Echolocation Clicks Research Articles

Modeling the acoustic repertoire of Cuvier's beaked whale clicks.

This paper investigates the evolution of spectral properties observed in Cuvier's beaked whale (Ziphius cavirostris) click trains recorded by fixed hydrophones in the Gulf of Mexico. In the context of deep water and high-frequency sounds and observed inter-click intervals, the authors assumed that the main effect responsible for the modification of the spectral content between adjacent clicks in the same click train is the source beam pattern. The spectral structure is studied by using the Wigner-Ville time-frequency distribution and is compared with the conventional Fourier spectrogram. The results show that the observed Cuvier's beaked whale clicks are a superposition of upsweep and downsweep chirps, unlike the currently accepted upsweep only structure of beaked whale clicks in bioacoustics literature. The spectral structure variations simulated by using a flat circular piston model as a beam pattern transmission model are consistent with the evolution of spectral click properties observed in experimental data. A better understanding of the properties of observed echolocation clicks of Cuvier's beaked whales will provide useful information for click annotations and, therefore, will contribute to improving accuracy of detecting, classifying, tracking, and estimating the density of Cuvier's beaked whales.

Monitoring of a Nearshore Small Dolphin Species Using Passive Acoustic Platforms and Supervised Machine Learning Techniques

Passive acoustic monitoring (PAM) is increasingly being adopted as a non-invasive method for the assessment of ocean ecological dynamics. PAM is an important sampling approach for acquiring critical information about marine mammals, especially in areas where data are lacking and where evaluations of threats for vulnerable populations are required. The Indo-Pacific humpback dolphin (IPHD, Sousa chinensis) is a coastal species which inhabits tropical and warm-temperate waters from the eastern Indian Ocean throughout Southeast Asia to central China. A new population of this species was recently discovered in waters southwest of Hainan Island, China. An array of passive acoustic platforms was deployed at depths of 10-20 m (the preferred habitat of humpback dolphins), across sites covering more than 100 km of coastline. In this study, we explored whether the acoustic data recorded by the array could be used to classify IPHD echolocation clicks, with the aim of investigating the spatiotemporal patterns of distribution and acoustic behavior of this species. A number of supervised machine learning algorithms were trained to automatically classify echolocation clicks from the different types of short-broadband pulses recorded. The best performance was reported by a cubic support vector machine (Cubic SVM), which was applied to 19,215 5-min recordings (similar to 4.2 TB), collected over a period of 75 days at six locations. Subsequently, using spectrogram visualization and audio listening, human operators confirmed the presence of clicks within the selected files. Additionally, other dolphin vocalizations (including whistles, buzzes, and burst pulses) and different sound sources (soniferous fishes, snapping shrimps, human activities) were also reported. The detection range of IPHD clicks was estimated using a transmission loss (IL) model and the performance of the trained classifier was compared with data synchronously collected by an acoustic data logger (A-tag). This study demonstrates that the distribution and habitat use of a coastal and resident dolphin species can be monitored over a large spatiotemporal scale, using an array of passive acoustic platforms and a data analysis protocol that includes both machine learning techniques and spectrogram inspection.

Echolocation Clicks of Irrawaddy Dolphins (Orcaella brevirostris) During Foraging in the Bay of Brunei, Malaysia

The source parameters of Irrawaddy dolphins’ echolocation click in the Bay of Brunei were estimated. Analysis of eight parameters shows that the Irrawaddy dolphins produce broadband echolocation clicks with mean click duration of 21.1 ± 7.2 μs. The clicks had a mean peak-to-peak apparent source level (ASLpp) of 201.2 ± 7.3 dB re 1 μPa (N = 350), a mean peak frequency of 116.5 ± 15.2 kHz, a mean centroid frequency of 115.4 ± 13.9 kHz, − 3 dB bandwidth of 51.8 ± 17.7 kHz, and − 10 dB bandwidth of 100.9 ± 20.3 kHz. Foraging behaviour was characterized by the high movement of the animals in various directions with no obvious pattern and frequent deep dives. The source parameters of the Brunei Bay Irrawaddy dolphins’ clicks from the present study were compared to those of the populations in Bangladesh and Thailand. The apparent source level and frequency range of clicks for the population in Brunei Bay were wider than those of the population in Bangladesh and Thailand. The variations in the measured parameters might be due to environmental factors or behaviour related.

