Fin Whale Vocalizations Research Articles

An overview of ambient sound using OOI hydrophone network

The NSF-funded Ocean Observatories Initiative (OOI) is a large-scale project that provides a unique observation of the ocean, collecting acoustic, meteorological, and oceanic datasets. Specifically, for hydrophone data, there are five low frequency hydrophones (Fs = 200 Hz), and six broadband hydrophones (Fs = 64 kHz) dispersed around the north-east pacific. This acoustic dataset presents numerous potential areas of study in the field of ambient sound analysis. In this talk, we analysis 6 years of acoustic data to identify prominent features that are present in the OOI acoustic dataset. Some notable occurrences that are observed in the dataset include volcanic activity, rain and wind noise, fin whale vocalizations, and shipping events. For all low frequency hydrophones and four of the six broadband hydrophones, we will present long-term spectrograms, time-series trends for different spectral bands, different acoustic features that are present in the dataset, and different statistical metrics about the acoustic environment. We find that 6-year, acoustic, trends vary depending on the location of the hydrophone and the spectral band that is observed. Some locations and spectral bands see increases in spectral levels while others see decreases in spectral levels over the course of the 6 years. [Work supported by ONR.]

Food talk: 40-Hz fin whale calls are associated with prey biomass.

Animals use varied acoustic signals that play critical roles in their lives. Understanding the function of these signals may inform about key life-history processes relevant for conservation. In the case of fin whales (Balaenoptera physalus), that produce different call types associated with different behaviours, several hypotheses have emerged regarding call function, but the topic still remains in its infancy. Here, we investigate the potential function of two fin whale vocalizations, the song-forming 20-Hz call and the 40-Hz call, by examining their production in relation to season, year and prey biomass. Our results showed that the production of 20-Hz calls was strongly influenced by season, with a clear peak during the breeding months, and secondarily by year, likely due to changes in whale abundance. These results support the reproductive function of the 20-Hz song used as an acoustic display. Conversely, season and year had no effect on variation in 40-Hz calling rates, but prey biomass did. This is the first study linking 40-Hz call activity to prey biomass, supporting the previously suggested food-associated function of this call. Understanding the functions of animal signals can help identifying functional habitats and predict the negative effects of human activities with important implications for conservation.

Structure from a whale song

Oceanography Probing the structure of the ocean crust requires a wave source. The most common source is an air gun, which is effective but potentially harmful for ocean life and not easy to use everywhere. Kuna and Nabělek found that fin whale songs can also be used as a seismic source for determining crustal structure. Fin whale vocalizations can be as loud as large ships and occur at frequencies useful for traveling through the ocean floor. These properties allow fin whale songs to be used for mapping out the density of ocean crust, a vital part of exploring the seafloor. Science , this issue p. [731][1] [1]: /lookup/doi/10.1126/science.abf3962

Seismic crustal imaging using fin whale songs.

Fin whale calls are among the strongest animal vocalizations that are detectable over great distances in the oceans. We analyze fin whale songs recorded at ocean-bottom seismometers in the northeast Pacific Ocean and show that in addition to the waterborne signal, the song recordings also contain signals reflected and refracted from crustal interfaces beneath the stations. With these data, we constrain the thickness and seismic velocity of the oceanic sediment and basaltic basement and the P-wave velocity of the gabbroic lower crust beneath and around the ocean bottom seismic stations. The abundant and globally available fin whale calls may be used to complement seismic studies in situations where conventional air-gun surveys are not available.

Remote sensing and mapping habitat features pertinent to fin whale life histories in coastal and offshore waters of Vancouver Island, British Columbia

