative results of beamforming as well as 2D localization performance will be presented. Application of data-adaptive beamforming techniques to bulk processing of these large monitoring data sets will be discussed. [Work supported by Shell Exploration and Production Company (SEPCO).]

Matched-field processing and modal filtered range estimates of bowhead whale calls detected in the Alaskan Beaufort Sea

In 2010, a 15-element autonomous vertical array was deployed roughly 35 km north of Kaktovik alongside a distributed array of directional autonomous seafloor acoustic recorders (DASARs). Matched-field processing and geoacoustic inversion techniques were used to extract the range, depth, bottom sound speed profile, density, and attenuation from one close-range whale call. The inversion localized the call at 1.2 km range and 44 m depth in 55 m deep water. The inverted propagation model was used to derive the group and phase velocities of the normal modes in the region surrounding the array. The vertical array spanned sufficient aperture in the water column to permit isolation of the first and second mode arrivals from any given call. A range- and frequency-dependent phase shift was applied to each modal arrival to remove geometric dispersion effects. The modeled range that time-aligned the modal arrivals was selected as the range estimate. The modal filtering technique is demonstrated on additional whale calls produced at 17.3 km and 35 km range from the vertical array, with the range estimates independently confirmed by triangulating bearings of call detections on surrounding DASARs. [Work supported by the North Pacific Research Board, Shell Exploration and Production Company, and Greeneridge Sciences Incorporated.]

Analyzing potential acoustic differences between different bowhead demographics in the Beaufort Sea.

Most bowhead whales from the Bering-Chukchi-Beaufort stock migrate westward from Arctic Canadian waters through the Alaskan Beaufort Sea each year from late August to late October. Since 2007 both aerial sighting and acoustic call data have been collected during these fall migrations, as part of a long-term monitoring effort to assess the potential impacts of North Slope oil industry activities on bowhead whales. The aerial sighting efforts recorded the position of bowhead individuals or groups, designating the animals as subadults or adults, and noted whether a calf was present. A five-site array of seven directional autonomous seafloor acoustic recorders (DASARs) each recorded the acoustic data. An automated detection algorithm was used to isolate individual bowhead whale calls and estimate their locations. Here simultaneous acoustic and visual data from 2007, 2008, and 2010 were merged to determine whether potential variations in calling behavior exist between subadults, adults, and adults with calves. Visual measurements of animal course and direction were used to place bounds on call times and locations that may be associated with that animal versus other animals sighted nearby. Potential differences in call rate and call type were statistically examined. [Work supported by the Shell Exploration and Production Company.]

Acousticians abuse Darwin: Multipath propagation of seismic airgun signals and bowhead whale calls in the Arctic Ocean, with evolutionary implications.

Evolutionary biologists have long noted that animals attempting long‐range communication often adjust their acoustic signals to compensate for or even exploit features of the acoustic propagation path. Here we present recordings of long‐range seismic exploration signals and bowhead whale calls detected on directional autonomous seafloor acoustic recorders deployed in the Beaufort Sea in 2008. Seismic signals that begin as broadband impulses geometrically disperse into FM downsweeps as they propagate over several kilometers range in shallow water waveguides parallel to the Alaskan coast, eventually becoming similar to bowhead FM sweeps in bandwidth and modulated structure, creating headaches for unfortunate scientists trying to create automated bowhead whale acoustic detectors. It is suggested that the similarity in structure between long‐range seismic signals and bowhead whales is not coincidental, but provides an enlightening example of how marine mammals, over evolutionary time scales, have optimized their acoustic signals to propagate long distances while retaining information. A broadband harmonic bowhead whale call is analyzed to examine the speculation. In general, stereotyped baleen whale calls might serve purposes beyond simple species identification in complex waveguide environments.

Acoustic propagation effects on airgun pulses and whale calls in the Beaufort Sea.

