BASSA: New software tool reveals hidden details in visualisation of low-frequency animal sounds

The study of animal sounds in biology and ecology relies heavily upon time-frequency (TF) visualisation, most commonly using the short-time Fourier transform (STFT) spectrogram. This method, however, has inherent bias towards either temporal or spectral details that can lead to misinterpretation of complex animal sounds. An ideal TF visualisation should accurately convey the structure of the sound in terms of both frequency and time, however, the STFT often cannot meet this requirement. We evaluate the accuracy of four TF visualisation methods (superlet transform [SLT], continuous wavelet transform [CWT] and two STFTs) using a synthetic test signal. We then apply these methods to visualise sounds of the Chagos blue whale, Asian elephant, southern cassowary, eastern whipbird, mulloway fish and the American crocodile. We show that the SLT visualises the test signal with 18.48%-28.08% less error than the other methods. A comparison between our visualisations of animal sounds and their literature descriptions indicates that the STFT's bias may have caused misinterpretations in describing pygmy blue whale songs and elephant rumbles. We suggest that use of the SLT to visualise low-frequency animal sounds may prevent such misinterpretations. Finally, we employ the SLT to develop 'BASSA', an open-source, GUI software application that offers a no-code, user-friendly tool for analysing short-duration recordings of low-frequency animal sounds for the Windows platform. The SLT visualises low-frequency animal sounds with improved accuracy, in a user-friendly format, minimising the risk of misinterpretation while requiring less technical expertise than the STFT. Using this method could propel advances in acoustics-driven studies of animal communication, vocal production methods, phonation and species identification.

Application of a seismic network to baleen whale call detection and localization in the Panama basin-a Bryde's whale example

Baleen whales use sounds of various characteristics for different tasks and interactions. This study focuses on recordings from the Costa Rica Rift, in the Eastern Tropical Pacific Ocean, made by 25 ocean-bottom seismographs and a vertical array of 12 hydrophones between January and February 2015. The whale calls observed are of two kinds: more commonly, repetitive 4-5 s-long signals separated into two frequency bands centered at ∼20 and ∼36 Hz; less commonly, a series of ∼0.5 to 1.0 s-long, lower amplitude signals with frequencies between 80 and 160 Hz. These characteristics are similar to calls attributed to Bryde's whales which are occasionally sighted in this region. In this study, the repetitive calls are detected using both the short-term average/long-term average approach and a network empirical subspace detector. In total, 188 and 1891 calls are obtained for each method, demonstrating the value of the subspace detector for highly similar signals. These signals are first localized using a non-linear grid search algorithm and then further relocalized using the double-difference technique. The high-resolution localizations reveal the presence of at least seven whales during the recording period, often crossing the instrument network from southwest to northeast.

Fin whale song evolution in the North Atlantic

Animal songs can change within and between populations as the result of different evolutionary processes. When these processes include cultural transmission, the social learning of information or behaviours from conspecifics, songs can undergo rapid evolutions because cultural novelties can emerge more frequently than genetic mutations. Understanding these song variations over large temporal and spatial scales can provide insights into the patterns, drivers and limits of song evolution that can ultimately inform on the species' capacity to adapt to rapidly changing acoustic environments. Here, we analysed changes in fin whale (Balaenoptera physalus) songs recorded over two decades across the central and eastern North Atlantic Ocean. We document a rapid replacement of song INIs (inter-note intervals) over just four singing seasons, that co-occurred with hybrid songs (with both INIs), and a clear geographic gradient in the occurrence of different song INIs during the transition period. We also found gradual changes in INIs and note frequencies over more than a decade with fin whales adopting song changes. These results provide evidence of vocal learning in fin whales and reveal patterns of song evolution that raise questions on the limits of song variation in this species.

Ecological inferences about marine mammals from passive acoustic data

Monitoring on the basis of sound recordings, or passive acoustic monitoring, can complement or serve as an alternative to real-time visual or aural monitoring of marine mammals and other animals by human observers. Passive acoustic data can support the estimation of common, individual-level ecological metrics, such as presence, detection-weighted occupancy, abundance and density, population viability and structure, and behaviour. Passive acoustic data also can support estimation of some community-level metrics, such as species richness and composition. The feasibility of estimation and certainty of estimates is highly context dependent, and understanding the factors that affect the reliability of measurements is useful for those considering whether to use passive acoustic data. Here, we review basic concepts and methods of passive acoustic sampling in marine systems that often are applicable to marine mammal research and conservation. Our ultimate aim is to facilitate collaboration among ecologists, bioacousticians, and data analysts. Ecological applications of passive acoustics require one to make decisions about sampling design, which in turn requires consideration of sound propagation, sampling of signals, and data storage. One also must make decisions about signal detection and classification and evaluation of the performance of algorithms for these tasks. Investment in the research and development of systems that automate detection and classification, including machine learning, are increasing. Passive acoustic monitoring is more reliable for detection of species presence than for estimation of other species-level metrics. Use of passive acoustic monitoring to distinguish among individual animals remains difficult. However, information about detection probability, vocalisation or cue rate, and relations between vocalisations and the number and behaviour of animals increases the feasibility of estimating abundance or density. Most sensor deployments are fixed in space or are sporadic, making temporal turnover in species composition more tractable to estimate than spatial turnover. Collaborations between acousticians and ecologists are most likely to be successful and rewarding when all partners critically examine and share a fundamental understanding of the target variables, sampling process, and analytical methods.

Basin scale coherence of Kauai-Beacon m-sequence transmissions received at Wake Island and Monterey, CA

The 75 Hz Kauai-Beacon source is well-situated for observing the North Pacific Ocean acoustically, and ongoing efforts enable transmissions and analysis of broadband signals in 2023 and beyond. This is the first demonstration of acoustic receiving along paths to Wake Island (∼3500 km) and Monterey Bay (∼4000 km). The 44 received m-sequence waveforms exhibit excellent phase stability with processing gain approaching the maximum theoretical gain evaluated over the 20 min signal transmission duration. The article concludes with a discussion on the future source utility and highlights research topics of interest, including observed Doppler (waveform dilation), thermometry, and tomography.