Current knowledge of the population size and movement of fin whales in the northeast Pacific is lacking, but is critical in implementing effective management actions. Here we used ocean gliders to survey the deep-coastal and offshore waters of Vancouver Island to add to what is known about fin whale habitat use in the Canadian Pacific. Passive acoustic recordings were made continuously, from which fin whale vocalizations were used as a proxy to presence. Oceanographic and preyscape variables were also taken to better delineate habitat units in horizontal and vertical space. Comparisons were made between on-shelf and shelf-break zones, and measures taken in- and out-side submarine canyons. Temporal comparisons between winter and spring were also possible.An on-shelf and shelf-break divide was seen, with cooler, fresher waters present closer to the coast for both deployments. The spring deployment showed fin whale acoustic presence focused around topographical features, including the shelf-break and canyons. The most prevalent cue was their 40-Hz call, which suggests foraging. Echosounder readings suggested denser aggregations of larger bodied zooplankton prey were present at canyon heads and along the shelf break. During the winter, calls in the acoustic record were more numerous than in the spring, and dominated by the 20-Hz call. A proportion of the recordings also showed ‘song’ patterning, which suggests breeding activity. Interpolated surfaces were used to examine the spatiotemporal relationships between the high-resolution habitat point data and presence of far-propagating fin whale calls. The predicted use of habitat by fin whales, as represented by their calls, was examined through Random Forest machine learning models using oceanographic and preyscape measures as input variables from these interpolations. Distance from the shelf break and preyscape variables were determined important in predicting whale habitat use in the spring, with model accuracy exceeding 80%. Call presence showed a much weaker relationship to environmental variables for the winter recordings with reduced accuracy of model predictions (13%). An understanding of habitat selection by fin whales may also aid in determining the importance of habitat units and areas that may critical to life history functions.

Comparing Performances of Five Distinct Automatic Classifiers for Fin Whale Vocalizations in Beamformed Spectrograms of Coherent Hydrophone Array

A large variety of sound sources in the ocean, including biological, geophysical, and man-made, can be simultaneously monitored over instantaneous continental-shelf scale regions via the passive ocean acoustic waveguide remote sensing (POAWRS) technique by employing a large-aperture densely-populated coherent hydrophone array system. Millions of acoustic signals received on the POAWRS system per day can make it challenging to identify individual sound sources. An automated classification system is necessary to enable sound sources to be recognized. Here, the objectives are to (i) gather a large training and test data set of fin whale vocalization and other acoustic signal detections; (ii) build multiple fin whale vocalization classifiers, including a logistic regression, support vector machine (SVM), decision tree, convolutional neural network (CNN), and long short-term memory (LSTM) network; (iii) evaluate and compare performance of these classifiers using multiple metrics including accuracy, precision, recall and F1-score; and (iv) integrate one of the classifiers into the existing POAWRS array and signal processing software. The findings presented here will (1) provide an automatic classifier for near real-time fin whale vocalization detection and recognition, useful in marine mammal monitoring applications; and (2) lay the foundation for building an automatic classifier applied for near real-time detection and recognition of a wide variety of biological, geophysical, and man-made sound sources typically detected by the POAWRS system in the ocean.

Temporal–spatial, spectral, and source level distributions of fin whale vocalizations in the Norwegian Sea observed with a coherent hydrophone array

AbstractTo better understand fin whale vocalization behaviour in the Norwegian and Barents Seas, a large-aperture densely sampled coherent hydrophone array was deployed in late winter 2014 to monitor their vocalizations instantaneously over wide areas via passive ocean acoustic waveguide remote sensing (POAWRS). Here, we (i) provide a time-frequency characterization for different call types observed (20 Hz pulses, 130 Hz upsweeps, 30–100 Hz downsweep chirps, and 18–19 Hz backbeats); (ii) compare their relative abundances in three different coastal regions off Alesund, Lofoten, and Northern Finnmark; (iii) estimate the temporal and spatial distributions; (iv) source level distributions; and (v) probability of detection (PoD) regions for the more abundant 20 Hz pulse and 130 Hz upsweep call types. The fin whale vocalizations received over the diel cycle (24 h) were significantly more abundant by a factor of roughly seven off Northern Finnmark than the other two regions, associated with fish feeding activities. The source levels are estimated to be 190.5±7.4 dB for the fin whale 20 Hz pulses and 170.3 ± 5.2 dB for the 130 Hz upsweeps. We find that fin whales are capable of producing each vocalization type either independently or simultaneously with other types, and the 20 Hz sound production in the fin whales involves a mechanism that generates a significantly less-intense second-order harmonic of the fundamental.

Density estimation using ocean-bottom seismometers in Cascadia

The Cascadia Initiative placed a large number of ocean-bottom seismic sensors off the coast of the US Pacific Northwest. These sensors recorded pulsed sounds from fin whales, recorded in conditions that in many cases make it impossible to resolve the sounds from individual whales. The sounds from multiple whales, however, combine to make a distinct rise in spectral energy in the 15–30 Hz band. This received energy can be used to estimate the population density of fin whales in the vicinity of each sensor over the 4-year duration of the project. Difficulties addressed include the variable rate and intensity of fin whale vocalizations, acoustic properties of the seafloor, and differences between seismic and acoustic sensors. [Funding from ONR.]

