Long-term Acoustic Recordings Research Articles

Unsupervised discovery of family specific vocal usage in the Mongolian gerbil.

In nature, animal vocalizations can provide crucial information about identity, including kinship and hierarchy. However, lab-based vocal behavior is typically studied during brief interactions between animals with no prior social relationship, and under environmental conditions with limited ethological relevance. Here, we address this gap by establishing long-term acoustic recordings from Mongolian gerbil families, a core social group that uses an array of sonic and ultrasonic vocalizations. Three separate gerbil families were transferred to an enlarged environment and continuous 20-day audio recordings were obtained. Using a variational autoencoder (VAE) to quantify 583,237 vocalizations, we show that gerbils exhibit a more elaborate vocal repertoire than has been previously reported and that vocal repertoire usage differs significantly by family. By performing gaussian mixture model clustering on the VAE latent space, we show that families preferentially use characteristic sets of vocal clusters and that these usage preferences remain stable over weeks. Furthermore, gerbils displayed family-specific transitions between vocal clusters. Since gerbils live naturally as extended families in complex underground burrows that are adjacent to other families, these results suggest the presence of a vocal dialect which could be exploited by animals to represent kinship. These findings position the Mongolian gerbil as a compelling animal model to study the neural basis of vocal communication and demonstrates the potential for using unsupervised machine learning with uninterrupted acoustic recordings to gain insights into naturalistic animal behavior.

Accurate species classification of Arctic toothed whale echolocation clicks using one-third octave ratios.

Passive acoustic monitoring has been an effective tool to study cetaceans in remote regions of the Arctic. Here, we advance methods to acoustically identify the only two Arctic toothed whales, the beluga (Delphinapterus leucas) and narwhal (Monodon monoceros), using echolocation clicks. Long-term acoustic recordings collected from moorings in Northwest Greenland were analyzed. Beluga and narwhal echolocation signals were distinguishable using spectrograms where beluga clicks had most energy >30 kHz and narwhal clicks had a sharp lower frequency limit near 20 kHz. Changes in one-third octave levels (TOL) between two pairs of one-third octave bands were compared from over one million click spectra. Narwhal clicks had a steep increase between the 16 and 25 kHz TOL bands that was absent in beluga click spectra. Conversely, beluga clicks had a steep increase between the 25 and 40 kHz TOL bands that was absent in narwhal click spectra. Random Forest classification models built using the 16 to 25 kHz and 25 to 40 kHz TOL ratios accurately predicted the species identity of 100% of acoustic events. Our findings support the use of echolocation TOL ratios in future automated click classifiers for acoustic monitoring of Arctic toothed whales and potentially for other odontocete species.

Observational study on the non-linear response of dolphins to the presence of vessels

With the large increase in human marine activity, our seas have become populated with vessels that can be overheard from distances of even 20 km. Prior investigations showed that such a dense presence of vessels impacts the behaviour of marine animals, and in particular dolphins. While previous explorations were based on a linear observation for changes in the features of dolphin whistles, in this work we examine non-linear responses of bottlenose dolphins (Tursiops Truncatus) to the presence of vessels. We explored the response of dolphins to vessels by continuously recording acoustic data using two long-term acoustic recorders deployed near a shipping lane and a dolphin habitat in Eilat, Israel. Using deep learning methods we detected a large number of 50,000 whistles, which were clustered to associate whistle traces and to characterize their features to discriminate vocalizations of dolphins: both structure and quantities. Using a non-linear classifier, the whistles were categorized into two classes representing the presence or absence of a nearby vessel. Although our database does not show linear observable change in the features of the whistles, we obtained true positive and true negative rates exceeding 90% accuracy on separate, left-out test sets. We argue that this success in classification serves as a statistical proof for a non-linear response of dolphins to the presence of vessels.

