The devil is in the detail: Quantifying vocal variation in a complex, multi-levelled, and rapidly evolving display

Development of a machine learning detector for North Atlantic humpback whale song

The study of humpback whale song using passive acoustic monitoring devices requires bioacousticians to manually review hours of audio recordings to annotate the signals. To vastly reduce the time of manual annotation through automation, a machine learning model was developed. Convolutional neural networks have made major advances in the previous decade, leading to a wide range of applications, including the detection of frequency modulated vocalizations by cetaceans. A large dataset of over 60 000 audio segments of 4 s length is collected from the North Atlantic and used to fine-tune an existing model for humpback whale song detection in the North Pacific (see Allen, Harvey, Harrell, Jansen, Merkens, Wall, Cattiau, and Oleson (2021). Front. Mar. Sci. 8, 607321). Furthermore, different data augmentation techniques (time-shift, noise augmentation, and masking) are used to artificially increase the variability within the training set. Retraining and augmentation yield F-score values of 0.88 on context window basis and 0.89 on hourly basis with false positive rates of 0.05 on context window basis and 0.01 on hourly basis. If necessary, usage and retraining of the existing model is made convenient by a framework (AcoDet, acoustic detector) built during this project. Combining the tools provided by this framework could save researchers hours of manual annotation time and, thus, accelerate their research.

Variability in humpback whale songs reveals how individuals can be distinctive when sharing a complex vocal display

Individually distinctive acoustic signals in animal vocal communication are taxonomically widespread, however, the investigation of these signal types in marine mammals has focused only on a few species. Humpback whale songs are a stereotyped, hierarchically structured vocal display performed by males, and hence thought to be sexually selected. Within a population, whales conform to a common version of the song despite the song constantly evolving. While humpback songs have been studied extensively at the population level, individual level variation has been rarely described, with inconclusive results. Here, we quantified inter- and intra-individual variability at different levels in the song hierarchy using songs from 25 singers across two song types from the eastern Australian population song of 2002 (12 singers), and the revolutionary song introduced in 2003 (13 singers). Inter-individual variability was found heterogeneously across all hierarchical levels of the song structure. In addition, distinct and individually specific patterns of song production were consistently recorded across song levels, with clear structural differences between the two song types. These results suggest that within the constraints of song conformity, males can produce individually distinctive patterns that could function as an advertisement to females to convey individual qualities.

Intra-individual variation in the songs of humpback whales suggests they are sonically searching for conspecifics

Observations of animals' vocal actions can provide important clues about how they communicate and about how they perceive and react to changing situations. Here, analyses of consecutive songs produced by singing humpback whales recorded off the coast of Hawaii revealed that singers constantly vary the acoustic qualities of their songs within prolonged song sessions. Unlike the progressive changes in song structure that singing humpback whales make across months and years, intra-individual acoustic variations within song sessions appear to be largely stochastic. Additionally, four sequentially produced song components (or "themes") were each found to vary in unique ways. The most extensively used theme was highly variable in overall duration within and across song sessions, but varied relatively little in frequency content. In contrast, the remaining themes varied greatly in frequency content, but showed less variation in duration. Analyses of variations in the amount of time singers spent producing the four themes suggest that the mechanisms that determine when singers transition between themes may be comparable to those that control when terrestrial animals move their eyes to fixate on different positions as they examine visual scenes. The dynamic changes that individual whales make to songs within song sessions are counterproductive if songs serve mainly to provide conspecifics with indications of a singer's fitness. Instead, within-session changes to the acoustic features of songs may serve to enhance a singer's capacity to echoically detect, localize, and track conspecifics from long distances.

Cognitive control of song production by humpback whales

Singing humpback whales are highly versatile vocalizers, producing complex sequences of sounds that they vary throughout adulthood. Past analyses of humpback whale song have emphasized yearly variations in structural features of songs made collectively by singers within a population with comparatively little attention given to the ways that individual singers vary consecutive songs. As a result, many researchers describe singing by humpback whales as a process in which singers produce sequences of repeating sound patterns. Here, we show that such characterizations misrepresent the degree to which humpback whales flexibly and dynamically control the production of sounds and sound patterns within song sessions. Singers recorded off the coast of Hawaii continuously morphed units along multiple acoustic dimensions, with the degree and direction of morphing varying across parallel streams of successive units. Individual singers also produced multiple phrase variants (structurally similar, but acoustically distinctive sequences) within song sessions. The precision with which individual singers maintained some acoustic properties of phrases and morphing trajectories while flexibly changing others suggests that singing humpback whales actively select and adjust acoustic elements of their songs in real time rather than simply repeating stereotyped sound patterns within song sessions.