Description and classification of echolocation clicks of Indian Ocean humpback (Sousa plumbea) and Indo-Pacific bottlenose (Tursiops aduncus) dolphins from Menai Bay, Zanzibar, East Africa.

Passive acoustic monitoring (PAM) is a powerful method to study the occurrence, movement and behavior of echolocating odontocetes (toothed whales) in the wild. However, in areas occupied by more than one species, echolocation clicks need to be classified into species. The present study investigated whether the echolocation clicks produced by small, at-risk, resident sympatric populations of Indian Ocean humpback dolphin (Sousa plumbea) and Indo-Pacific bottlenose dolphin (Tursiops aduncus) in Menai Bay, Zanzibar, East Africa, could be classified to allow species specific monitoring. Underwater sounds of S. plumbea and T. aduncus groups were recorded using a SoundTrap 202HF in January and June-August 2015. Eight acoustic parameters, i.e. -10 dB duration, peak, centroid, lower -3 and lower -10 dB frequencies, and -3 dB, -10 dB and root-mean-squared bandwidth, were used to describe and compare the two species’ echolocation clicks. Statistical analyses showed that S. plumbea clicks had significantly higher peak, centroid, lower -3 and lower -10 dB frequencies compared to T. aduncus, whereas duration and bandwidth parameters were similar for the two species. Random Forest (RF) classifiers were applied to determine parameters that could be used to classify the two species from echolocation clicks and achieved 28.6% and 90.2% correct species classification rates for S. plumbea and T. aduncus, respectively. Both species were classified at a higher rate than expected at random, however the identified classifiers would only be useful for T. aduncus monitoring. The frequency and bandwidth parameters provided most power for species classification. Further study is necessary to identify useful classifiers for S. plumbea. This study represents a first step in acoustic description and classification of S. plumbea and T. aduncus in the western Indian Ocean region, with potential application for future acoustic monitoring of species-specific temporal and spatial occurrence in these sympatric species.

Harbor Porpoise (Phocoena phocoena) Reaction to a 3D Seismic Airgun Survey in the North Sea

The most common cetacean in the North Sea is the harbour porpoise (Phocoena phocoena). Underwater noise is increasingly recognised as a source of impact on the marine environment, and 3D seismic and airguns were are one some of the first man-made high intensity sound sources to receive attention with respect to potential impact on marine mammals. In this study, we investigate the effects of a 3D seismic survey on harbour porpoise echolocation activity in the Danish sector of the North Sea. This was achieved by deploying porpoise click detectors (C-PODs) and sound recorders (SM2M and SM3M) both inside and adjacent to the seismic survey area, before, during and after the survey over a (total study time wasduration of nine months in total). Three echolocation parameters were analysed: number of clicks per minute, minutes with porpoise echolocation click trains and feeding buzz frequency in relation to all minutes with click trains. Decreases in echolocation signals were detected up to 8-12 km from the active airguns, which may indicate temporary displacement of porpoises or a change in porpoise echolocation behaviour. No general displacement of harbour porpoises away from the seismic survey area could be detected when comparing to reference stations 15 km away from any seismic activity. However, nNo general displacement of harbour porpoises away from the seismic survey area could be detected when comparing to reference stations 15 km away from any seismic activity. decreases in recorded echolocation signals were detected up to 8-12 km from the active airguns, which indicate temporary displacement of porpoises and/or a change in porpoise echolocation behaviour. Our results add to the understanding that underwater noise has the potential to affect temporarily foraging efficiency in porpoises. While the effect of seismic surveys on harbour porpoise behaviour was smaller than what has been found for pile-driving, the cumulative effect of the different anthropogenic impacts should could be performed assessed by evaluation of potential population level consequences.

DetEdit: A graphical user interface for annotating and editing events detected in long-term acoustic monitoring data.