Temporal evolution of the Mediterranean fin whale song

We present an analysis of fin whale (Balaenoptera physalus) songs on passive acoustic recordings from the Pelagos Sanctuary (Western Mediterranean Basin). The recordings were gathered between 2008 and 2018 using 2 different hydrophone stations. We show how 20 Hz fin whale pulses can be automatically detected using a low complexity convolutional neural network (CNN) despite data variability (different recording devices exposed to diverse noises). The pulses were further classified into the two categories described in past studies and inter pulse intervals (IPI) were measured. The results confirm previous observations on the local relationship between pulse type and IPI with substantially more data. Furthermore we show inter-annual shifts in IPI and an intra-annual trend in pulse center frequency. This study provides new elements of comparison for the understanding of long term fin whale song trends worldwide.

Mathematical models of long term evolution of blue whale song types' frequencies

The linear decrease in the frequency of blue whale songs around the world is, to date, an unexplained phenomenon. We show it can be reproduced by a mathematical model considering two antagonistic behavioral trends: first, a bias towards conformity in the song, and second, a tendency to try and sing lower than the other whales. We check the robustness of our model by considering some more complex premises. First, different hierarchical relations between the singers are explored, adapting methods used in the flocking motion studies. Then a population-dependant simulation shows that even considering the gradual addition of new whales, the evolution is still globally linear. Finally, we show that intra-annual variations surging from different causes can be naturally incorporated into the model. We then conclude that, unlike other explanations, a cultural hypothesis seems compatible with the observed linearity of the blue whales's songs frequency shift.

Update on frequency decline of Northeast Pacific blue whale (Balaenoptera musculus) calls

Worldwide, the frequency (pitch) of blue whale (Balaenoptera musculus) calls has been decreasing since first recorded in the 1960s. This frequency decline occurs over annual and inter-annual timescales and has recently been documented in other baleen whale species, yet it remains unexplained. In the Northeast Pacific, blue whales produce two calls, or units, that, when regularly repeated, are referred to as song: A and B calls. In this population, frequency decline has thus far only been examined in B calls. In this work, passive acoustic data collected in the Southern California Bight from 2006 to 2019 were examined to determine if A calls are also declining in frequency and whether the call pulse rate was similarly impacted. Additionally, frequency measurements were made for B calls to determine whether the rate of frequency decline is the same as was calculated when this phenomenon was first reported in 2009. We found that A calls decreased at a rate of 0.32 Hz yr^-1 during this period and that B calls were still decreasing, albeit at a slower rate (0.27 Hz yr^-1) than reported previously. The A call pulse rate also declined over the course of the study, at a rate of 0.006 pulses/s yr^-1. With this updated information, we consider the various theories that have been proposed to explain frequency decline in blue whales. We conclude that no current theory adequately accounts for all aspects of this phenomenon and consider the role that individual perception of song frequency may play. To understand the cause behind call frequency decline, future studies might want to explore the function of these songs and the mechanism for their synchronization. The ubiquitous nature of the frequency shift phenomenon may indicate a consistent level of vocal plasticity and fine auditory processing abilities across baleen whale species.

Characterization of fin whale song off the Western Antarctic Peninsula

Song is produced by a variety of terrestrial and marine animals and is particularly common among baleen whales. Fin whale (Balaenoptera physalus) song is comprised of relatively simple 20 Hz pulses produced at regular intervals. The timing of these intervals, in addition to the presence and frequency of overtones, appears to be unique to each population. The purpose of this study was to characterize Western Antarctic Peninsula fin whale song and describe temporal pattern variations in song type and occurrence. Recordings were collected in the area from 2001-2004 and again 2014-2016. One song type was identified with a primary inter-pulse interval (IPI) of approximately 14 s and secondary IPI of 12.5 s. This song occurred in three pattern variants: singlet, doublet, and long triplet. The interval between pulses increased by 1.5 s between recording periods while the frequency of the overtones decreased from 89 Hz to 86 Hz. Song was never recorded in August and while it was recorded at other times in some years, it was consistently present in recordings from April through June across all years. While multiple pattern variants were present each year, singlets were generally the most prevalent variant. Doublets and triplets occurred from February through June, with highest levels of variants in February. In later years the triplet variant presence increased and in 2016 it comprised 53% of recorded song bouts. Further research is needed to understand the reasons why song changes over time and to examine the feasibility of using song to delineate and identify populations.

Seasonal Trends and Diel Patterns of Downsweep and SEP Calls in Chilean Blue Whales

To learn more about the occurrence and behaviour of a recently discovered population of blue whales, passive acoustic data were collected between January 2012 and April 2013 in the Chiloense ecoregion of southern Chile. Automatic detectors and manual auditing were used to detect blue whale songs (SEP calls) and D calls, which were then analysed to gain insights into temporal calling patterns. We found that D call rates were extremely low during winter (June–August) but gradually increased in spring and summer, decreasing again later during fall. SEP calls were absent for most winter and spring months (July–November) but increased in summer and fall, peaking between March and April. Thus, our results support previous studies documenting the austral summer residency of blue whales in this region, while suggesting that some individuals stay longer, highlighting the importance of this area as a blue whale habitat. We also investigated the daily occurrence of each call type and found that D calls occurred more frequently during dusk and night hours compared to dawn and day periods, whereas SEP calls did not show any significant diel patterns. Overall, these findings help to understand the occurrence and behaviour of endangered Chilean blue whales, enhancing our ability to develop conservation strategies in this important Southern Hemisphere habitat.