Humpback whale song revolutions continue to spread from the central into the eastern South Pacific

Cultural transmission of behaviour is an important aspect of many animal communities ranging from humans to birds. Male humpback whales (Megaptera novaeangliae) sing a repetitive, stereotyped, socially learnt and culturally transmitted song display that slowly evolves each year. Most males within a population sing the same, slow-evolving song type; but in the South Pacific, song 'revolutions' have led to rapid and complete replacement of one song type by another introduced from a neighbouring population. Songs spread eastwards, from eastern Australia to French Polynesia, but the easterly extent of this transmission was unknown. Here, we investigated whether song revolutions continue to spread from the central (French Polynesia) into the eastern (Ecuador) South Pacific region. Similarity analyses using three consecutive years of song data (2016-2018) revealed that song themes recorded in 2016-2018 French Polynesian song matched song themes sung in 2018 Ecuadorian song, suggesting continued easterly transmission of song to Ecuador, and vocal connectivity across the entire South Pacific Ocean basin. This study demonstrates songs first identified in western populations can be transmitted across the entire South Pacific, supporting the potential for a circumpolar Southern Hemisphere cultural transmission of song and a vocal culture rivalled in its extent only by our own.

Humpback whale song revolutions continue to spread from the central into the eastern South Pacific

Cultural transmission of behaviour is an important aspect of many animal communities ranging from humans to birds. Male humpback whales (Megaptera novaeangliae) sing a repetitive, stereotyped, socially learnt and culturally transmitted song display that slowly evolves each year. Most males within a population sing the same, slow-evolving song type; but in the South Pacific, song 'revolutions' have led to rapid and complete replacement of one song type by another introduced from a neighbouring population. Songs spread eastwards, from eastern Australia to French Polynesia, but the easterly extent of this transmission was unknown. Here, we investigated whether song revolutions continue to spread from the central (French Polynesia) into the eastern (Ecuador) South Pacific region. Similarity analyses using three consecutive years of song data (2016-2018) revealed that song themes recorded in 2016-2018 French Polynesian song matched song themes sung in 2018 Ecuadorian song, suggesting continued easterly transmission of song to Ecuador, and vocal connectivity across the entire South Pacific Ocean basin. This study demonstrates songs first identified in western populations can be transmitted across the entire South Pacific, supporting the potential for a circumpolar Southern Hemisphere cultural transmission of song and a vocal culture rivalled in its extent only by our own.

Song complexity is maintained during inter-population cultural transmission of humpback whale songs

Among animal species, the songs of male humpback whales (Megaptera novaeangliae) are a rare example of social learning between entire populations. Understanding fine-scale similarity in song patterns and structural features will better clarify how accurately songs are learned during inter-population transmission. Here, six distinct song types (2009–2015) transmitted from the east Australian to New Caledonian populations were quantitatively analysed using fine-scale song features. Results found that New Caledonian whales learned each song type with high accuracy regardless of the pattern’s complexity. However, there were rare instances of themes (stereotyped patterns of sound units) only sung by a single population. These occurred more often in progressively changing ‘evolutionary’ songs compared to rapidly changing ‘revolutionary’ songs. Our results suggest that populations do not need to reduce complexity to accurately learn song patterns. Populations may also incorporate changes and embellishments into songs in the form of themes which are suggested to be learnt as distinct segments. Maintaining complex song patterns with such accuracy suggests significant acoustic contact, supporting the hypothesis that song learning may occur on shared feeding grounds or migration routes. This study improves the understanding of inter-population mechanisms for large-scale cultural transmission in animals.

When does cultural evolution become cumulative culture? A case study of humpback whale song

Culture presents a second inheritance system by which innovations can be transmitted between generations and among individuals. Some vocal behaviours present compelling examples of cultural evolution. Where modifications accumulate over time, such a process can become cumulative cultural evolution. The existence of cumulative cultural evolution in non-human animals is controversial. When physical products of such a process do not exist, modifications may not be clearly visible over time. Here, we investigate whether the constantly evolving songs of humpback whales (Megaptera novaeangliae) are indicative of cumulative cultural evolution. Using nine years of song data recorded from the New Caledonian humpback whale population, we quantified song evolution and complexity, and formally evaluated this process in light of criteria for cumulative cultural evolution. Song accumulates changes shown by an increase in complexity, but this process is punctuated by rapid loss of song material. While such changes tentatively satisfy the core criteria for cumulative cultural evolution, this claim hinges on the assumption that novel songs are preferred by females. While parsimonious, until such time as studies can link fitness benefits (reproductive success) to individual singers, any claims that humpback whale song evolution represents a form of cumulative cultural evolution may remain open to interpretation. This article is part of a discussion meeting issue 'The emergence of collective knowledge and cumulative culture in animals, humans and machines'.