Passive acoustic monitoring has become an important data collection method, yielding massive datasets replete with biological, environmental and anthropogenic information. Automated signal detectors and classifiers are needed to identify events within these datasets, such as the presence of species-specific sounds or anthropogenic noise. These automated methods, however, are rarely a complete substitute for expert analyst review. The ability to visualize and annotate acoustic events efficiently can enhance scientific insights from large, previously intractable datasets. A MATLAB-based graphical user interface, called DetEdit, was developed to accelerate the editing and annotating of automated detections from extensive acoustic datasets. This tool is highly-configurable and multipurpose, with uses ranging from annotation and classification of individual signals or signal-clusters and evaluation of signal properties, to identification of false detections and false positive rate estimation. DetEdit allows users to step through acoustic events, displaying a range of signal features, including time series of received levels, long-term spectral averages, time intervals between detections, and scatter plots of peak frequency, RMS, and peak-to-peak received levels. Additionally, it displays either individual, or averaged sound pressure waveforms, and power spectra within each acoustic event. These views simultaneously provide analysts with signal-level detail and encounter-level context. DetEdit creates datasets of signal labels for further analyses, such as training classifiers and quantifying occurrence, abundances, or trends. Although designed for evaluating underwater-recorded odontocete echolocation click detections, DetEdit can be adapted to almost any stereotyped impulsive signal. Our software package complements available tools for the bioacoustic community and is provided open source at https://github.com/MarineBioAcousticsRC/DetEdit.

Lateralized sound production in the beluga whale (Delphinapterus leucas).

Like other toothed whales, belugas produce sound through pneumatic actuation of two phonic lip pairs, but it is unclear whether both pairs are actuated concurrently to generate a single sound (the dual actuation hypothesis) or laterally in the production of their rich vocal repertoires. Here, using suction cup hydrophones on the head of a trained beluga whale, we measured seven different communication signal types and echolocation clicks in order to test the hypothesis that belugas produce distinct sounds unilaterally. We show that, like other delphinoids, belugas produce echolocation clicks with the right phonic lips and tonal sounds from the left. We also demonstrate for the first time that the left phonic lips are responsible for generating communication signals other than tonal sounds. Thus, our findings provide empirical support for functionalized laterality in delphinoid sound production, in keeping with the functional laterality hypothesis of vocal-motor control in toothed whales.

Classification of odontocete echolocation clicks using convolutional neural network

A method based on a convolutional neural network for the automatic classification of odontocete echolocation clicks is presented. The proposed convolutional neural network comprises six layers: three one-dimensional convolutional layers, two fully connected layers, and a softmax classification layer. Rectified linear units were chosen as the activation function for each convolutional layer. The input to the first convolutional layer is the raw time signal of an echolocation click. Species prediction was performed for groups of m clicks, and two strategies for species label prediction were explored: the majority vote and maximum posterior. Two datasets were used to evaluate the classification performance of the proposed algorithm. Experiments showed that the convolutional neural network can model odontocete species from the raw time signal of echolocation clicks. With the increase in m, the classification accuracy of the proposed method improved. The proposed method can be employed in passive acoustic monitoring to classify different delphinid species and facilitate future studies on odontocetes.

Sounds associated with foraging and prey capture in individual fish-eating killer whales, Orcinus orca.

Foraging behavior in odontocetes is fundamentally tied to the use of sound. Resident-type killer whales use echolocation to locate and capture elusive salmonid prey. In this investigation, acoustic recording tags were suction cup-attached to endangered Southern Resident killer whales to describe their acoustic behavior during different phases of foraging that, along with detections of prey handling sounds (e.g., crunches) and observed predation events, allow confirmation of prey capture. Echolocation click trains were categorized based on the inter-click interval (ICI) according to hypothesized foraging function. Whales produced slow click trains (ICI >100 ms) at shallowest depths but over the largest change of depth, fast click trains (10 ms < ICI ≤100 ms) at intermediate depths, and buzz trains (ICI ≤10 ms) at deepest depths over the smallest depth change. These results align with hypotheses regarding biosonar use to search, pursue and capture prey. Males exhibited a higher probability of producing slow click trains, buzzes and prey handling sounds, indicating higher levels of prey searching and capture to support the energy requirement of their larger body size. These findings identify relevant acoustic indicators of subsurface foraging behaviors of killer whales, enabling investigations of human impacts on sound use and foraging.

Behavioural laterality in foraging bottlenose dolphins (Tursiops truncatus).

Lateralized behaviour is found in humans and a wide variety of other species. At a population level, lateralization of behaviour suggests hemispheric specialization may underlie this behaviour. As in other cetaceans, dolphins exhibit a strong right-side bias in foraging behaviour. Common bottlenose dolphins in The Bahamas use a foraging technique termed ‘crater feeding’, in which they swim slowly along the ocean floor, scanning the substrate using echolocation, and then bury their rostrums into the sand to obtain prey. The bottlenose dolphins off Bimini, The Bahamas, frequently execute a sharp turn before burying their rostrums in the sand. Based on data collected from 2012 to 2018, we report a significant right-side (left turn) bias in these dolphins. Out of 709 turns recorded from at least 27 different individuals, 99.44% (n = 705) were to the left (right side and right eye down) [z = 3.275, p = 0.001]. Only one individual turned right (left side and left eye down, 4/4 turns). We hypothesize that this right-side bias may be due in part to the possible laterization of echolocation production mechanisms, the dolphins' use of the right set of phonic lips to produce echolocation clicks, and a right eye (left hemisphere) advantage in visual discrimination and visuospatial processing.