A real-time underwater soundscape classification system

The study of underwater soundscapes, pioneered by Dr. Jeff Nystuen, includes information about geophonic, biological, and anthropogenic processes, many of which have distinct spectral characteristics. The contributors to the underwater ambient sound field can be quantified and classified using knowledge of these spectra. Long-term acoustic data recordings from a wide variety of depths and locations with high sampling frequencies have been analyzed to develop a robust soundscape classification algorithm based on Dr. Nystuen’s methods. Each one-minute of data has been evaluated to classify the soundscape into wind, rainfall, drizzle, heavy shipping, light shipping, other vessel activity, and biological phenomena. The power spectral density (PSD) level at twelve frequencies in the range of 0.03–30 kHz, as well as the spectral slope for the frequency range between 8 and 15kHz and kurtosis are used for the passive classification algorithm. After classification, the wind speed was quantified as a cubic function of PSD at 6 kHz and recording depth. The wind speed estimated from the acoustics compared very well to satellite data for speeds lower than 15 m/s. The classification algorithms are being embedded on a processor using Xilinx’s Zynq System-on-Chip that produces a 32-kHz hybrid millidecade spectrum in real-time on a logarithmic scale.

A workflow for the automated detection and classification of female gibbon calls from long-term acoustic recordings

Passive acoustic monitoring (PAM) allows for the study of vocal animals on temporal and spatial scales difficult to achieve using only human observers. Recent improvements in recording technology, data storage, and battery capacity have led to increased use of PAM. One of the main obstacles in implementing wide-scale PAM programs is the lack of open-source programs that efficiently process terabytes of sound recordings and do not require large amounts of training data. Here we describe a workflow for detecting, classifying, and visualizing female Northern grey gibbon calls in Sabah, Malaysia. Our approach detects sound events using band-limited energy summation and does binary classification of these events (gibbon female or not) using machine learning algorithms (support vector machine and random forest). We then applied an unsupervised approach (affinity propagation clustering) to see if we could further differentiate between true and false positives or the number of gibbon females in our dataset. We used this workflow to address three questions: (1) does this automated approach provide reliable estimates of temporal patterns of gibbon calling activity; (2) can unsupervised approaches be applied as a post-processing step to improve the performance of the system; and (3) can unsupervised approaches be used to estimate how many female individuals (or clusters) there are in our study area? We found that performance plateaued with >160 clips of training data for each of our two classes. Using optimized settings, our automated approach achieved a satisfactory performance (F1 score ~ 80%). The unsupervised approach did not effectively differentiate between true and false positives or return clusters that appear to correspond to the number of females in our study area. Our results indicate that more work needs to be done before unsupervised approaches can be reliably used to estimate the number of individual animals occupying an area from PAM data. Future work applying these methods across sites and different gibbon species and comparisons to deep learning approaches will be crucial for future gibbon conservation initiatives across Southeast Asia.

Homage to Reptiles and Amphibians as Model Systems: One Ecologist's View

Reptiles and amphibians are surprisingly understudied in ecology. In this research perspective, I examine five areas of research in ecology, associated with the research areas and organisms that have attracted my interest as an ecologist for over 40 years. After discussing the forces and studies that shaped my interests and approaches I used as an ecologist, I then examine these fields, focusing on the utility of herptiles as study systems in each case. I first describe the use of reptiles and amphibians, but especially reptiles, as study organisms with which to examine the mechanisms of effects of anthropogenic habitat modifications. I then move on to describe the use of lizards in life history and performance evolution studies, including their use in studies of skin functional morphology and adhesion. After these descriptions, I briefly discuss the use of frogs in understanding disease ecology and cane toads as a system in which to study the control of invasive organisms. Finally, I discuss the use of long-term acoustic recordings of frog choruses to reveal ecological patterns. I put these areas into the framework of my own scientific trajectory from Canada to Australia and from life history studies on single species to preparing for pattern seeking in years-long, continent-wide acoustic recordings. All along, amphibians and reptiles have been outstanding model systems with which to make headway on a wide variety of ecological questions.