Sounding the Call for a Global Library of Underwater Biological Sounds

Aquatic environments encompass the world’s most extensive habitats, rich with sounds produced by a diversity of animals. Passive acoustic monitoring (PAM) is an increasingly accessible remote sensing technology that uses hydrophones to listen to the underwater world and represents an unprecedented, non-invasive method to monitor underwater environments. This information can assist in the delineation of biologically important areas via detection of sound-producing species or characterization of ecosystem type and condition, inferred from the acoustic properties of the local soundscape. At a time when worldwide biodiversity is in significant decline and underwater soundscapes are being altered as a result of anthropogenic impacts, there is a need to document, quantify, and understand biotic sound sources–potentially before they disappear. A significant step toward these goals is the development of a web-based, open-access platform that provides: (1) a reference library of known and unknown biological sound sources (by integrating and expanding existing libraries around the world); (2) a data repository portal for annotated and unannotated audio recordings of single sources and of soundscapes; (3) a training platform for artificial intelligence algorithms for signal detection and classification; and (4) a citizen science-based application for public users. Although individually, these resources are often met on regional and taxa-specific scales, many are not sustained and, collectively, an enduring global database with an integrated platform has not been realized. We discuss the benefits such a program can provide, previous calls for global data-sharing and reference libraries, and the challenges that need to be overcome to bring together bio- and ecoacousticians, bioinformaticians, propagation experts, web engineers, and signal processing specialists (e.g., artificial intelligence) with the necessary support and funding to build a sustainable and scalable platform that could address the needs of all contributors and stakeholders into the future.

Performance metrics for marine mammal signal detection and classification

Automatic algorithms for the detection and classification of sound are essential to the analysis of acoustic datasets with long duration. Metrics are needed to assess the performance characteristics of these algorithms. Four metrics for performance evaluation are discussed here: receiver-operating-characteristic (ROC) curves, detection-error-trade-off (DET) curves, precision-recall (PR) curves, and cost curves. These metrics were applied to the generalized power law detector for blue whale D calls [Helble, Ierley, D'Spain, Roch, and Hildebrand (2012). J. Acoust. Soc. Am. 131(4), 2682-2699] and the click-clustering neural-net algorithm for Cuvier's beaked whale echolocation click detection [Frasier, Roch, Soldevilla, Wiggins, Garrison, and Hildebrand (2017). PLoS Comp. Biol. 13(12), e1005823] using data prepared for the 2015 Detection, Classification, Localization and Density Estimation Workshop. Detection class imbalance, particularly the situation of rare occurrence, is common for long-term passive acoustic monitoring datasets and is a factor in the performance of ROC and DET curves with regard to the impact of false positive detections. PR curves overcome this shortcoming when calculated for individual detections and do not rely on the reporting of true negatives. Cost curves provide additional insight on the effective operating range for the detector based on the a priori probability of occurrence. Use of more than a single metric is helpful in understanding the performance of a detection algorithm.

Geographically distinct blue whale song variants in the Northeast Pacific

The Northeast Pacific (NEP) population of blue whales Balaenoptera musculus musculus is currently managed as a single stock. We investigated the fine-scale frequency characteristics of 1 NEP blue whale song unit, the B call. We analyzed B calls from passive acoustic data collected between 2010 and 2013 at 2 low-latitude sites, Palmyra Atoll and the Hawaiian Islands, and 3 higher-latitude sites, off southern California, off Washington state and in the Gulf of Alaska. Frequency measurements were extracted along the contour of the third harmonic from each call, and data from each region were compared. Calls from the Gulf of Alaska and Hawai‘i presented a downshift in frequency, beginning just past the midway point of the contour, which was not present in calls recorded from southern California or Palmyra Atoll. Calls from Washington displayed intermediate characteristics between those from the other 2 high-latitude sites. Cluster analysis resulted in consistent grouping of call contours from Washington and southern California, in what we termed the NEP B1 variant, while contours from Hawai‘i and the Gulf of Alaska were grouped together, as a NEP B2 variant. Frequency differences were also observed among the variants; the Gulf of Alaska displayed the highest frequency on average, followed by Washington, then southern California. Consistent with other studies, a yearly decline in the frequency of B calls was also observed. This discovery of at least 2 geographically distinct variants provides the first evidence of vocally distinct subpopulations within the NEP, indicating the possibility of a need for finer-scale population segmentation.

All units are equal in humpback whale songs, but some are more equal than others

Flexible production and perception of vocalizations is linked to an impressive array of cognitive capacities including language acquisition by humans, song learning by birds, biosonar in bats, and vocal imitation by cetaceans. Here, we characterize a portion of the repertoire of one of the most impressive vocalizers in nature: the humpback whale. Qualitative and quantitative analyses of sounds (units) produced by humpback whales revealed that singers gradually morphed streams of units along multiple acoustic dimensions within songs, maintaining the continuity of spectral content across subjectively dissimilar unit "types." Singers consistently produced some unit forms more frequently and intensely than others, suggesting that units are functionally heterogeneous. The precision with which singing humpback whales continuously adjusted the acoustic characteristics of units shows that they possess exquisite vocal control mechanisms and vocal flexibility beyond what is seen in most animals other than humans. The gradual morphing of units within songs that we observed is inconsistent with past claims that humpback whales construct songs from a fixed repertoire of discrete unit types. These findings challenge the results of past studies based on fixed-unit classification methods and argue for the development of new metrics for characterizing the graded structure of units. The specific vocal variations that singers produced suggest that humpback whale songs are unlikely to provide detailed information about a singer's reproductive fitness, but can reveal the precise locations and movements of singers from long distances and may enhance the effectiveness of units as sonar signals.

