Repertoire and classification of non-song calls in Southeast Alaskan humpback whales (Megaptera novaeangliae)

Acoustic behavior of humpback whale calves on the feeding ground: Comparisons across age and implications for vocal development

Studying sound production at different developmental stages can provide insight into the processes involved in vocal ontogeny. Humpback whales (Megaptera novaeangliae) are a known vocal learning species, but their vocal development is poorly understood. While studies of humpback whale calves in the early stages of their lives on the breeding grounds and migration routes exist, little is known about the behavior of these immature, dependent animals by the time they reach the feeding grounds. In this study, we used data from groups of North Atlantic humpback whales in the Gulf of Maine in which all members were simultaneously carrying acoustic recording tags attached with suction cups. This allowed for assignment of likely caller identity using the relative received levels of calls across tags. We analyzed data from 3 calves and 13 adults. There were high levels of call rate variation among these individuals and the results represent preliminary descriptions of calf behavior. Our analysis suggests that, in contrast to the breeding grounds or on migration, calves are no longer acoustically cryptic by the time they reach their feeding ground. Calves and adults both produce calls in bouts, but there may be some differences in bout parameters like inter-call intervals and bout durations. Calves were able to produce most of the adult vocal repertoire but used different call types in different proportions. Finally, we found evidence of immature call types in calves, akin to protosyllables used in babbling in other mammals, including humans. Overall, the sound production of humpback whale calves on the feeding grounds appears to be already similar to that of adults, but with differences in line with ontogenetic changes observed in other vocal learning species.

Vocal complexity in the long calls of Bornean orangutans

Vocal complexity is central to many evolutionary hypotheses about animal communication. Yet, quantifying and comparing complexity remains a challenge, particularly when vocal types are highly graded. Male Bornean orangutans (Pongo pygmaeus wurmbii) produce complex and variable "long call" vocalizations comprising multiple sound types that vary within and among individuals. Previous studies described six distinct call (or pulse) types within these complex vocalizations, but none quantified their discreteness or the ability of human observers to reliably classify them. We studied the long calls of 13 individuals to: (1) evaluate and quantify the reliability of audio-visual classification by three well-trained observers, (2) distinguish among call types using supervised classification and unsupervised clustering, and (3) compare the performance of different feature sets. Using 46 acoustic features, we used machine learning (i.e., support vector machines, affinity propagation, and fuzzy c-means) to identify call types and assess their discreteness. We additionally used Uniform Manifold Approximation and Projection (UMAP) to visualize the separation of pulses using both extracted features and spectrogram representations. Supervised approaches showed low inter-observer reliability and poor classification accuracy, indicating that pulse types were not discrete. We propose an updated pulse classification approach that is highly reproducible across observers and exhibits strong classification accuracy using support vector machines. Although the low number of call types suggests long calls are fairly simple, the continuous gradation of sounds seems to greatly boost the complexity of this system. This work responds to calls for more quantitative research to define call types and quantify gradedness in animal vocal systems and highlights the need for a more comprehensive framework for studying vocal complexity vis-à-vis graded repertoires.

Interactive bioacoustic playback as a tool for detecting and exploring nonhuman intelligence: "conversing" with an Alaskan humpback whale

Here we report on a rare and opportunistic acoustic turn-taking with an adult female humpback whale, known as Twain, in Southeast Alaska. Post hoc acoustic and statistical analyses of a 20-min acoustic exchange between the broadcast of a recorded contact call, known as a 'whup/throp', with call responses by Twain revealed an intentional human-whale acoustic (and behavioral) interaction. Our results show that Twain participated both physically and acoustically in three phases of interaction (Phase 1: Engagement, Phase 2: Agitation, Phase 3: Disengagement), independently determined by blind observers reporting on surface behavior and respiratory activity of the interacting whale. A close examination of both changes to the latency between Twain's calls and the temporal matching to the latency of the exemplar across phases indicated that Twain was actively engaged in the exchange during Phase 1 (Engagement), less so during Phase 2 (Agitation), and disengaged during Phase 3 (Disengagement). These results, while preliminary, point to several key considerations for effective playback design, namely the importance of salient, dynamic and adaptive playbacks, that should be utilized in experimentation with whales and other interactive nonhuman species.