Quiet oceans: United States Virgin Islands coral reef soundscapes in the age of coronavirus

The COVID-19 pandemic has led to unprecedented changes in human behavior around the world. These changes have impacted local soundscapes in both terrestrial and marine environments resulting in a global ‘quieting’ event that has allowed for the opportunity to evaluate the activity of soniferous organisms in unique acoustic environments. Across the world’s oceans, patterns of commercial, research, and recreational vessel traffic have all been altered, with corresponding alterations in the anthropogenic contribution to underwater noise. To assess these changes we used long-term acoustic recorders at three different coral reef sites on the Southern coast of the island of St. John in the United States Virgin Islands (USVI). We compared conditions pre-coronavirus outbreak (2019) and post-outbreak (2020) and found a decrease in the incidence of boat noise over the course of the pandemic. Because anthropogenic noise tends to overlap with the low frequencies used by marine fishes and invertebrates for sound production and detection, we predicted that changes in boat noise patterns would lead to concurrent changes in biological sound production of coral reef organisms. Understanding the impacts of this “quiet time” on marine organismal activity in the USVI could play a role in worldwide coral reef acoustic studies and conservation efforts.

Machine learning analysis reveals relationship between pomacentrid calls and environmental cues

Sound production rates of fishes can be used as an indicator for coral reef health, providing an opportunity to utilize long-term acoustic recordings to assess environmental change. As acoustic datasets become more common, computational techniques need to be developed to facilitate analysis of the massive data files produced by long-term monitoring. Machine learning techniques demonstrate an advantage in the identification of fish sounds over manual sampling approaches. Here we evaluated the ability of convolutional neural networks to identify and monitor call patterns for pomacentrids (damselfishes) in a tropical reef region of the western Pacific. A stationary hydrophone was deployed for 39 mo (2014-2018) in the National Park of American Samoa to continuously record the local marine acoustic environment. A neural network was trained—achieving 94% identification accuracy of pomacentrids—to demonstrate the applicability of machine learning in fish acoustics and ecology. The distribution of sound production was found to vary on diel and interannual timescales. Additionally, the distribution of sound production was correlated with wind speed, water temperature, tidal amplitude, and sound pressure level. This research has broad implications for state-of-the-art acoustic analysis and promises to be an efficient, scalable asset for ecological research, environmental monitoring, and conservation planning.

Seasonal trends in underwater ambient noise near St. Lawrence Island and the Bering Strait

We measured spatial and temporal patterns of ambient noise in dynamic, relatively pristine Arctic marine habitats and evaluate the contributions of environmental and human noise sources. Long-term acoustic recorders were deployed around St. Lawrence Island and the Bering Strait region within key feeding and migratory corridors for protected species that are inherently important to Native Alaskan cultures. Over 3000 h of data from 14 recorders at nine sites were obtained from October 2014 to June 2017. Spatial and temporal ambient noise patterns were quantified with percentile statistics in 1/3rd-octave bands (0.02–8 kHz). Ice presence strongly influenced ambient noise by influencing the physical environment and presence of marine mammals. High variability in noise was observed within and between sites, largely as a function of ice presence and associated factors. Acute contributions of biological and anthropogenic sources to local ambient noise are compared to monthly averages, demonstrating how they influence Arctic soundscapes.

Characterization of the acoustic community of vocal fishes in the Azores.

Sounds produced by teleost fishes are an important component of marine soundscapes, making passive acoustic monitoring (PAM) an effective way to map the presence of vocal fishes with a minimal impact on ecosystems. Based on a literature review, we list the known soniferous fish species occurring in Azorean waters and compile their sounds. We also describe new fish sounds recorded in Azores seamounts. From the literature, we identified 20 vocal fish species present in Azores. We analysed long-term acoustic recordings carried out since 2008 in Condor and Princesa Alice seamounts and describe 20 new putative fish sound sequences. Although we propose candidates as the source of some vocalizations, this study puts into evidence the myriad of fish sounds lacking species identification. In addition to identifying new sound sequences, we provide the first marine fish sound library for Azores. Our acoustic library will allow to monitor soniferous fish species for conservation and management purposes.