Source level and vocalizing depth estimation of two blue whale subspecies in the western Indian Ocean from single sensor observations

The source level (SL) and vocalizing source depth (SD) of individuals from two blue whale (BW) subspecies, an Antarctic blue whale (Balaenoptera musculus intermedia; ABW) and a Madagascar pygmy blue whale (Balaenoptera musculus brevicauda; MPBW) are estimated from a single bottom-mounted hydrophone in the western Indian Ocean. Stereotyped units (male) are automatically detected and the range is estimated from the time delay between the direct and lowest-order multiply-reflected acoustic paths (multipath-ranging). Allowing for geometric spreading and the Lloyd's mirror effect (range-, depth-, and frequency-dependent) SL and SD are estimated by minimizing the SL variance over a series of units from the same individual over time (and hence also range). The average estimated SL of 188.5 ± 2.1 dB re 1μPa measured between [25-30] Hz for the ABW and 176.8 ± 1.8 dB re. 1μPa measured between [22-27] Hz for the MPBW agree with values published for other geographical areas. Units were vocalized at estimated depths of 25.0 ± 3.7 and 32.7 ± 5.7 m for the ABW Unit A and C and, ≃20 m for the MPBW. The measurements show that these BW calls series are stereotyped in frequency, amplitude, and depth.

Detecting, classifying, and counting blue whale calls with Siamese neural networks

The goal of this project is to use acoustic signatures to detect, classify, and count the calls of four acoustic populations of blue whales so that, ultimately, the conservation status of each population can be better assessed. We used manual annotations from 350 h of audio recordings from the underwater hydrophones in the Indian Ocean to build a deep learning model to detect, classify, and count the calls from four acoustic song types. The method we used was Siamese neural networks (SNN), a class of neural network architectures that are used to find the similarity of the inputs by comparing their feature vectors, finding that they outperformed the more widely used convolutional neural networks (CNN). Specifically, the SNN outperform a CNN with 2% accuracy improvement in population classification and 1.7%-6.4% accuracy improvement in call count estimation for each blue whale population. In addition, even though we treat the call count estimation problem as a classification task and encode the number of calls in each spectrogram as a categorical variable, SNN surprisingly learned the ordinal relationship among them. SNN are robust and are shown here to be an effective way to automatically mine large acoustic datasets for blue whale calls.

Multiple pygmy blue whale acoustic populations in the Indian Ocean: whale song identifies a possible new population

Blue whales were brought to the edge of extinction by commercial whaling in the twentieth century and their recovery rate in the Southern Hemisphere has been slow; they remain endangered. Blue whales, although the largest animals on Earth, are difficult to study in the Southern Hemisphere, thus their population structure, distribution and migration remain poorly known. Fortunately, blue whales produce powerful and stereotyped songs, which prove an effective clue for monitoring their different ‘acoustic populations.’ The DGD-Chagos song has been previously reported in the central Indian Ocean. A comparison of this song with the pygmy blue and Omura’s whale songs shows that the Chagos song are likely produced by a distinct previously unknown pygmy blue whale population. These songs are a large part of the underwater soundscape in the tropical Indian Ocean and have been so for nearly two decades. Seasonal differences in song detections among our six recording sites suggest that the Chagos whales migrate from the eastern to western central Indian Ocean, around the Chagos Archipelago, then further east, up to the north of Western Australia, and possibly further north, as far as Sri Lanka. The Indian Ocean holds a greater diversity of blue whale populations than thought previously.

Long-term frequency changes of a potential baleen whale call from the central Indian Ocean during 2002-2019

The multiple baleen whales of the central Indian Ocean use distinct calls, with their acoustic signatures marking their respective geographic distributions. This paper uses observations from Diego Garcia to track long-term changes (2002-2019) in calls produced by an unidentified whale. The calls around 20-45 Hz consist of closely spaced frequency tones that resemble a comb, followed by a downsweep. The observations show that while the average comb frequencies steadily increase, the downsweep portion decreases. Some frequencies disappear, while new ones appear. These frequency-observations are different from similar studies of other baleen whales in the region, which mostly show a decrease.

An open access dataset for developing automated detectors of Antarctic baleen whale sounds and performance evaluation of two commonly used detectors

Since 2001, hundreds of thousands of hours of underwater acoustic recordings have been made throughout the Southern Ocean south of 60° S. Detailed analysis of the occurrence of marine mammal sounds in these circumpolar recordings could provide novel insights into their ecology, but manual inspection of the entirety of all recordings would be prohibitively time consuming and expensive. Automated signal processing methods have now developed to the point that they can be applied to these data in a cost-effective manner. However training and evaluating the efficacy of these automated signal processing methods still requires a representative annotated library of sounds to identify the true presence and absence of different sound types. This work presents such a library of annotated recordings for the purpose of training and evaluating automated detectors of Antarctic blue and fin whale calls. Creation of the library has focused on the annotation of a representative sample of recordings to ensure that automated algorithms can be developed and tested across a broad range of instruments, locations, environmental conditions, and years. To demonstrate the utility of the library, we characterise the performance of two automated detection algorithms that have been commonly used to detect stereotyped calls of blue and fin whales. The availability of this library will facilitate development of improved detectors for the acoustic presence of Southern Ocean blue and fin whales. It can also be expanded upon to facilitate standardization of subsequent analysis of spatiotemporal trends in call-density of these circumpolar species.

A new blue whale song-type described for the Arabian Sea and Western Indian Ocean

Blue whales Balaenoptera musculus in the Indian Ocean (IO) are currently thought to represent 2 or 3 subspecies (B. m. intermedia, B. m. brevicauda, B. m. indica), and believed to be structured into 4 populations, each with a diagnostic song-type. Here we describe a previously unreported song-type that implies the probable existence of a population that has been undetected or conflated with another population. The novel song-type was recorded off Oman in the northern IO/Arabian Sea, off the western Chagos Archipelago in the equatorial central IO, and off Madagascar in the southwestern IO. As this is the only blue whale song that has been identified in the western Arabian Sea, we label it the ‘Northwest Indian Ocean’ song-type to distinguish it from other regional song-types. Spatiotemporal variation suggested a distribution west of 70°E, with potential affinity for the northern IO/Arabian Sea, and only minor presence in the southwestern IO. Timing of presence off Oman suggested that intensive illegal Soviet whaling that took 1294 blue whales in the 1960s likely targeted this population, as opposed to the more widely distributed ‘Sri Lanka’ acoustic population as previously assumed. Based upon geographic distribution and potential aseasonal reproduction found in the Soviet catch data, we suggest that if there is a northern IO subspecies (B. m. indica), it is likely this population. Moreover, the potentially restricted range, intensive historic whaling, and the fact that the song-type has been previously undetected, suggests a small population that is in critical need of status assessment and conservation action.