Early life experience and sex influence acoustic repertoire use in wild-born, but hand-reared, captive cheetahs (Acinonyx jubatus)

Early deprivation of adult influence is known to have long-lasting effects on social abilities, notably communication skills, as adults play a key role in guiding and regulating the behavior of youngsters, including acoustic repertoire use in species in which vocal production is not learned. Cheetahs grow up alongside their mother for 18 months, thus maternal influences on the development of social skills are likely to be crucial. Here, we investigated the impact of early maternal deprivation on vocal production and use in 12 wild-born cheetahs, rescued and subsequently hand-reared either at an early (less than 2 months) or a later stage of development. We could distinguish 16 sound types, produced mostly singly but sometimes in repeated or multitype sound sequences. The repertoire of these cheetahs did not differ fundamentally from that described in other studies on adult cheetahs, but statistical analyses revealed a concurrent effect of both early experience and sex on repertoire use. More specifically, early-reared males were characterized by a high proportion of Purr, Meow, and Stutter; early-reared females Mew, Growl, Hoot, Sneeze, and Hiss; late-reared males Meow, Mew, Growl, and Howl; and late-reared females mostly Meow. Our study demonstrates therefore the long-term effects of maternal deprivation on communication skills in a limited-vocal learner and its differential effect according to sex, in line with known social differences and potential differential maternal investment. More generally, it emphasizes the critical importance to consider the past history of the subjects (e.g., captive/wild-born, mother/hand-reared, early/late-mother-deprived, etc.) when studying social behavior, notably acoustic communication.

Social calls in humpback whale mother-calf groups off Sainte Marie breeding ground (Madagascar, Indian Ocean)

Humpback whales (Megaptera novaeangliae) use vocalizations during diverse social interactions or activities such as foraging or mating. Unlike songs produced only by males, social calls are produced by all types of individuals (adult males and females, juveniles and calves). Several studies have described social calls in the humpback whale's breeding and the feeding grounds and from different geographic areas. We aimed to investigate for the first time the vocal repertoire of humpback whale mother-calf groups during the breeding season off Sainte Marie island, Madagascar, South Western Indian Ocean using data collected in 2013, 2014, 2016, and 2017. We recorded social calls using Acousonde tags deployed on the mother or the calf in mother-calf groups. A total of 21 deployments were analyzed. We visually and aurally identified 30 social call types and classified them into five categories: low, medium, high-frequency sounds, amplitude-modulated sounds, and pulsed sounds. The aural-visual classifications have been validated using random forest (RF) analyses. Low-frequency sounds constituted 46% of all social calls, mid-frequency 35%, and high frequency 10%. Amplitude-modulated sounds constituted 8% of all vocalizations, and pulsed sounds constituted 1%. While some social call types seemed specific to our study area, others presented similarities with social calls described in other geographic areas, on breeding and foraging grounds, and during migrating routes. Among the call types described in this study, nine call types were also found in humpback whale songs recorded in the same region. The 30 call types highlight the diversity of the social calls recorded in mother-calf groups and thus the importance of acoustic interactions in the relationships between the mother and her calf and between the mother-calf pair and escorts.

Detection of humpback whale (Megaptera novaeangliae) non-song vocalizations around the Vema Seamount, southeast Atlantic Ocean

Humpback whales are a cosmopolitan, highly vocal species. Investigated here are their vocalizations recorded at the Vema Seamount (31°38'S, 08°20'E) from moored hydrophones in the austral spring of 2019. During the 11-d recording period over 600 non-song calls were detected. Calls were predominantly detected at night over three consecutive days. The most common calls were low, frequency-modulated sounds (whups). An impulsive sound (gunshot) previously unknown in humpback whales was also detected. The location and timing of the calls suggests that humpback whales may be using the Vema Seamount as a temporary stop on their migration to their polar feeding grounds.

Cryptic vocal behavior of foraging humpback whales on feeding grounds in West Greenland

Male humpback whales (Megaptera novaeangliae) sing in mating aggregations in the form of song displays, but much less is known about how both sexes use sound on their feeding grounds. Here, we test different hypotheses about the function of vocalizations in 14 foraging humpback whales tagged with sound and movement recording Dtags in Greenland. We show that this population of foraging humpback whales have an overall low call rate of 11.9 calls h^-1 (inter-quartile range = 12.1) with no support for the hypotheses that they employ sound in the localization or manipulation of prey nor in the coordination of lunge feeding. The calls had a mean received level of 135 ± 5dB re 1 μPa, which is some 30 dB lower than maximum levels of song recorded on similar deployed tags, suggesting a much smaller active space of these vocalizations. This reduced active space might, in concert with low call rates, serve to mitigate eavesdropping by predatory killer whales or conspecifics competing for the same prey resources. We conclude that feeding humpback whales in Greenland produce low level, infrequent calls suggesting that calling is not a prerequisite for successful feeding, but likely serves to mediate within group social interactions.