Deep-diving pilot whales make cheap, but powerful, echolocation clicks with 50 \xb5L of air

Echolocating toothed whales produce powerful clicks pneumatically to detect prey in the deep sea where this long-range sensory channel makes them formidable top predators. However, air supplies for sound production compress with depth following Boyle’s law suggesting that deep-diving whales must use very small air volumes per echolocation click to facilitate continuous sensory flow in foraging dives. Here we test this hypothesis by analysing click-induced acoustic resonances in the nasal air sacs, recorded by biologging tags. Using 27000 clicks from 102 dives of 23 tagged pilot whales (Globicephala macrorhynchus), we show that click production requires only 50 µL of air/click at 500 m depth increasing gradually to 100 µL at 1000 m. With such small air volumes, the metabolic cost of sound production is on the order of 40 J per dive which is a negligible fraction of the field metabolic rate. Nonetheless, whales must make frequent pauses in echolocation to recycle air between nasal sacs. Thus, frugal use of air and periodic recycling of very limited air volumes enable pilot whales, and likely other toothed whales, to echolocate cheaply and almost continuously throughout foraging dives, providing them with a strong sensory advantage in diverse aquatic habitats.

Interacting effects of vessel noise and shallow river depth elevate metabolic stress in Ganges river dolphins

In riverine ‘soundscapes’, complex interactions between sound, substrate type, and depth create difficulties in assessing impacts of anthropogenic noise pollution on freshwater fauna. Underwater noise from vessels can negatively affect endangered Ganges river dolphins (Platanista gangetica), which are ‘almost blind’ and rely entirely on high-frequency echolocation clicks to sense their environment. We conducted field-based acoustic recordings and modelling to assess acoustic responses of Platanista to underwater noise exposure from vessels in the Ganga River (India), which is now being transformed into a major waterway. Dolphins showed enhanced activity during acute noise exposure and suppressed activity during chronic exposure. Increase in ambient noise levels altered dolphin acoustic responses, strongly masked echolocation clicks, and more than doubled metabolic stress. Noise impacts were further aggravated during dry-season river depth reduction. Maintaining ecological flows, downscaling of vessel traffic, and propeller modifications to reduce cavitation noise, could help mitigate noise impacts on Ganges river dolphins.

Whistle Classification of Sympatric False Killer Whale Populations in Hawaiian Waters Yields Low Accuracy Rates

Cetaceans are ecologically important marine predators, and designating individuals to distinct populations can be challenging. Passive acoustic monitoring provides an approach to classify cetaceans to populations using their vocalizations. In the Hawaiian Archipelago, three genetically distinct, sympatric false killer whale (Pseudorca crassidens) populations coexist: a broadly distributed pelagic population and two island-associated populations, an endangered main Hawaiian Islands (MHI) population and a Northwestern Hawaiian Islands (NWHI) population. The mechanisms that sustain the genetic separation between these overlapping populations are unknown but previous studies suggest that the acoustic diversity between populations may correspond to genetic differences. Here, we investigated whether false killer whale whistles could be correctly classified to population based on their characteristics to serve as a method of identifying populations when genetic or photographic-identification data are unavailable. Acoustic data were collected during line-transect surveys using towed hydrophone arrays. We measured 50 time and frequency parameters from whistles in 16 false killer whale encounters identified to population and used those measures to train and test random forest classification models. Random Forest models that included three populations correctly classified 42% of individual whistles overall and resulted in a low kappa coefficient, κ = 0.15, indicating low agreement between models and the true population. Whistles from the MHI population showed the highest correct classification rate (52%) compared to pelagic and NWHI whistles (42% and 36%, respectively). Pairwise random forest models classifying pelagic and MHI whistles proved slightly more accurate (62% accuracy, κ = 0.24), though a similar pelagic-NWHI model did not (56% accuracy, κ = 0.12). Results suggest that the time-frequency whistle characteristics are not suitable to confidently classify encounters to a specific false killer whale population, although certain features of whistles produced by the endangered MHI population allow for overall higher classification accuracy. Inclusion of other vocalization types, such as echolocation clicks, and alternative whistle variables may improve correct classification success for these sympatric populations.