Acoustic seasonality, behaviour and detection ranges of Antarctic blue and fin whales under different sea ice conditions off Antarctica

Descriptions of seasonal occurrence and behaviour of Antarctic blue and fin whales in the Southern Ocean are of pivotal importance for the effective conservation and management of these endangered species. We used an autonomous acoustic recorder to collect bioacoustic data from January through September 2014 to describe the seasonal occurrence, behaviour and detection ranges of Antarctic blue and fin whale calls off the Maud Rise, Antarctica. From 2479 h of recordings, we detected D- and Z-calls plus the 27 Hz chorus of blue whales, the 20 and 99 Hz pulses of fin whales and the 18-28 Hz chorus of blue and fin whales. Blue whale calls were detected throughout the hydrophone deployment period with a peak occurrence in February, indicating continuous presence of whales in a broad Southern Ocean area (given the modelled detection ranges). Fin whale calls were detected from January through July when sea ice was present on the latter dates. No temporal segregation in peaks of diel calling rates of blue and fin whales was observed in autumn, but a clear temporal segregation was apparent in summer. Acoustic propagation models suggest that blue and fin whale calls can be heard as far as 1700 km from the hydrophone position in spring. Random forest models ranked month of the year as the most important predictor of call occurrence and call rates (i.e. behaviour) for these whales. Our work highlights areas around the Maud Rise as important habitats for blue and fin whales in the Southern Ocean.

Inter-annual decrease in pulse rate and peak frequency of Southeast Pacific blue whale song types

A decrease in the frequency of two southeast Pacific blue whale song types was examined over decades, using acoustic data from several different sources in the eastern Pacific Ocean ranging between the Equator and Chilean Patagonia. The pulse rate of the song units as well as their peak frequency were measured using two different methods (summed auto-correlation and Fourier transform). The sources of error associated with each measurement were assessed. There was a linear decline in both parameters for the more common song type (southeast Pacific song type n.2) between 1997 to 2017. An abbreviated analysis, also showed a frequency decline in the scarcer southeast Pacific song type n.1 between 1970 to 2014, revealing that both song types are declining at similar rates. We discussed the use of measuring both pulse rate and peak frequency to examine the frequency decline. Finally, a comparison of the rates of frequency decline with other song types reported in the literature and a discussion on the reasons of the frequency shift are presented.

Comment on "A standardized method of classifying pulsed sounds and its application to pulse rate measurement of blue whale southeast Pacific song units" [J. Acoust. Soc. Am. 146, 2145-2154 (2019)]

Patris, Malige, Glotin, Asch, and Buchan [(2019). Acoust. Soc. Am. 146, 2145-2154] proposed a technique for classifying and describing pulsed sounds produced by whales that can improve the precision and objectivity of acoustic measurements from song units. Their analyses revealed that blue whales produce units at precise pulse rates. The structure and precision of the song phrase they describe is remarkably similar to what is seen in song production by chickadees. In both species, precise control of shifts in pulse rate may produce reverberation-related cues that enable listeners to localize singers. The techniques developed by Patris and colleagues thus can provide more accurate measures as well as insights into how animals produce and use songs.

Machine learning in acoustics: Theory and applications

Acoustic data provide scientific and engineering insights in fields ranging from biology and communications to ocean and Earth science. We survey the recent advances and transformative potential of machine learning (ML), including deep learning, in the field of acoustics. ML is a broad family of techniques, which are often based in statistics, for automatically detecting and utilizing patterns in data. Relative to conventional acoustics and signal processing, ML is data-driven. Given sufficient training data, ML can discover complex relationships between features and desired labels or actions, or between features themselves. With large volumes of training data, ML can discover models describing complex acoustic phenomena such as human speech and reverberation. ML in acoustics is rapidly developing with compelling results and significant future promise. We first introduce ML, then highlight ML developments in four acoustics research areas: source localization in speech processing, source localization in ocean acoustics, bioacoustics, and environmental sounds in everyday scenes.

A standardized method of classifying pulsed sounds and its application to pulse rate measurement of blue whale southeast Pacific song units

Pulsed sounds are an interesting example of complex biological sounds, frequent in cetaceans' vocalizations. A two-group classification of these sounds is proposed: tonal or non-tonal. Through two simple mathematical models, it is shown that this classification can lead to better techniques for measuring the pulse rate. This classification is thus useful for improving measurement accuracy, but can also help in formulating hypotheses regarding mechanisms of sound production. This method of classification is applied to south Pacific blue whale vocalizations and it is found that the pulse rate corresponds to the fundamental frequency (not expressed in the spectrum) of the song. Thus, the hypothesis that the sound is produced by only one organ and then filtered by the body of the giant is reinforced.

Two unit analysis of Sri Lankan pygmy blue whale song over a decade

Sri Lankan pygmy blue whale song consists of three repeated units: (1) low frequency pulsive unit, (2) frequency modulated (FM) upsweep, and (3) long tonal downsweep. The Unit 2 FM unit has up to three visible upsweeps with energy concentrated at approximately 40, 50, and 60 Hz, while the Unit 3 (∼100 Hz) tonal downsweep is the most distinct unit lasting 20-30 s. Spectral characteristics of the Units 2 and 3 song elements, along with ocean sound levels, were analyzed in the Indian Ocean from 2002 to 2013. The peak frequency of the tonal Unit 3 calls decreased from approximately 106.5 to 100.7 Hz over a decade corresponding to a 5.4% decrease. Over the same time period, the frequency content of the Unit 2 upsweeps did not change as dramatically with only a 3.1% change. Ambient sound levels in the vocalization bands did not exhibit equivalent patterns in amplitude trends. Analysis showed no increase in the ambient sound or compensated peak amplitude levels of the tonal downsweeps, eliminating the presence of a Lombard effect. Here it is proposed that each song unit may convey different information and thus may be responding to different selective pressures.