Potential impact of mid-frequency active sonar on whales from passive acoustic monitoring data

Behavioral response studies of tagged whales have documented an adverse reaction to mid-frequency active sonar (MFAS). The relationship between MFAS and the acoustic behavior of whales is complex and requires accounting for natural variability in species presence. We examine the acoustic response of Cuvier’s beaked whales (Ziphius cavirostris) and blue whales (Balaenoptera musculus) from long-term acoustic recordings at three sites in Southern California waters from 2009 to 2015. Presence and absence were marked in 1-min bins for all signal types. Generalized estimating equations (GEEs) were used to model relationships between species acoustic detection and temporal and sonar covariates while accounting for residual temporal autocorrelation. Both species showed considerable temporal variability in calling. Modeling at one site showed decreased detections of beaked whale echolocation clicks when sonar signals had higher received levels at the recorders and when the proportion of bins containing sonar increased. On cessation of sonar signals, beaked whale detection increased over periods of up to about a week. Analyses of blue whale response and a comparison of multiple sites will also be presented. This work represents substantial progress in our understanding of the impact of MFAS on cetaceans, which is crucial for effective context-dependent management of anthropogenic activities.

Ghost bats exhibit informative daily and seasonal temporal patterns in the production of social vocalisations

The ghost bat (Macroderma gigas) is a colonial and highly vocal species that is impacted by human visitation of caves. The ability to document behaviours inside the roost by recording vocalisations could provide an important new tool for the management of this disturbance-prone species by removing the need for in-person confirmation of reproductive activity, and, in turn, identifying roosts of conservation importance. To assess whether vocalisations are indicators of daily and seasonal behavioural events, we aimed to determine whether total vocal activity significantly varied by time of day and time of year and, further, how the relative frequencies of occurrence of three common social vocalisations (‘Chirp-trill’, ‘Squabble’ and ‘Ultrasonic Social’) aligned with previously reported seasonal reproductive behaviour. We recorded sound inside the largest known maternity roost, extracted all vocal signals and classified them into types using semiautomated methods. Total vocal activity varied significantly by time of day and time of year, peaking around sunrise and sunset, and during the mating and nursing seasons. The relative frequencies of occurrence of vocalisation types varied significantly seasonally, with the Chirp-trill and Squabble produced most during the mating season and first flight periods, whereas the Ultrasonic Social peaked during parturition and weaning periods. This timing aligns with a previously suggested vocalisation function, providing further evidence that these signals are important in mating and maternity behaviours. Further, this suggests that peaks in the relative frequency of occurrence of distinct social vocalisations may act as indicators of in-roost reproductive and pup development behaviours and provides a low-disturbance, semiautomated method for using long-term acoustic recordings to study and monitor behaviour in this sensitive species.

The seasonal occupancy and diel behaviour of Antarctic sperm whales revealed by acoustic monitoring

The seasonal occupancy and diel behaviour of sperm whales (Physeter macrocephalus) was investigated using data from long-term acoustic recorders deployed off east Antarctica. An automated method for investigating acoustic presence of sperm whales was developed, characterised, and applied to multi-year acoustic datasets at three locations. Instead of focusing on the acoustic properties of detected clicks, the method relied solely on the inter-click-interval (ICI) for determining presence within an hour-long recording. Parameters for our classifier were informed by knowledge of typical vocal behaviour of sperm whales. Sperm whales were detected predominantly from Dec-Feb, occasionally in Nov, Mar, Apr, and May, but never in the Austral winter or early spring months. Ice cover was found to have a statistically significant negative effect on sperm whale presence. In ice-free months sperm whales were detected more often during daylight hours and were seldom detected at night, and this effect was also statistically significant. Seasonal presence at the three east Antarctic recording sites were in accord with what has been inferred from 20th century whale catches off western Antarctica and from stomach contents of whales caught off South Africa.