Fine-scale classification of odontocete echolocation clicks using deep learning

All toothed whale species produce echolocation clicks, which can be used for passive acoustic monitoring. However in biodiverse habitats, such as off the coast of southern California, it is common for acoustic encounters with multiple species to overlap in time. This creates a challenge for acoustic classification in the context of long term monitoring. Machine learning methods may facilitate classification at the level of individual detections. A generic echolocation click detector was applied to passive acoustic recordings from the 2013 DCL dataset, which has manual labels of five odontocete species encounters available as a ground truth. We evaluated two different methods for training deep neural networks to classify the detections to species, using either individual click waveforms or clustered sets of similar echolocation clicks. Results show that deep neural networks have the potential to accurately classify clicks to species at fine scales, allowing for improved handling of complex acoustic environments and multi-species encounters. We suggest that click-level classification can facilitate more advanced quantitative analysis of passive acoustic recordings, and that it can be done efficiently using deep learning.

Developing a click type “library” for Hawaiian odontocetes using machine learning methods

Though odontocete echolocation clicks are highly useful in passive acoustic monitoring, their acoustic features vary depending on behavior, orientation, and location relative to the receiver. Using unsupervised machine learning, regional sets of echolocation click types can be identified that allow for automated labeling of bioacoustic data while accounting for within-type signal variability. This methodology derives labels for clusters of click types based on parameters including spectral shape, inter-click interval, and click duration. Here we present classification results of echolocation clicks produced by Hawaiian odontocetes. Unsupervised methods were used to establish a set of click types for 10 years of recordings (200 kHz) collected from a bottom-mounted hydrophone off the coast of Kona, HI. These automated labels were compared with manual labels for species with established and recognizable clicks, which allowed for ground-truthing of some click types identified by the unsupervised machine learning method. Towed array data (500 kHz) was used to assign click types to species with visual verification, expanding the set of known species-specific click types for the region. The click types and method established here will be useful in future analyses of Hawaiian odontocete density estimation, abundance, and distribution patterns using acoustic data without requiring visual confirmation.

Time of arrival difference estimation for narrow band high frequency echolocation clicks.

Algorithms are presented for the accurate time of arrival difference estimation of high frequency narrow band echolocation clicks from Harbor Porpoise. These clicks typically have a center frequency of around 130 kHz (wavelength ∼1.2 cm) and duration of <0.1 ms. When using hydrophones spaced centimeters apart, spatial aliasing can cause large errors on inter-hydrophone timing measurements due to the incorrect peak in the cross-correlation function of two signals being selected. It is shown that at sample rates of less than about 6 times the fundamental frequency, the incorrect correlation peak will be selected in 55% of measurements leading to large errors in time of arrival estimates. For clicks with a SNR > 10 dB these errors can be reduced by over two orders of magnitude through a combination of up-sampling the data and parabolic interpolation of peaks in the cross-correlation functions.

A paradigm to explore bistatic matching in the bottlenose dolphin (Tursiops truncatus)

Match-to-sample (MTS) has been demonstrated in dolphins using echolocation and across sensory modalities (e.g., visual to echolocation and vice-versa). To assess potential for bistatic MTS, a pilot study was conducted in which a dolphin listened to one of three recorded playbacks of echoes obtained from targets of different shape and material composition and was then asked to echo locate on a set of physical alternatives and choose the corresponding target. Training began with two alternatives and 50% errorless trials, where only the matching physical target was presented. The proportion of errorless trials was progressively reduced to 25%, 12.5%, and then 0% as a criterion of three sessions with >80% correct matching was met. Thereafter, all three alternatives were presented with 0% errorless trials. The dolphin consistently performed above chance over 34 sessions—average correct matching rate was 76% (range 67%–96%)—even though the projected sample echoes did not faithfully reproduce echoes obtained when the dolphin insonified the targets with echolocation clicks. Future testing with novel stimulus sets would be required to determine if and how a dolphin might spontaneously generalize between projected echo facsimiles and real world targets.Match-to-sample (MTS) has been demonstrated in dolphins using echolocation and across sensory modalities (e.g., visual to echolocation and vice-versa). To assess potential for bistatic MTS, a pilot study was conducted in which a dolphin listened to one of three recorded playbacks of echoes obtained from targets of different shape and material composition and was then asked to echo locate on a set of physical alternatives and choose the corresponding target. Training began with two alternatives and 50% errorless trials, where only the matching physical target was presented. The proportion of errorless trials was progressively reduced to 25%, 12.5%, and then 0% as a criterion of three sessions with >80% correct matching was met. Thereafter, all three alternatives were presented with 0% errorless trials. The dolphin consistently performed above chance over 34 sessions—average correct matching rate was 76% (range 67%–96%)—even though the projected sample echoes did not faithfully reproduce echoes obtained whe...