Comparing neural networks with conventional classifiers for fin whale vocalizations in beamformed spectrograms of coherent hydrophone array

A large variety of sound sources in the ocean, including biological, geophysical and man-made activities can be simultaneously monitored over instantaneous continental-shelf scale regions via the passive ocean acoustic waveguide remote sensing (POAWRS) technique by employing a large-aperture densely-sampled coherent hydrophone array. Millions of acoustic signals received on the POAWRS system per day can make it challenging to identify individual sound sources. An automated classification system is necessary to enable sound sources to be recognized. Here a large training data set of fin whale and other vocalizations are gathered after manual inspection and labelling. Next, multiple classifiers including neural networks, logistic regression, support vector machine (SVM) and decision tree are built and tested for identifying the fin whale and other vocalizations from the enormous amounts of acoustic signals detected per day. The neural network classifier will use beamformed spectrograms to classify acoustic ...

Temporal-spatial and source level distributions for a repertoire of fin whale vocalizations in the Norwegian Sea observed with a coherent hydrophone array

The vocalizations of the fin whale are detected, characterized and localized over instantaneous wide areas of the Norwegian and Barents Seas using a large-aperture densely-sampled coherent hydrohone array via the passive ocean acoustic waveguide remote sensing (POAWRS) technique from observations in late winter to early spring 2014. The fin whale vocalizations are comprised of their characteristic 20 Hz pulses, interspersed by 130 Hz upsweeps, less abundant 30-100 Hz downsweep chirps and 18-19 Hz centered backbeats. The time-frequency characteristics of these vocalization types and their diel occurrence rate time-series are quantified in three distinct regions of the Norwegian Sea, off the coasts of Alesund, Lofoten, and the Northern Finnmark. Their vocalization rates are a factor of roughly 5 times and 17 times larger respectively for the 20 Hz and 130 Hz centered vocalizations off Northern Finnmark than the other regions. The vocalization rate spatial density distributions, source level distributions an...

Spatial and temporal trends in fin whale vocalizations recorded in the NE Pacific Ocean between 2003-2013

In order to study the long-term stability of fin whale (Balaenoptera physalus) singing behavior, the frequency and inter-pulse interval of fin whale 20 Hz vocalizations were observed over 10 years from 2003–2013 from bottom mounted hydrophones and seismometers in the northeast Pacific Ocean. The instrument locations extended from 40°N to 48°N and 130°W to 125°W with water depths ranging from 1500–4000 m. The inter-pulse interval (IPI) of fin whale song sequences was observed to increase at a rate of 0.54 seconds/year over the decade of observation. During the same time period, peak frequency decreased at a rate of 0.17 Hz/year. Two primary call patterns were observed. During the earlier years, the more commonly observed pattern had a single frequency and single IPI. In later years, a doublet pattern emerged, with two dominant frequencies and IPIs. Many call sequences in the intervening years appeared to represent a transitional state between the two patterns. The overall trend was consistent across the entire geographical span, although some regional differences exist. Understanding changes in acoustic behavior over long time periods is needed to help establish whether acoustic characteristics can be used to help determine population identity in a widely distributed, difficult to study species such as the fin whale.

Analysis of fin whale vocalizations south of Rhode Island

Fin whale vocalizations were recorded south of Rhode Island during late summer through early fall of 2015 using a number of underwater recording systems. These systems were deployed to monitor broadband noise, including pile driving, from construction of the Block Island Wind Farm. Two vertical hydrophone array moorings were deployed in approximately 40 m of water each with four hydrophones. Additionally, a tetrahedral array was deployed in about 30 m of water just above the seabed. The tetrahedral array consisted of four hydrophones spaced 0.5 m apart. The spacing between each of these recording systems was approximately 7.5 km. 20-Hz fin whale vocalizations were recorded numerous times on all of the sensors both during and after construction was completed. An analysis and localization effort of these signals was performed to estimate the source level, directionality and the track of the whale over the period of the vocalizations. The results of this analysis will be discussed. [Work supported by the BOEM.]