Sparse representation-based classification of mysticete calls

This paper presents an automatic classification method dedicated to mysticete calls. This method relies on sparse representations which assume that mysticete calls lie in a linear subspace described by a dictionary-based representation. The classifier accounts for noise by refusing to assign the observed signal to a given class if it is not included into the linear subspace spanned by the dictionaries of mysticete calls. Rejection of noise is achieved without feature learning. In addition, the proposed method is modular in that, call classes can be appended to or removed from the classifier without requiring retraining. The classifier is easy to design since it relies on a few parameters. Experiments on five types of mysticete calls are presented. It includes Antarctic blue whale Z-calls, two types of "Madagascar" pygmy blue whale calls, fin whale 20 Hz calls and North-Pacific blue whale D-calls. On this dataset, containing 2185 calls and 15 000 noise samples, an average recall of 96.4% is obtained and 93.3% of the noise data (persistent and transient) are correctly rejected by the classifier.

On the reliability of acoustic annotations and automatic detections of Antarctic blue whale calls under different acoustic conditions

Evaluation of the performance of computer-based algorithms to automatically detect mammalian vocalizations often relies on comparisons between detector outputs and a reference data set, generally obtained by manual annotation of acoustic recordings. To explore the reproducibility of these annotations, inter- and intra-analyst variability in manually annotated Antarctic blue whale (ABW) Z-calls are investigated by two analysts in acoustic data from two ocean basins representing different scenarios in terms of call abundance and background noise. Manual annotations exhibit strong inter- and intra-analyst variability, with less than 50% agreement between analysts. This variability is mainly caused by the difficulty of reliably and reproducibly distinguishing single calls in an ABW chorus made of overlaying distant calls. Furthermore, the performance of two automated detectors, based on spectrogram correlation or subspace-detection strategy, is evaluated by comparing detector output to a "conservative" manually annotated reference data set, which comprises only analysts' matching events. This study highlights the need for a standardized approach for human annotations and automatic detections, including a quantitative description of their performance, to improve the comparability of acoustic data, which is particularly relevant in the context of collaborative approaches in collecting and analyzing large passive acoustic data sets.

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.

Decadal-scale frequency shift of migrating bowhead whale calls in the shallow Beaufort Sea

Automated and manual acoustic localizations of bowhead whale calls in the Beaufort Sea were used to estimate the minimum frequency attained by their highly variable FM-modulated call repertoire during seven westerly fall migrations. Analyses of 13 355 manual and 100 009 automated call localizations found that between 2008 and 2014 the proportion of calls that dipped below 75 Hz increased from 27% to 41%, shifting the mean value of the minimum frequency distribution from 94 to 84 Hz. Multivariate regression analyses using both generalized linear models and generalized estimating equations found that this frequency shift persisted even when accounting for ten other factors, including calling depth, call range, call type, noise level, signal-to-noise ratio, local water depth (site), airgun activity, and call spatial density. No single call type was responsible for the observed shift, but so-called "complex" calls experienced larger percentage downward shifts. By contrast, the call source level distribution remained stable over the same period. The observed frequency shift also could not be explained by migration corridor shifts, relative changes in call detectability between different frequency bands, long-term degradation in the automated airgun detector, physiological growth in the population, or behavioral responses to increasing population density (estimated via call density).

Identification of two potential whale calls in the southern Indian Ocean, and their geographic and seasonal occurrence

Since passive acoustic monitoring is widely used, unidentified acoustic signals from marine mammals are commonly reported. The signal characteristics and emission patterns are the main clues to identify the possible sources. In this study, the authors describe two previously unidentified sounds, recorded at up to five widely-spaced sites (30 × 30 degree area) in the southern Indian Ocean, in 2007 and between 2010 and 2015. The first reported signal (M-call) consists of a single tonal unit near 22 Hz and lasting about 10 s, repeated with an interval longer than 2 min. This signal is only detected in 2007. The second signal (P-call) is also a tonal unit of 10 s, repeated every 160 s, but at a frequency near 27 Hz. Its yearly number increased greatly between 2007 and 2010, and moderately since then. Based on their characteristics and seasonal patterns, this study shows that both signals are clearly distinct from any known calls of blue whale subspecies and populations dwelling in the southern Indian Ocean. However, they display similarities with blue whale vocalizations. More particularly, the P-call can be mistaken for the first tonal unit of the Antarctic blue whale Z-call.

Fin whale song variability in southern California and the Gulf of California

Songs are distinct, patterned sounds produced by a variety of animals including baleen whales. Fin whale songs, which consist of short pulses repeated at regular interpulse intervals (IPIs), have been suggested as a tool to distinguish populations. Fin whale songs were analyzed from data collected from 2000–2012 in Southern California and from 2004–2010 in the Gulf of California using autonomous acoustic recorders. IPIs were measured for each identifiable song sequence during two random days of each month with recordings. Four distinct song types were identified: long doublet, short doublet, long triplet, and short triplet. Long and short doublets were the dominant songs in Southern California, while long and short triplets were dominant in the Gulf of California. An abrupt change in song type occurred in both areas during the monitoring period. We argue that each song type is unique to a population and these changes represent a shift in the primary population in the monitoring area. Occasional temporal and spatial song overlap indicated some exchange or visitation among populations. Fin whales appear to synchronize and gradually modify song rhythm over long time scales. A better understanding of the evolutionary and ecological importance of songs to fin whale populations is needed.