Global cultural evolutionary model of humpback whale song

Humpback whale song is an extraordinary example of vocal cultural behaviour. In northern populations, the complex songs show long-lasting traditions that slowly evolve, while in the South Pacific, periodic revolutions occur when songs are adopted from neighbouring populations and rapidly spread. In this species, vocal learning cannot be studied in the laboratory, learning is instead inferred from the songs' complexity and patterns of transmission. Here, we used individual-based cultural evolutionary simulations of the entire Southern and Northern Hemisphere humpback whale populations to formalize this process of inference. We modelled processes of song mutation and patterns of contact among populations and compared our model with patterns of song theme sharing measured in South Pacific populations. Low levels of mutation in combination with rare population interactions were sufficient to closely fit the pattern of diversity in the South Pacific, including the distinctive pattern of west-to-east revolutions. Interestingly, the same learning parameters that gave rise to revolutions in the Southern Hemisphere simulations gave rise to evolutionary patterns of cultural evolution in the Northern Hemisphere populations. Our study demonstrates how cultural evolutionary approaches can be used to make inferences about the learning processes underlying cultural transmission and how they might generate emergent population-level processes.

This article is part of the theme issue ‘Vocal learning in animals and humans’.

Contribution to unravel variability in bowhead whale songs and better understand its ecological significance

Since the first studies on bowhead whale singing behaviour, song variations have been consistently reported. However, there has been little discussion regarding variability in bowhead whale singing display and its ecological significance. Unlike the better studied humpback whales, bowhead whales do not appear to share songs at population level, but several studies have reported song sharing within clusters of animals. Over the winter season 2013-2014, in an unstudied wintering ground off Northeast Greenland, 13 song groups sharing similar hierarchical structure and units were identified. Unit types were assessed through multidimensional maps, showing well separated clusters corresponding to manually labelled units, and revealing the presence of unit subtypes. Units presented contrasting levels of variability over their acoustic parameters, suggesting that bowhead whales keep consistency in some units while using a continuum in values of frequency, duration and modulation parameters for other unit types. Those findings emphasise the need to account for variability in song analysis to better understand the behavioural ecology of this endangered species. Additionally, shifting from song toward units or phrase-based analysis, as it has been suggested for humpback whales, offers the opportunity to identify and track similarities in songs over temporal and geographical scales relevant to population monitoring.

Migratory insights from singing humpback whales recorded around central New Zealand

The migration routes of wide-ranging species can be difficult to study, particularly at sea. In the western South Pacific, migratory routes of humpback whales between breeding and feeding areas are unclear. Male humpback whales sing a population-specific song, which can be used to match singers on migration to a breeding population. To investigate migratory routes and breeding area connections, passive acoustic recorders were deployed in the central New Zealand migratory corridor (2016); recorded humpback whale song was compared to song from the closest breeding populations of East Australia and New Caledonia (2015-2017). Singing northbound whales migrated past New Zealand from June to August via the east coast of the South Island and Cook Strait. Few song detections were made along the east coast of the North Island. New Zealand song matched New Caledonia song, suggesting a migratory destination, but connectivity to East Australia could not be ruled out. Two song types were present in New Zealand, illustrating the potential for easterly song transmission from East Australia to New Caledonia in this shared migratory corridor. This study enhances our understanding of western South Pacific humpback whale breeding population connectivity, and provides novel insights into the dynamic transmission of song culture.

Cultural Transmission, Evolution, and Revolution in Vocal Displays: Insights From Bird and Whale Song

Culture, defined as shared behavior or information within a community acquired through some form of social learning from conspecifics, is now suggested to act as a second inheritance system. Cultural processes are important in a wide variety of vertebrate species. Birdsong provides a classic example of cultural processes: cultural transmission, where changes in a shared song are learned from surrounding conspecifics, and cultural evolution, where the patterns of songs change through time. This form of cultural transmission of information has features that are different in speed and form from genetic transmission. More recently, culture, vocal traditions, and an extreme form of song evolution have been documented in cetaceans. Humpback whale song “revolutions,” where the single population-wide shared song type is rapidly replaced by a new, novel song type introduced from a neighboring population, represents an extraordinary example of ocean basin-wide cultural transmission rivaled in its geographic extent only by humans. In this review, we examine the cultural evolutions and revolutions present in some birdsong and whale song, respectively. By taking a comparative approach to these cultural processes, we review the existing evidence to understand the similarities and differences for their patterns of expression and the underlying drivers, including anthropogenic influences, which may shape them. Finally, we encourage future studies to explore the role of innovation vs. production errors in song evolution, the fitness information present in song, and how human-induced changes in population sizes, trajectories, and migratory connections facilitating cultural transmission may be driving song revolutions.