Detection and localization of whale vocalizations in archived data of seafloor cabled observatories in eastern Japan

Through the examination of archived OBS (Ocean Bottom Seismometer) waveform data of seafloor cabled observatories in the eastern Japan, fin whale vocalizations were detected. Most of those vocalizations were observed with a single OBS because of the sparseness of the OBS deployment. At the observatory off Kushiro-Tokachi in Hokkaido, the fin whale vocalizations were localized using waveform data of both OBS and hydrophone at the same site (Iwase, 2015). At the observatory off Kamaishi in Tohoku District composed of OBSs without a hydrophone, seismic wave velocities in sediment, which are necessary parameters in estimating horizontal range of sound source from apparent emergence angle of transmitted wave at seafloor, were estimated in situ based on air gun signal observation (Iwase, 2016). By using those results, the fin whale vocalizations were localized. At the observatory off Hatsushima Island in Sagami Bay, sperm whale vocalizations were detected in the audible sound signals of a single hydrophone which were recorded on the soundtrack of video tapes with seafloor images. To date, more than 7000 video tapes recorded since 1993 have been archived, and the result of analysis in progress will be shown at the presentation.

Analysis of fin whale vocalizations south of Rhode Island

Fin whale vocalizations were recorded south of Rhode Island during late summer through early fall of 2015 using a number of underwater recording systems. These systems were deployed to monitor broadband noise, including pile driving, from construction of the Block Island Wind Farm. Two vertical hydrophone array moorings were deployed in approximately 40 meters of water each with four hydrophones. Additionally, a tetrahedral array was deployed in about 30 meters of water just above the seabed. The tetrahedral array consisted of four hydrophones spaced 0.5 meters apart. The spacing between each of these recording systems was approximately 7.5 km. 20-Hz fin whale vocalizations were recorded numerous times on all of the sensors both during and after construction was completed. An analysis and localization effort of these signals was performed to estimate the source level, directionality, and the track of the whale over the period of the vocalizations. The results of this analysis will be discussed. [Work supported by the BOEM.]

Characterization of very low frequency ambient noise by principal component analysis

Long-term acoustic monitoring datasets compose a dynamic mixture of many source types, creating a unique noise field in each measurement location and for each recording time. Principal Component Analysis (PCA), and the correlation matrix from which the Principal Components are derived, offer an effective means of estimating the fundamental acoustic contributors, and their portion of the noise field. An improved characterization of ambient noise data sets is discussed, by representing the sound field in a smaller parameter space of principal components, each component reflecting the spectral characteristics of a specific sound source (i.e. fin whale vocalizations). This procedure leads to a definitive parsing of the contributing source types in the measured spectra. The datasets used are recordings from the Comprehensive Nuclear-Test Ban Treaty Organization’s (CTBTO) hydro-acoustic monitoring system.

Annual Acoustic Presence of Fin Whale (Balaenoptera physalus) Offshore Eastern Sicily, Central Mediterranean Sea.

In recent years, an increasing number of surveys have definitively confirmed the seasonal presence of fin whales (Balaenoptera physalus) in highly productive regions of the Mediterranean Sea. Despite this, very little is yet known about the routes that the species seasonally follows within the Mediterranean basin and, particularly, in the Ionian area. The present study assesses for the first time fin whale acoustic presence offshore Eastern Sicily (Ionian Sea), throughout the processing of about 10 months of continuous acoustic monitoring. The recording of fin whale vocalizations was made possible by the cabled deep-sea multidisciplinary observatory, “NEMO-SN1”, deployed 25 km off the Catania harbor at a depth of about 2,100 meters. NEMO-SN1 is an operational node of the European Multidisciplinary Seafloor and water-column Observatory (EMSO) Research Infrastructure. The observatory was equipped with a low-frequency hydrophone (bandwidth: 0.05 Hz–1 kHz, sampling rate: 2 kHz) which continuously acquired data from July 2012 to May 2013. About 7,200 hours of acoustic data were analyzed by means of spectrogram display. Calls with the typical structure and patterns associated to the Mediterranean fin whale population were identified and monitored in the area for the first time. Furthermore, a background noise analysis within the fin whale communication frequency band (17.9–22.5 Hz) was conducted to investigate possible detection-masking effects. The study confirms the hypothesis that fin whales are present in the Ionian Sea throughout all seasons, with peaks in call detection rate during spring and summer months. The analysis also demonstrates that calls were more frequently detected in low background noise conditions. Further analysis will be performed to understand whether observed levels of noise limit the acoustic detection of the fin whales vocalizations, or whether the animals vocalize less in the presence of high background noise.