A pulsed-air model of blue whale B call vocalizations

Blue whale sound production has been thought to occur by Helmholtz resonance via air flowing from the lungs into the upper respiratory spaces. This implies that the frequency of blue whale vocalizations might be directly proportional to the size of their sound-producing organs. Here we present a sound production mechanism where the fundamental and overtone frequencies of blue whale B calls can be well modeled using a series of short-duration (<1 s) wavelets. We propose that the likely source of these wavelets are pneumatic pulses caused by opening and closing of respiratory valves during air recirculation between the lungs and laryngeal sac. This vocal production model is similar to those proposed for humpback whales, where valve open/closure and vocal fold oscillation is passively driven by airflow between the lungs and upper respiratory spaces, and implies call frequencies could be actively changed by the animal to center fundamental tones at different frequency bands during the call series.

"Spot" call: A common sound from an unidentified great whale in Australian temperate waters

Underwater passive acoustic recordings in the Southern and Indian Oceans off Australia from 2002 to 2016 have regularly captured a tonal signal of about 10 s duration at 22-28 Hz with a symmetrical bell-shaped envelope. The sound is often accompanied by short, higher frequency downsweeps and repeated at irregular intervals varying from 120 to 200 s. It is termed the "spot" call according to its appearance in spectrograms of long-time averaging. Although similar to the first part of an Antarctic blue whale Z-call, evidence suggests the call is produced by another great whale, with the source as yet not identified.

The influence of sea ice, wind speed and marine mammals on Southern Ocean ambient sound

This paper describes the natural variability of ambient sound in the Southern Ocean, an acoustically pristine marine mammal habitat. Over a 3-year period, two autonomous recorders were moored along the Greenwich meridian to collect underwater passive acoustic data. Ambient sound levels were strongly affected by the annual variation of the sea-ice cover, which decouples local wind speed and sound levels during austral winter. With increasing sea-ice concentration, area and thickness, sound levels decreased while the contribution of distant sources increased. Marine mammal sounds formed a substantial part of the overall acoustic environment, comprising calls produced by Antarctic blue whales (Balaenoptera musculus intermedia), fin whales (Balaenoptera physalus), Antarctic minke whales (Balaenoptera bonaerensis) and leopard seals (Hydrurga leptonyx). The combined sound energy of a group or population vocalizing during extended periods contributed species-specific peaks to the ambient sound spectra. The temporal and spatial variation in the contribution of marine mammals to ambient sound suggests annual patterns in migration and behaviour. The Antarctic blue and fin whale contributions were loudest in austral autumn, whereas the Antarctic minke whale contribution was loudest during austral winter and repeatedly showed a diel pattern that coincided with the diel vertical migration of zooplankton.

The influence of sea ice, wind speed and marine mammals on Southern Ocean ambient sound

This paper describes the natural variability of ambient sound in the Southern Ocean, an acoustically pristine marine mammal habitat. Over a 3-year period, two autonomous recorders were moored along the Greenwich meridian to collect underwater passive acoustic data. Ambient sound levels were strongly affected by the annual variation of the sea-ice cover, which decouples local wind speed and sound levels during austral winter. With increasing sea-ice concentration, area and thickness, sound levels decreased while the contribution of distant sources increased. Marine mammal sounds formed a substantial part of the overall acoustic environment, comprising calls produced by Antarctic blue whales (Balaenoptera musculus intermedia), fin whales (Balaenoptera physalus), Antarctic minke whales (Balaenoptera bonaerensis) and leopard seals (Hydrurga leptonyx). The combined sound energy of a group or population vocalizing during extended periods contributed species-specific peaks to the ambient sound spectra. The temporal and spatial variation in the contribution of marine mammals to ambient sound suggests annual patterns in migration and behaviour. The Antarctic blue and fin whale contributions were loudest in austral autumn, whereas the Antarctic minke whale contribution was loudest during austral winter and repeatedly showed a diel pattern that coincided with the diel vertical migration of zooplankton.

Characterizing Chilean blue whale vocalizations with DTAGs: a test of using tag accelerometers for caller identification

Vocal behavior of blue whales (Balaenoptera musculus) in the Gulf of Corcovado, Chile, was analyzed using both audio and accelerometer data from digital acoustic recording tags (DTAGs). Over the course of three austral summers (2014, 2015, 2016), seventeen tags were deployed, yielding 124 hours of data. We report the occurrence of Southeast Pacific type 2 (SEP2) calls, which exhibit peak frequencies, durations, and timing consistent with previous recordings made using towed and moored hydrophones. We also describe tonal downswept (D) calls, which have not been previously described for this population. Since being able to accurately assign vocalizations to individual whales is fundamental for studying communication and for estimating population densities from call rates, we further examine the feasibility of using high-resolution DTAG accelerometers to identify low-frequency calls produced by tagged blue whales. We cross-correlated acoustic signals with simultaneous tri-axial accelerometer readings in order to analyze the phase match as well as the amplitude of accelerometer signals associated with low-frequency calls, which provides a quantitative method of determining if a call is associated with a detectable acceleration signal. Our results suggest that vocalizations from nearby individuals are also capable of registering accelerometer signals in the tagged whale's DTAG record. We cross-correlate acceleration vectors between calls to explore the possibility of using signature acceleration patterns associated with sounds produced within the tagged whale as a new method of identifying which accelerometer-detectable calls originate from the tagged animal.

Seasonal and Diel Vocalization Patterns of Antarctic Blue Whale (Balaenoptera musculus intermedia) in the Southern Indian Ocean: A Multi-Year and Multi-Site Study

Passive acoustic monitoring is an efficient way to provide insights on the ecology of large whales. This approach allows for long-term and species-specific monitoring over large areas. In this study, we examined six years (2010 to 2015) of continuous acoustic recordings at up to seven different locations in the Central and Southern Indian Basin to assess the peak periods of presence, seasonality and migration movements of Antarctic blue whales (Balaenoptera musculus intermedia). An automated method is used to detect the Antarctic blue whale stereotyped call, known as Z-call. Detection results are analyzed in terms of distribution, seasonal presence and diel pattern of emission at each site. Z-calls are detected year-round at each site, except for one located in the equatorial Indian Ocean, and display highly seasonal distribution. This seasonality is stable across years for every site, but varies between sites. Z-calls are mainly detected during autumn and spring at the subantarctic locations, suggesting that these sites are on the Antarctic blue whale migration routes, and mostly during winter at the subtropical sites. In addition to these seasonal trends, there is a significant diel pattern in Z-call emission, with more Z-calls in daytime than in nighttime. This diel pattern may be related to the blue whale feeding ecology.