In 2008, vocalizing bowhead whales (Balaena mysticetus) conducted their fall migration in the Beaufort Sea amidst short‐ and long‐range seismic exploration activities. Whale calls and airgun pulses were recorded on five arrays of Directional Autonomous Seafloor Acoustic Recorders (DASARs) deployed along 280 km of Alaska’s North Slope. One particular sequence of pulses was recorded from the Louis S. St. Laurent, a Canadian geophysical vessel conducting surveys in deep offshore waters in collaboration with the United States Geological Survey (USGS), thus providing a fix on the location of some distant airgun activity. The distant airgun signals display an interesting selection of long‐range propagation effects, which will be discussed and analyzed in the context of range‐dependent propagation modeling in a combined deep‐water upward‐refracting and shallow‐water bottom‐interacting environment. [Work supported by the U.S. Geological Survey.]

Effects of seismic exploration activities on bowhead whale call distribution in the Alaskan Beaufort Sea.

Airgun use during seismic exploration activities is known to cause a decrease in call detection rates (used here as a proxy for calling rates) in migrating bowhead whales. However, the received airgun sound levels at the whales, at which behavioral changes occur, are not known. To address this question 40 directional autonomous seafloor acoustic recorders (DASARs) were deployed in autumn 2008 at five sites along the migration corridor of bowhead whales during ongoing marine seismic exploration activities. Over 440 000 whale calls were detected and localized by triangulation. Concurrently, over 100 000 airgun pulses were detected and analyzed. The study area was divided into a hexagonal grid (hexagon width 1.75 km). The received level (SPL and cumulative SEL) of airgun sounds was modeled for each hexagon in the study area and each 15‐min period over the entire season, using information on the timing and location of airgun activities, the size of the airgun array being used, the depth of the DASARs recording the airgun sounds, and other covariates. Logistic regression was then used to model occurrence of whale calls as a function of received level, distance offshore, and other covariates. [Work supported by Shell Exploration and Production Company.]

Bowhead Whale (<i>Balaena mysticetus</i>) Migration and Calling Behaviour in the Alaskan Beaufort Sea, Autumn 2001–04: An Acoustic Localization Study

The westward migration of bowhead whales (Bering Sea stock) was studied during autumn 2001–04 as part of an oil industry monitoring program. An array of Directional Autonomous Seafloor Acoustic Recorders (DASARs) was deployed northeast of the Northstar oil production island near Prudhoe Bay, Alaska. Underwater sounds were recorded continuously for 24–35 days per year, mainly in September. More than 130 000 bowhead calls were detected, and the directional capability of DASARs allowed triangulation of whale position for ~93 500 calls. The migration pathway was closer to shore in 2003–04 than in 2001–02. Calls were clumped in space and time, and there was significantly more calling at night than by day. From 65% to 82% of calls were simple frequency-modulated calls, and the percentage of complex calls was positively related to the daily number of calls. No songs were detected, but in 2004 there were numerous call sequences consisting of repeated identical calls in series and generally lasting up to a few minutes. The DASAR methodology provides detailed information on the temporal and spatial distribution of calling whales and on their acoustic repertoire.

Effects of seismic exploration in the Beaufort Sea on bowhead whale call distributions.

We analyzed locations of bowhead whale (Balaena mysticetus) calls in the Beaufort Sea during the fall migrations of 2007 and 2008, including periods of seismic exploration. Analysis examined relationships between received sound levels from airgun pulses and both spatial and temporal distributions of whale call locations. Locations were estimated by triangulation of calls received on Directional Autonomous Seafloor Acoustic Recorders (DASARs) arranged in five arrays along 280 km of Alaska’s coast. Received levels either were measured directly by DASARs or estimated from transmission loss equations. For analysis, a grid was overlaid on the study region, and the 7-week study period was divided into short, regular intervals. Whale calls were counted and received levels estimated within each grid cell and time interval. The relationship between call count and received level was assessed using statistical models that accounted for anthropogenic sounds, date, time of day, bathymetry, and east-west position along the coast. Furthermore, the models incorporated adjustments for detectability (DASARs are less likely to detect more distant calls) and spatial and temporal covariances. Preliminary results indicate declines in calling rate and temporary onshore and offshore shifts in location as received levels increase. [Work supported by Shell Exploration and Production Company.]

Large-scale evaluation of automated detection and localization of bowhead whale calls in the Beaufort Sea.