Fin whale vocalizations observed with ocean bottom seismometers of cabled observatories off east Japan Pacific Ocean

Fin whale vocalizations were found in the archived waveform data from both hydrophones and ocean bottom seismometers (OBSs) of a cabled observatory off Kushiro-Tokachi in Hokkaido. A fin whale was localized on the basis of the incident orientation estimated with a single OBS and the time difference of multipath arrival of sound pressure data from a hydrophone. Furthermore, several fin whale vocalizations were found in the archived OBS waveform data from other cabled observatories off east Japan Pacific Ocean. These findings suggest that the cabled OBSs would be significant apparatuses for real-time monitoring of the presence of baleen whales around Japan.

Sources and levels of ambient ocean sound near the Antarctic Peninsula.

Arrays of hydrophones were deployed within the Bransfield Strait and Scotia Sea (Antarctic Peninsula region) from 2005 to 2009 to record ambient ocean sound at frequencies of up to 125 and 500 Hz. Icequakes, which are broadband, short duration signals derived from fracturing of large free-floating icebergs, are a prominent feature of the ocean soundscape. Icequake activity peaks during austral summer and is minimum during winter, likely following freeze-thaw cycles. Iceberg grounding and rapid disintegration also releases significant acoustic energy, equivalent to large-scale geophysical events. Overall ambient sound levels can be as much as ~10–20 dB higher in the open, deep ocean of the Scotia Sea compared to the relatively shallow Bransfield Strait. Noise levels become lowest during the austral winter, as sea-ice cover suppresses wind and wave noise. Ambient noise levels are highest during austral spring and summer, as surface noise, ice cracking and biological activity intensifies. Vocalizations of blue (Balaenoptera musculus) and fin (B. physalus) whales also dominate the long-term spectra records in the 15–28 and 89 Hz bands. Blue whale call energy is a maximum during austral summer-fall in the Drake Passage and Bransfield Strait when ambient noise levels are a maximum and sea-ice cover is a minimum. Fin whale vocalizations were also most common during austral summer-early fall months in both the Bransfield Strait and Scotia Sea. The hydrophone data overall do not show sustained anthropogenic sources (ships and airguns), likely due to low coastal traffic and the typically rough weather and sea conditions of the Southern Ocean.

Fin whale vocalization classification and abundance estimation

Several thousand fin whale vocalizations from multiple fin individuals were passively recorded by a high-resolution coherent hydrophone array system in the Gulf of Maine in Fall 2006. The recorded fin whale vocalizations have relatively short durations roughly 0.4 s and frequencies ranging from 15 to 40 Hz. Here we classify the fin whale vocalizations and apply the results to estimate the minimum number of vocalizing fin individuals detected by our hydrophone array. The horizontal azimuth or bearing of each fin whale vocalization is first determined by beamforming. Each beamformed fin whale vocalization spectrogram is next characterized by several features such as center frequency, upper and lower frequency limits, as well as amplitude-weighted mean frequency. The vocalizations are then classified using k-mean clustering into several distinct vocal types. The vocalization clustering result is then combined with the bearing-time trajectory information for a consecutive sequence of vocalizations to provide an estimate of the minimum number of vocalizing fin individuals detected.

An automatic single station multipath ranging technique for 20 Hz fin whale calls: Applications for ocean bottom seismometer data in the northeast Pacific

Ocean bottom seismometers (OBSs) deployed to monitor earthquakes also record 20 Hz fin whale vocalizations. Because OBSs are often deployed too far apart to record a call on more than one instrument, and call detections are dependent on the acoustic environment, techniques are required to determine call density from a single instrument. One approach is to use point transect distance sampling, which relies on estimates of horizontal range to each call. An automated technique has been developed to estimate range based on the amplitude and relative timing of multipath arrivals assuming a uniform water depth. The method has been evaluated using independently located calls recorded by a seismic network on the Endeavour Segment of the Juan de Fuca Ridge, a region of rough and partially sedimented seafloor. Results suggest that the method works well to ranges of ~5 km but at larger ranges ambiguities arise. Calls at >10 km range are sometimes located between 5 and 10 km because the times are erroneously fit by one too few water column multiples and the amplitudes have too much scatter to be diagnostic of range. The technique is presently being applied in different acoustic settings to determine whether it is more reliable in flat sedimented regions.