A study of vocal nonlinearities in humpback whale songs: from production mechanisms to acoustic analysis

Although mammalian vocalizations are predominantly harmonically structured, they can exhibit an acoustic complexity with nonlinear vocal sounds, including deterministic chaos and frequency jumps. Such sounds are normative events in mammalian vocalizations, and can be directly traceable to the nonlinear nature of vocal-fold dynamics underlying typical mammalian sound production. In this study, we give qualitative descriptions and quantitative analyses of nonlinearities in the song repertoire of humpback whales from the Ste Marie channel (Madagascar) to provide more insight into the potential communication functions and underlying production mechanisms of these features. A low-dimensional biomechanical modeling of the whale’s U-fold (vocal folds homolog) is used to relate specific vocal mechanisms to nonlinear vocal features. Recordings of living humpback whales were searched for occurrences of vocal nonlinearities (instabilities). Temporal distributions of nonlinearities were assessed within sound units, and between different songs. The anatomical production sources of vocal nonlinearities and the communication context of their occurrences in recordings are discussed. Our results show that vocal nonlinearities may be a communication strategy that conveys information about the whale’s body size and physical fitness, and thus may be an important component of humpback whale songs.

Observations and Bayesian location methodology of transient acoustic signals (likely blue whales) in the Indian Ocean, using a hydrophone triplet

A notable sequence of calls was encountered, spanning several days in January 2003, in the central part of the Indian Ocean on a hydrophone triplet recording acoustic data at a 250 Hz sampling rate. This paper presents signal processing methods applied to the waveform data to detect, group, extract amplitude and bearing estimates for the recorded signals. An approximate location for the source of the sequence of calls is inferred from extracting the features from the waveform. As the source approaches the hydrophone triplet, the source level (SL) of the calls is estimated at 187 ± 6 dB re: 1 μPa-1 m in the 15-60 Hz frequency range. The calls are attributed to a subgroup of blue whales, Balaenoptera musculus, with a characteristic acoustic signature. A Bayesian location method using probabilistic models for bearing and amplitude is demonstrated on the calls sequence. The method is applied to the case of detection at a single triad of hydrophones and results in a probability distribution map for the origin of the calls. It can be extended to detections at multiple triads and because of the Bayesian formulation, additional modeling complexity can be built-in as needed.

Are Masking-Based Models of Risk Useful?

As our understanding of directly observable effects from anthropogenic sound exposure has improved, concern about "unobservable" effects such as stress and masking have received greater attention. Equal energy models of masking such as power spectrum models have the appeal of simplicity, but do they offer biologically realistic assessments of the risk of masking? Data relevant to masking such as critical ratios, critical bandwidths, temporal resolution, and directional resolution along with what is known about general mammalian antimasking mechanisms all argue for a much more complicated view of masking when making decisions about the risk of masking inherent in a given anthropogenic sound exposure scenario.

Is low frequency ocean sound increasing globally?

Low frequency sound has increased in the Northeast Pacific Ocean over the past 60 yr [Ross (1993) Acoust. Bull. 18, 5-8; (2005) IEEE J. Ocean. Eng. 30, 257-261; Andrew, Howe, Mercer, and Dzieciuch (2002) J. Acoust. Soc. Am. 129, 642-651; McDonald, Hildebrand, and Wiggins (2006) J. Acoust. Soc. Am. 120, 711-717; Chapman and Price (2011) J. Acoust. Soc. Am. 129, EL161-EL165] and in the Indian Ocean over the past decade, [Miksis-Olds, Bradley, and Niu (2013) J. Acoust. Soc. Am. 134, 3464-3475]. More recently, Andrew, Howe, and Mercer's [(2011) J. Acoust. Soc. Am. 129, 642-651] observations in the Northeast Pacific show a level or slightly decreasing trend in low frequency noise. It remains unclear what the low frequency trends are in other regions of the world. In this work, data from the Comprehensive Nuclear-Test Ban Treaty Organization International Monitoring System was used to examine the rate and magnitude of change in low frequency sound (5-115 Hz) over the past decade in the South Atlantic and Equatorial Pacific Oceans. The dominant source observed in the South Atlantic was seismic air gun signals, while shipping and biologic sources contributed more to the acoustic environment at the Equatorial Pacific location. Sound levels over the past 5-6 yr in the Equatorial Pacific have decreased. Decreases were also observed in the ambient sound floor in the South Atlantic Ocean. Based on these observations, it does not appear that low frequency sound levels are increasing globally.

Automated detection of Antarctic blue whale calls

This paper addresses the problem of automated detection of Z-calls emitted by Antarctic blue whales (B. m. intermedia). The proposed solution is based on a subspace detector of sigmoidal-frequency signals with unknown time-varying amplitude. This detection strategy takes into account frequency variations of blue whale calls as well as the presence of other transient sounds that can interfere with Z-calls (such as airguns or other whale calls). The proposed method has been tested on more than 105 h of acoustic data containing about 2200 Z-calls (as found by an experienced human operator). This method is shown to have a correct-detection rate of up to more than 15% better than the extensible bioacoustic tool package, a spectrogram-based correlation detector commonly used to study blue whales. Because the proposed method relies on subspace detection, it does not suffer from some drawbacks of correlation-based detectors. In particular, it does not require the choice of an a priori fixed and subjective template. The analytic expression of the detection performance is also derived, which provides crucial information for higher level analyses such as animal density estimation from acoustic data. Finally, the detection threshold automatically adapts to the soundscape in order not to violate a user-specified false alarm rate.