Identification of species-specific delphinid echolocation click types

Delphinid echolocation clicks are often recorded in large, high-quality datasets collected by moored autonomous passive acoustic sensors, but the particular species present during these acoustic encounters is generally unknown. We combined two acoustic data types with visual observations to obtain species identity for some of these clicks. We assigned species identity to acoustic encounters in eight years of recordings at five western North Atlantic mooring sites using sighting data from contemporaneous shipboard and aerial visual surveys. We also assigned species identity to acoustic encounters in towed acoustic array data collected concurrent with visual surveys during four ship-based cetacean surveys in the western North Atlantic and the Gulf of Mexico. 114 mooring encounters with seven species and 181 towed array encounters with nine species were labeled this way. We identified recurring click types during all encounters using an unsupervised clustering algorithm to groups clicks based on pairwise spectral distances. Labeled click types thus identified in the moored data were compared to labeled click types identified in the towed array data to determine their consistency and utility for delphinid species discrimination in moored passive acoustic data. Results for bottlenose dolphin, Atlantic spotted dolphin, Risso’s dolphin, and pilot whales will be presented.

Inter and intra specific variation in echolocation signals among odontocete species in the Northwest Atlantic, the Temperate Pacific and Hawaii

Odontocete species use echolocation signals (clicks) to forage and navigate. The aim of this study is to explore inter- and intra-specific variation in clicks among odontocete species in the Northwest Atlantic, Temperate Pacific, and Hawaii. Clicks were examined for seven species of odontocetes—bottlenose dolphins, common dolphins, striped dolphins, rough-toothed dolphins, pilot whales, Risso’s dolphins, and Cuvier’s beaked whales. Specially developed PAMGuard tools were used to automatically measure a suite of click parameters. Seven parameters were compared within and between species; duration, 10 dB bandwidth, 3 dB bandwidth, center frequency, peak frequency, sweep rate, and number of zero crossings. Significant differences in duration, center, and peak frequency were evident between species within study areas (Dunn’s test with Benjamini-Hochberg adjustment <0.05). Geographic variation in click parameters between the three study regions was also examined. Results showed statistically significant pair-wise differences between geographic regions for almost all echolocation click parameters and species (Dunn’s test with Benjamini-Hochberg adjustment <0.05). Results suggest species-specific differences in clicks among odontocetes and indicate that geographic variation exists for multiple species. The ecological significance of these findings will be discussed along with implications for classifier development.

Deep Machine Learning Techniques for the Detection and Classification of Sperm Whale Bioacoustics

We implemented Machine Learning (ML) techniques to advance the study of sperm whale (Physeter macrocephalus) bioacoustics. This entailed employing Convolutional Neural Networks (CNNs) to construct an echolocation click detector designed to classify spectrograms generated from sperm whale acoustic data according to the presence or absence of a click. The click detector achieved 99.5% accuracy in classifying 650 spectrograms. The successful application of CNNs to clicks reveals the potential of future studies to train CNN-based architectures to extract finer-scale details from cetacean spectrograms. Long short-term memory and gated recurrent unit recurrent neural networks were trained to perform classification tasks, including (1) “coda type classification” where we obtained 97.5% accuracy in categorizing 23 coda types from a Dominica dataset containing 8,719 codas and 93.6% accuracy in categorizing 43 coda types from an Eastern Tropical Pacific (ETP) dataset with 16,995 codas; (2) “vocal clan classification” where we obtained 95.3% accuracy for two clan classes from Dominica and 93.1% for four ETP clan types; and (3) “individual whale identification” where we obtained 99.4% accuracy using two Dominica sperm whales. These results demonstrate the feasibility of applying ML to sperm whale bioacoustics and establish the validity of constructing neural networks to learn meaningful representations of whale vocalizations.