Acoustic and biological trends on coral reefs off Maui, Hawaii

Coral reefs are characterized by high biodiversity, and evidence suggests that reef soundscapes reflect local species assemblages. To investigate how sounds produced on a given reef relate to abiotic and biotic parameters and how that relationship may change over time, an observational study was conducted between September 2014 and January 2016 at seven Hawaiian reefs that varied in coral cover, rugosity, and fish assemblages. The reefs were equipped with temperature loggers and acoustic recording devices that recorded on a 10% duty cycle. Benthic and fish visual survey data were collected four times over the course of the study. On average, reefs ranged from 0 to 80% live coral cover, although changes between surveys were noted, in particular during the major El Nino-related bleaching event of October 2015. Acoustic analyses focused on two frequency bands (50–1200 and 1.8–20.5 kHz) that corresponded to the dominant spectral features of the major sound-producing taxa on these reefs, fish, and snapping shrimp, respectively. In the low-frequency band, the presence of humpback whales (December–May) was a major contributor to sound level, whereas in the high-frequency band sound level closely tracked water temperature. On shorter timescales, the magnitude of the diel trend in sound production was greater than that of the lunar trend, but both varied in strength among reefs, which may reflect differences in the species assemblages present. Results indicated that the magnitude of the diel trend was related to fish densities at low frequencies and coral cover at high frequencies; however, the strength of these relationships varied by season. Thus, long-term acoustic recordings capture the substantial acoustic variability present in coral-reef ecosystems and provide insight into the presence and relative abundance of sound-producing organisms.

Spatial and temporal occurrence of killer whale ecotypes off the outer coast of Washington State, USA

Three killer whale Orcinus orca ecotypes inhabit the northeastern Pacific: residents, transients, and offshores. To investigate intraspecific differences in spatial and temporal occur-rence off the outer coast of Washington State, USA, 2 long-term acoustic recorders were deployed from July 2004 to August 2013: one off the continental shelf in Quinault Canyon (QC) and the other on the shelf, off Cape Elizabeth (CE). Acoustic encounters containing pulsed calls were analyzed for call types attributable to specific ecotypes, as no calls are shared between ecotypes. Both sites showed killer whale presence year-round, although site CE had a higher number of days with en- counters overall. Transients were the most common ecotype at both sites and were encountered mainly during the spring and early summer. Residents were encountered primarily at site CE and showed potential seasonal segregation between the 2 resident communities, with northern resi- dents present mainly during summer and early fall when southern residents were not encountered. Offshore encounters were higher at site QC, with little evidence for seasonality. Spatial and temporal variability of residents and transients matches the distribution of their prey and can potentially be used for further inferences about prey preferences for different transient groups.

Acoustic and biological trends on coral reefs off Maui, Hawaii

Coral reef soundscapes comprise a range of biological sounds. To investigate how the sounds produced on a given reef relate to the species present, 7 Hawaiian reefs that varied in their species assemblages were equipped with acoustic recorders operating on a 10% duty cycle for 16 months, starting in September 2014. Benthic and fish visual surveys were conducted 4 times over the course of the study. Acoustic analyses were carried out in 2 frequency bands (50-1200 Hz and 1.8-20.5 kHz) that corresponded with the spectral features of the major sound-producing taxa on these reefs, fish and snapping shrimp, respectively. In the low-frequency band, the presence of humpback whales (December-May) was the major driver of sound level, whereas in the high-frequency band sound level closely tracked water temperature. On shorter timescales, the magnitude of the diel trend varied in strength among reefs, which may reflect differences in the species assemblages present. Regression trees indicated that, at low frequencies, the relationship between species assemblages and acoustic parameters varied by season; however, at high frequencies, a given reef was generally most like itself over time. Thus, long-term acoustic recordings can capture and distill the substantial acoustic variability present in coral reef ecosystems.