An automated procedure has been developed for automated detection and localization of bowhead whale sounds with arbitrary frequency-modulated tones and applied to 2008 data collected from 41 directional autonomous seafloor acoustic recorders deployed over a 280-km swath in the Beaufort Sea. The procedure has seven sequential stages: an incoherent spectral band detector, an interval estimator, a feature extractor, a feed-forward neural network classifier, a stage to flag common calls among recorders, bearing extraction, and final localization. The three-layer neural network has ten hidden-layer units and uses 20 features extracted from the spectrogram classifying signals as either whale calls or other sounds. Manual analysis was conducted on recordings from six non-consecutive days chosen to span a variety of seismic activity levels. The first 12 h of each day yielded 141 796 true calls and 1.15×106 other signals that were used to train the network, with 60% of that data used to adjust the weights and 40% used to evaluate convergence. The performance of both the neural network stage and the complete tracking algorithm on the remaining 12 h of manually-analyzed data from each day will be presented. [Work supported by Shell Exploration and Production Company.]

Automated classification of frequency-modulated bowhead whale calls using contour tracing and image segmentation methods.

In 2007, directional autonomous seafloor acoustic recorders (DASARs) were deployed at 35 locations over a 280 km swath in the Beaufort Sea in order to monitor potential changes in bowhead whale (Balaena mysticetus) locations and/or acoustic activity during seismic exploration activities. The large amount of acoustic data generated motivated the development of computer-aided methods to assist in detecting and classifying bowhead whale calls. In the classification stage, bowhead whale calls were divided into six categories: (1) upsweeps, (2) downsweeps, (3) constant calls, (4) u-shaped and (5) n-shaped undulated calls, and (6) complex calls. The frequency range, duration, and fine structure of individual calls vary considerably, creating difficulties when using match-filtering or spectrogram correlation methods. A manually reviewed test data set was assembled, containing examples from each call category, arranged by signal-to-noise ratio. The data set was then used to test several methods (based on image segmentation techniques) for extracting relevant parameters from the signal for subsequent classification. An optimization procedure was then used to generate receiver operating characteristic curves and thus determine appropriate decision boundaries for optimum statistical classifiers. [Work supported by Shell Exploration and Production Company.]

Effects of sounds from seismic exploration on the calling behavior of bowhead whales.

The westward migration of bowhead whales (Balaena mysticetus) was studied during summer and autumn of 2007 to examine the effects of airgun sounds on the whale calling behavior. Whale calls were recorded by 35 directional autonomous seafloor acoustic recorders (DASARs), placed in groups of seven recorders at five locations in the Beaufort Sea covering an east‐west span of 280 km. The directional capability of DASARs allowed triangulation of an estimated whale position for about 130 000 calls. Call detection rates and call locations were compared to the timing of seismic operations—which took place in the center of the study area—and the estimated received levels of airgun sounds at the whale call locations. The analyses showed that seismic operations led to a significant decrease in call detection rates used as a proxy for calling rates. Within about 30 km of the seismic activities, received sound pressure levels from airgun pulses at call locations were greater than 140 dB re 1 μPa for about 20%–40% of calls. Quantile regression analyses showed that seismic activities were correlated to statistically significant shifts in the whales’ distance from the shore, either offshore or inshore. [Study funded by Shell Exploration and Production Company, Alaska.]

Effects of the sounds from an artificial oil production island on bowhead whale calling behavior

The westward migration of bowhead whales (Balaena mysticetus) was studied during autumn 2001–2004 to examine the effects of sounds from an artificial oil production island (Northstar) on whale calling behavior. Whale calls were recorded by an array of directional autonomous seafloor acoustic recorders (DASARs) located 6.4–21.5 km northeast of Northstar in the Beaufort Sea. Simultaneously, a continuous record of sounds produced by Northstar was obtained ∼450 m north of the island. More than 130,000 bowhead calls were detected, and were classified as to their type, including five types of simple calls and one complex call category. In addition, the directional capability of DASARs allowed triangulation of an estimated whale position for a majority of the calls. The primary objectives of the study were to assess any effects of Northstar sounds and environmental‐physical covariates on the duration, midfrequency, frequency range, type of call, and call detection rates. The analyses showed that an increase in transient sounds from Northstar (i.e., boats) resulted in significantly shorter calls. In addition, call detection rates were significantly higher for whales upstream of the array, and the use of complex calls increased significantly as the whales swam westward past Northstar. [Study funded by BP Exploration, Alaska.]