Allopatric humpback whales of differing generations share call types between foraging and wintering grounds

Humpback whales (Megaptera novaeangliae) are a cosmopolitan baleen whale species with geographically isolated lineages. Despite last sharing an ancestor ~ 2-3 million years ago, Atlantic and Pacific foraging populations share five call types. Whether these call types are also shared between allopatric breeding and foraging populations is unclear, but would provide further evidence that some call types are ubiquitous and fixed. We investigated whether these five call types were present on a contemporary foraging ground (Newfoundland, 2015-2016) and a historic breeding ground (Hawaii, 1981-1982). Calls were classified using aural/visual (AV) characteristics; 16 relevant acoustic variables were measured and a Principal Component Analysis (PCA) was used to examine within-call and between-population variation. To assess whether between-population variation influenced classification, all 16 variables were included in classification and regression tree (CART) and random forest analyses (RF). All five call types were identified in both populations. Between-population variation in combined acoustic variables (PC1, PC2, PC3) was lower within call types than among call types, and high agreement between AV and quantitative classification (CART: 83% agreement; RF: 77% agreement) suggested that acoustic characteristics were more similar within than among call types. Findings indicate that these five call types are shared across allopatric populations, generations, and behavioural contexts.

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.

Fuzzy clustering as a tool to differentiate between discrete and graded call types

Animals may communicate potential information to conspecifics using stereotyped "discrete" calls and variable "graded" calls. However, animal vocal research often centers on identifying the number of call types in a repertoire rather than quantifying the amount of gradation. Here, fuzzy clustering was applied to the social call repertoire of a species with a complex communication system, the humpback whale (Megaptera novaeangliae). Of 26 call types, 6 were classified as discrete, 7 as graded, and 13 as intermediate. These results indicate that humpback whales have a graded call repertoire, and fuzzy clustering may be a useful approach to investigate this variability.

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.

Seasonal trends and primary contributors to the low-frequency soundscape of the Cordell Bank National Marine Sanctuary

Passive acoustic monitoring of ocean soundscapes can provide information on ecosystem status for those tasked with protecting marine resources. In 2015, the National Oceanic and Atmospheric Administration (NOAA) established a long-term, continuous, low-frequency (10 Hz-2 kHz) passive acoustic monitoring site in the Cordell Bank National Marine Sanctuary (CBNMS), located offshore of the central United States of America (U.S.) west coast, near San Francisco, CA. The California Current flows southward along the coast in this area, supporting a diverse community of marine animals, including several baleen whale species. Acoustic data analysis revealed that both large vessels and vocalizing baleen whales contribute to the ambient soundscape of the CBNMS. Sound levels fluctuated by month with the highest levels in the fall and lowest levels in the summer. Throughout the year, very low-frequency (10-100 Hz) sound levels were most variable. Vessels and whales overlap in their contributions to ambient sound levels within this range, although vessel contributions were more omnipresent, while seasonal peaks were associated with vocalizing whales. This characterization of low-frequency ambient sound levels in the CBNMS establishes initial baselines for an important component of this site's underwater soundscape. Standardized monitoring of soundscapes directly supports NOAA's ability to evaluate and report on conditions within national marine sanctuaries.

Some things never change: multi-decadal stability in humpback whale calling repertoire on Southeast Alaskan foraging grounds

Investigating long term trends in acoustic communication is essential for understanding the role of sound in social species. Humpback whales are an acoustically plastic species known for producing rapidly-evolving song and a suite of non-song vocalizations (“calls”) containing some call types that exhibit short-term stability. By comparing the earliest known acoustic recordings of humpback whales in Southeast Alaska (from the 1970’s) with recordings collected in the 1990’s, 2000’s, and 2010’s, we investigated the long-term repertoire stability of calls on Southeast Alaskan foraging grounds. Of the sixteen previously described humpback whale call types produced in Southeast Alaska, twelve were detected in both 1976 and 2012, indicating stability over a 36-year time period; eight call types were present in all four decades and every call type was present in at least three decades. We conclude that the conservation of call types at this temporal scale is indicative of multi-generational persistence and confirms that acoustic communication in humpback whales is comprised of some highly stable call elements in strong contrast to ever-changing song.