Network analysis reveals underlying syntactic features in a vocally learnt mammalian display, humpback whale song

Vocal communication systems have a set of rules that govern the arrangement of acoustic signals, broadly defined as 'syntax'. However, there is a limited understanding of potentially shared or analogous rules across vocal displays in different taxa. Recent work on songbirds has investigated syntax using network-based modelling. This technique quantifies features such as connectivity (adjacent signals in a sequence) and recurring patterns. Here, we apply network-based modelling to the complex, hierarchically structured songs of humpback whales (Megaptera novaeangliae) from east Australia. Given the song's annual evolving pattern and the cultural conformity of males within a population, network modelling captured the patterns of multiple song types over 13 consecutive years. Song arrangements in each year displayed clear 'small-world' network structure, characterized by clusters of highly connected sounds. Transitions between these connected sounds further suggested a combination of both structural stability and variability. Small-world network structure within humpback songs may facilitate the characteristic and persistent vocal learning observed. Similar small-world structures and transition patterns are found in several birdsong displays, indicating common syntactic patterns among vocal learning in multiple taxa. Understanding the syntactic rules governing vocal displays in multiple, independently evolving lineages may indicate what rules or structural features are important to the evolution of complex communication, including human language.

Migratory convergence facilitates cultural transmission of humpback whale song

Cultural transmission of behaviour is important in a wide variety of vertebrate taxa from birds to humans. Vocal traditions and vocal learning provide a strong foundation for studying culture and its transmission in both humans and cetaceans. Male humpback whales (Megaptera novaeangliae) perform complex, culturally transmitted song displays that can change both evolutionarily (through accumulations of small changes) or revolutionarily (where a population rapidly adopts a novel song). The degree of coordination and conformity underlying song revolutions makes their study of particular interest. Acoustic contact on migratory routes may provide a mechanism for cultural revolutions of song, yet these areas of contact remain uncertain. Here, we compared songs recorded from the Kermadec Islands, a recently discovered migratory stopover, to multiple South Pacific wintering grounds. Similarities in song themes from the Kermadec Islands and multiple wintering locations (from New Caledonia across to the Cook Islands) suggest a location allowing cultural transmission of song eastward across the South Pacific, active song learning (hybrid songs) and the potential for cultural convergence after acoustic isolation at the wintering grounds. As with the correlations in humans between genes, communication and migration, the migration patterns of humpback whales are written into their songs.

Song variation of the South Eastern Indian Ocean pygmy blue whale population in the Perth Canyon, Western Australia

Sea noise collected over 2003 to 2017 from the Perth Canyon, Western Australia was analysed for variation in the South Eastern Indian Ocean pygmy blue whale song structure. The primary song-types were: P3, a three unit phrase (I, II and III) repeated with an inter-song interval (ISI) of 170-194 s; P2, a phrase consisting of only units II & III repeated every 84-96 s; and P1 with a phrase consisting of only unit II repeated every 45-49 s. The different ISI values were approximate multiples of each other within a season. When comparing data from each season, across seasons, the ISI value for each song increased significantly through time (all fits had p << 0.001), at 0.30 s/Year (95%CI 0.217-0.383), 0.8 s/Year (95%CI 0.655-1.025) and 1.73 s/Year (95%CI 1.264-2.196) for the P1, P2 and P3 songs respectively. The proportions of each song-type averaged at 21.5, 24.2 and 56% for P1, P2 and P3 occurrence respectively and these ratios could vary by up to ± 8% (95% CI) amongst years. On some occasions animals changed the P3 ISI to be significantly shorter (120-160 s) or longer (220-280 s). Hybrid song patterns occurred where animals combined multiple phrase types into a repeated song. In recent years whales introduced further complexity by splitting song units. This variability of song-type and proportions implies abundance measure for this whale sub population based on song detection needs to factor in trends in song variability to make data comparable between seasons. Further, such variability in song production by a sub population of pygmy blue whales raises questions as to the stability of the song types that are used to delineate populations. The high level of song variability may be driven by an increasing number of background whale callers creating 'noise' and so forcing animals to alter song in order to 'stand out' amongst the crowd.