Fish bioacoustics, big data, and changing ocean soundscapes

For many species of vocalizing fishes, reproductive calls form sustained choruses lasting for weeks to months, and are the dominant biological sounds in marine and freshwater habitats worldwide. The timing, duration, and magnitude of these choruses is strongly influenced by environmental factors and may further be influenced by increase in anthropogenic noise. Decades of pioneering work on fish hearing has established a solid scientific foundation to inform what and how fishes hear. By combining this foundation on the auditory system with long-term acoustic recordings of wild populations, we can link organismal-level and population-level processes in fishes to evaluate how they produce, perceive, and respond to sound. The scientific community is now in a place to begin understanding mechanisms and magnitudes of anthropogenic noise impacts on fishes and evaluate how human activities may influence fish populations. An integrative and comparative approach to the study of fish bioacoustics demonstrates how fundamental principles of auditory physiology, behavior, and ecology can inform our interpretation of changing ocean soundscapes.

Information obtained from long-term acoustic recordings: applying bioacoustic techniques for monitoring wetland birds during breeding season

Digital acoustic recording devices allow for recording of vocal activity of animals over long time periods and therefore open new perspectives for monitoring. In this paper results of a bioacoustic monitoring of birds in a wetland restoration area in Mecklenburg-Western Pomerania in Northeastern Germany are presented. Ninety minutes of sound recordings were made daily from March 29 until May 30 2012 using a four-channel recording setup. The recording schedule was mainly designed for nocturnal birds but also included dawn and dusk choruses. By listening to the recordings, supported by spectrographic visualization, 61 bird species could be detected. Despite the fact that most of the recordings were made during the night, no period of absolute silence was detected. The occurrence data allow some conclusions on the state of the bird species, e.g., breeding bird, resting during migration, or guest only. The bioacoustic monitoring revealed the occurrence of rare nocturnal birds such as Little Crake (Porzana parva), Spotted Crake (Porzana porzana) and Short-eared Owl (Asio flammeus). For the Eurasian Bittern (Botaurus stellaris) data on vocal activity and population size could be obtained. The study has shown that automated acoustic recordings can be applied for long-term monitoring of bird populations in remote areas.

Acoustic signalling for mate attraction in crickets: Abdominal ganglia control the timing of the calling song pattern

Decoding the neural basis of behaviour requires analysing how the nervous system is organised and how the temporal structure of motor patterns emerges from its activity. The stereotypical patterns of the calling song behaviour of male crickets, which consists of chirps and pulses, is an ideal model to study this question. We applied selective lesions to the abdominal nervous system of field crickets and performed long-term acoustic recordings of the songs. Specific lesions to connectives or ganglia abolish singing or reliably alter the temporal features of the chirps and pulses. Singing motor control appears to be organised in a modular and hierarchically fashion, where more posterior ganglia control the timing of the chirp pattern and structure and anterior ganglia the timing of the pulses. This modular organisation may provide the substrate for song variants underlying calling, courtship and rivalry behaviour and for the species-specific song patterns in extant crickets.

Red shift, blue shift: investigating Doppler shifts, blubber thickness, and migration as explanations of seasonal variation in the tonality of Antarctic blue whale song.

The song of Antarctic blue whales (Balaenoptera musculus intermedia) comprises repeated, stereotyped, low-frequency calls. Measurements of these calls from recordings spanning many years have revealed a long-term linear decline as well as an intra-annual pattern in tonal frequency. While a number of hypotheses for this long-term decline have been investigated, including changes in population structure, changes in the physical environment, and changes in the behaviour of the whales, there have been relatively few attempts to explain the intra-annual pattern. An additional hypothesis that has not yet been investigated is that differences in the observed frequency from each call are due to the Doppler effect. The assumptions and implications of the Doppler effect on whale song are investigated using 1) vessel-based acoustic recordings of Antarctic blue whales with simultaneous observation of whale movement and 2) long-term acoustic recordings from both the subtropics and Antarctic. Results from vessel-based recordings of Antarctic blue whales indicate that variation in peak-frequency between calls produced by an individual whale was greater than would be expected by the movement of the whale alone. Furthermore, analysis of intra-annual frequency shift at Antarctic recording stations indicates that the Doppler effect is unlikely to fully explain the observations of intra-annual pattern in the frequency of Antarctic blue whale song. However, data do show cyclical changes in frequency in conjunction with season, thus suggesting that there might be a relationship among tonal frequency, body condition, and migration to and from Antarctic feeding grounds.