Biphonation, deterministic chaos, sidebands and subharmonics are four non-linear phenomena (NLP) that have been identified as common additions in the phonations of animals. NLP have been hypothesised to communicate urgency, caller identification, fitness and arousal/valence states for a variety of species but have yet to be studied in detail for bottlenose dolphins. For this study, the signature whistles of nine bottlenose dolphins residing at the US Navy Marine Mammal Program (MMP) were opportunistically recorded during routine periods of separation from conspecifics. NLP were found to be common additions onto the spectral structure of signature whistles, occurring in 53% of recorded whistles (340/642). Sidebands were the most common NLP type produced. Although less frequently emitted, biphonations were characterised by a significantly longer persistence than the other NLP types. Age had a negative correlation with overall NLP presence, and more specifically, sideband presence. Individual differences in NLP use existed between dolphins; however, all dolphins were recorded producing a minimum of two NLP types. We describe NLP prevalence in dolphin whistles in order to provide a useful baseline for continued research to further identify changes in NLP across behavioural and/or health conditions.

Bottlenose dolphins often conceal behavioral signs of illness until they reach an advanced stage. Motivated by the efficacy of vocal biomarkers in human health diagnostics, we utilized supervised machine learning methods to assess various model architectures’ effectiveness in classifying dolphin health status from the acoustic features of their whistles. A gradient boosting classifier achieved a 72.3% accuracy in distinguishing between normal and abnormal health states—a significant improvement over chance (permutation test; 1000 iterations, p < 0.001). The model was trained on 30,693 whistles from 15 dolphins and the test set (15%) totaled 3612 ‘normal’ and 1775 ‘abnormal’ whistles. The classifier identified the health status of the dolphin from the whistles features with 72.3% accuracy, 73.2% recall, 56.1% precision, and a 63.5% F1 score. These findings suggest the encoding of internal health information within dolphin whistle features, with indications that the severity of illness correlates with classification accuracy, notably in its success for identifying ‘critical’ cases (94.2%). The successful development of this diagnostic tool holds promise for furnishing a passive, non-invasive, and cost-effective means for early disease detection in bottlenose dolphins.

Bottlenose dolphins (Tursiops truncatus) rely on frequency- and amplitude-modulated whistles to communicate, and noise exposure can inhibit the success of acoustic communication through masking or causing behavioral changes in the animal. At the US Navy Marine Mammal Program (MMP) in San Diego, CA, dolphins are housed in netted enclosures in the San Diego Bay and exposed to noise from vessels, unmanned underwater vehicles, and other remote sensing devices. The acoustic behavior of 20 dolphins was monitored and whistle rates during noise events were quantified. Whistle rates during the onset of the event (i.e., the first 5 min) did not significantly differ from the pre-onset (5 min immediately preceding). Whistle rates were also not significantly different for the entire duration of the event compared to a matched control period. The noise’s frequency range (i.e., control, mid-frequency (0–20 kHz) or high-frequency (21–80 kHz)), signal-to-noise ratio, and sound pressure level were not significantly related to the dolphins’ whistle rate. Considering this is a location of frequent and moderate noise output, these results lend support to established guidelines on anthropogenic noise exposure for cetaceans, suggesting that moderate noise exposure levels may not impact communication efforts in bottlenose dolphins.

Killer whales (Orcinus orca) produce a variety of acoustic signal types used for communication: clicks, whistles, and pulsed calls. Discrete pulsed calls are highly stereotyped, repetitive, and unique to individual pods found around the world. Discriminating amongst pod specific calls can help determine population structure in killer whales and is used to track pod movements around oceans. Killer whale presence in the Canadian Arctic has increased substantially, but we have limited understanding of their ecology, movements, and stock identity. Two autonomous passive acoustic monitoring (PAM) hydrophones were deployed in the waters of Eclipse Sound and Milne Inlet, in northern Baffin Island, Nunavut, Canada, in August and September 2017. Eleven killer whale pulsed call types, three multiphonic and eight monophonic, are proposed and described using manual whistle contour extraction and feature normalization. Automated detection of echolocation clicks between 20 and 48 kHz demonstrated little to no overlap between killer whale calls and echolocation presumed to be narwhal, which suggests that narwhal remain audibly inconspicuous when killer whales are present. Describing the acoustic repertoire of killer whales seasonally present in the Canadian Arctic will aid in understanding their acoustic behaviour, seasonal movements, and ecological impacts. The calls described here provide a basis for future acoustic comparisons across the North Atlantic and aid in characterizing killer whale demographics and ecology, particularly for pods making seasonal incursions into Arctic waters.

(1) Background: When a human or animal is recovering from general anesthesia, their medical team uses several behavioral and physiological parameters to assess their emergence from the unconscious state to complete wakefulness. However, the return of auditory and acoustic behaviors indicative of the complete return of consciousness in humans can be difficult to assess in a completely aquatic non-human mammal. Dolphins produce sound using the nasal system while using both passive auditory and active biological sonar (echolocation) to navigate and interrogate their environment. The sounds generated by dolphins, such as whistles and clicks, however, can be difficult to hear when the animal is submerged. (2) Methods: We implemented a system to audibly and visually (i.e., using spectrograms) monitor the underwater acoustic behavior of dolphins recovering from anesthesia. (3) Results: Eleven of the twelve recorded dolphins began echolocating within 92 min (Mean = 00:43:41 HH:MM:SS) following spontaneous respirations. In all cases, the dolphins echolocated prior to whistling (Mean = 04:57:47). The return of echolocation was significantly correlated to the return of the righting reflex (Mean = 1:13:44), a commonly used behavioral indicator of dolphin emergence. (4) Conclusions: We suggest that acoustic monitoring for the onset of click production may be a useful supplement to the established medical and behavioral biomarkers of restoring consciousness following anesthesia in bottlenose dolphins.

Bottlenose dolphins produce several types of whistle contours, including signature whistles, shared whistles, copies of conspecifics' signature whistles, and variant whistles. While signature whistles as individual identifiers are well studied, less has been demonstrated for identifying dolphins from non‐signature whistle types. Shared whistles, which are produced by multiple dolphins, and signature whistle copies are particularly difficult to classify. Variant whistles remain largely unexplored for individual identification. This study evaluated the use of a convolutional neural network operating on raw amplitude vectors to classify which dolphin produced the whistle (20 different dolphins), regardless of the contour type, from the raw time‐series of whistle recordings. When trained on all whistle types, the model achieved strong performance in classifying whistle producers (85.20%), with the highest accuracies for signature (96.10%) and shared whistles (92.50%), moderate accuracy for variant whistles (72.10%), and lowest accuracy for signature whistle copies (17.20%). To evaluate the trade‐off between classification accuracy and ecological relevance, a model run solely on signature whistles achieved 93.70% accuracy. This study represents a first step toward caller identification across whistle types for well‐studied populations, rather than solely relying on individually unique signature whistles alone. These results also demonstrate the potential for deep learning (DL) and raw audio time series to support future advances in passive acoustic monitoring of odontocetes.