Cultural revolutions reduce complexity in the songs of humpback whales

Much evidence for non-human culture comes from vocally learned displays, such as the vocal dialects and song displays of birds and cetaceans. While many oscine birds use song complexity to assess male fitness, the role of complexity in humpback whale (Megaptera novaeangliae) song is uncertain owing to population-wide conformity to one song pattern. Although songs change gradually each year, the eastern Australian population also completely replaces their song every few years in cultural 'revolutions'. Revolutions involve learning large amounts of novel material introduced from the Western Australian population. We examined two measures of song structure, complexity and entropy, in the eastern Australian population over 13 consecutive years. These measures aimed to identify the role of complexity and information content in the vocal learning processes of humpback whales. Complexity was quantified at two hierarchical levels: the entire sequence of individual sound 'units' and the stereotyped arrangements of units which comprise a 'theme'. Complexity increased as songs evolved over time but decreased when revolutions occurred. No correlation between complexity and entropy estimates suggests that changes to complexity may represent embellishment to the song which could allow males to stand out amidst population-wide conformity. The consistent reduction in complexity during song revolutions suggests a potential limit to the social learning capacity of novel material in humpback whales.

Culturally transmitted song exchange between humpback whales (Megaptera novaeangliae) in the southeast Atlantic and southwest Indian Ocean basins

In migratory marine species, investigating population connectivity and structure can be challenging given barriers to dispersal are less evident and multiple factors may influence individual movement patterns. Male humpback whales sing a song display that can provide insights into contemporary connectivity patterns, as there can be a cultural exchange of a single, population-wide shared song type with neighbouring populations in acoustic contact. Here, we investigated song exchange between populations located on the east and west coasts of Africa using 5 years of concurrent data (2001-2005). Songs were qualitatively and quantitatively transcribed by measuring acoustic features of all song units and then compared using both Dice's similarity index and the Levenshtein distance similarity index (LSI) to quantitatively calculate song similarity. Song similarity varied among individuals and potentially between populations depending on the year (Dice: 36-100%, LSI: 21-100%), suggesting varying levels of population connectivity and/or interchange among years. The high degree of song sharing indicated in this study further supports genetic studies that demonstrate interchange between these two populations and reinforces the emerging picture of broad-scale connectivity in Southern Hemisphere populations. Further research incorporating additional populations and years would be invaluable for better understanding of fine-scale, song interchange patterns between Southern Hemisphere male humpback whales.

Using self-organizing maps to classify humpback whale song units and quantify their similarity

Classification of vocal signals can be undertaken using a wide variety of qualitative and quantitative techniques. Using east Australian humpback whale song from 2002 to 2014, a subset of vocal signals was acoustically measured and then classified using a Self-Organizing Map (SOM). The SOM created (1) an acoustic dictionary of units representing the song's repertoire, and (2) Cartesian distance measurements among all unit types (SOM nodes). Utilizing the SOM dictionary as a guide, additional song recordings from east Australia were rapidly (manually) transcribed. To assess the similarity in song sequences, the Cartesian distance output from the SOM was applied in Levenshtein distance similarity analyses as a weighting factor to better incorporate unit similarity in the calculation (previously a qualitative process). SOMs provide a more robust and repeatable means of categorizing acoustic signals along with a clear quantitative measurement of sound type similarity based on acoustic features. This method can be utilized for a wide variety of acoustic databases especially those containing very large datasets and can be applied across the vocalization research community to help address concerns surrounding inconsistency in manual classification.

Song hybridization events during revolutionary song change provide insights into cultural transmission in humpback whales

Cultural processes occur in a wide variety of animal taxa, from insects to cetaceans. The songs of humpback whales are one of the most striking examples of the transmission of a cultural trait and social learning in any nonhuman animal. To understand how songs are learned, we investigate rare cases of song hybridization, where parts of an existing song are spliced with a new one, likely before an individual totally adopts the new song. Song unit sequences were extracted from over 9,300 phrases recorded during two song revolutions across the South Pacific Ocean, allowing fine-scale analysis of composition and sequencing. In hybrid songs the current and new songs were spliced together in two specific ways: (i) singers placed a single hybrid phrase, in which content from both songs were combined, between the two song types when transitioning from one to the other, and/or (ii) singers spliced complete themes from the revolutionary song into the current song. Sequence analysis indicated that both processes were governed by structural similarity rules. Hybrid phrases or theme substitutions occurred at points in the songs where both songs contained "similar sounds arranged in a similar pattern." Songs appear to be learned as segments (themes/phrase types), akin to birdsong and human language acquisition, and these can be combined in predictable ways if the underlying structural pattern is similar. These snapshots of song change provide insights into the mechanisms underlying song learning in humpback whales, and comparative perspectives on the evolution of human language and culture.