The communication space of humpback whale social sounds in wind-dominated noise

In animal social networks, a large acoustic communication space tends to involve complex networks. Signal masking may reduce this space, leading to detrimental effects on the animal's ability to obtain important social information. Humpback whales use acoustic social sounds (vocal sounds and surface-generated sounds from breaching or fin slapping) for within- and between-group communication. In this study, changes in various sound parameters (e.g., signal-above-noise and frequency content) of received humpback whale social sounds were statistically modeled against the combined effect of increasing wind-dominated noise and distance from the source (whale) to produce masking models. Behavioral data on vocalizing groups were also used to inform these models. The acoustic communication space, in this shallow water (<50 m) environment, extended to approximately 4 km from the signaler in median wind noise. However, the majority of behavioral interactions occurred within 2 km of the signaler. Surface-generated signals propagated better and likely function to maintain this space in higher wind noise. This study provides a basic wind-noise masking model for social communication signals in humpback whales which can be updated as more information on humpback auditory capabilities, and potential masking effects of anthropogenic noise sources, becomes available.

More of the same: allopatric humpback whale populations share acoustic repertoire

BACKGROUND

Humpback whales (Megaptera novaeangliae) are a widespread, vocal baleen whale best known for producing song, a complex, repetitive, geographically distinct acoustic signal sung by males, predominantly in a breeding context. Humpback whales worldwide also produce non-song vocalizations ("calls") throughout their migratory range, some of which are stable across generations.

METHODS

We looked for evidence that temporally stable call types are shared by two allopatric humpback whale populations while on their northern hemisphere foraging grounds in order to test the hypothesis that some calls, in strong contrast to song, are innate within the humpback whale acoustic repertoire.

RESULTS

Despite being geographically and genetically distinct populations, humpback whales in Southeast Alaska (North Pacific Ocean) share at least five call types with conspecifics in Massachusetts Bay (North Atlantic Ocean).

DISCUSSION

This study is the first to identify call types shared by allopatric populations, and provides evidence that some call types may be innate.

Source levels of foraging humpback whale calls

Humpback whales produce a wide range of low- to mid-frequency vocalizations throughout their migratory range. Non-song "calls" dominate this species' vocal repertoire while on high-latitude foraging grounds. The source levels of 426 humpback whale calls in four vocal classes were estimated using a four-element planar array deployed in Glacier Bay National Park and Preserve, Southeast Alaska. There was no significant difference in source levels between humpback whale vocal classes. The mean call source level was 137 dB_RMS re 1 μPa @ 1 m in the bandwidth of the call (range 113-157 dB_RMS re 1 μPa @ 1 m), where bandwidth is defined as the frequency range from the lowest to the highest frequency component of the call. These values represent a robust estimate of humpback whale source levels on foraging grounds and should append earlier estimates.

Potential motivational information encoded within humpback whale non-song vocal sounds

Acoustic signals in terrestrial animals follow motivational-structural rules to inform receivers of the signaler's motivational state, valence and level of arousal. Low-frequency "harsh" signals are produced in aggressive contexts, whereas high-frequency tonal sounds are produced in fearful/appeasement contexts. Using the non-song social call catalogue of humpback whales (Megaptera novaeangliae), this study tested for potential motivational-structural rules within the call catalogue of a baleen whale species. A total of 32 groups within different social contexts (ranging from stable, low arousal groups, such as a female with her calf, to affiliating, higher arousal, groups containing multiple males competing for access to the central female) were visually and acoustically tracked as they migrated southwards along the eastern coast of Australia. Social calls separated into four main cluster types, with signal structures in two categories consistent with "aggressive" signals and, "fearful/appeasement" signals in terrestrial animals. The group's use of signals within these clusters matched their context in that presumed low arousal non-affiliating groups almost exclusively used "low-arousal" signals (a cluster of low frequency unmodulated or upsweep sounds). Affiliating groups used a higher proportion of an intermediate cluster of signal types deemed "higher arousal" signals and groups containing three or more adults used a higher proportion of "aggressive" signal types.

Determining forward speed from accelerometer jiggle in aquatic environments

How fast animals move is critical to understanding their energetic requirements, locomotor capacity, and foraging performance, yet current methods for measuring speed via animal-attached devices are not universally applicable. Here we present and evaluate a new method that relates forward speed to the stochastic motion of biologging devices since tag jiggle, the amplitude of the tag vibrations as measured by high sample rate accelerometers, increases exponentially with increasing speed. We successfully tested this method in a flow tank using two types of biologging devices and tested the method in situ on wild cetaceans spanning ∼3 to &gt;20 m in length using two types of suction cup-attached and two types of dart-attached tag. This technique provides some advantages over other approaches for determining speed as it is device-orientation independent and relies only on a pressure sensor and a high sample rate accelerometer, sensors that are nearly universal across biologging device types.