Bioacoustic Surveys Research Articles

All thresholds barred: direct estimation of call density in bioacoustic data

Passive acoustic monitoring (PAM) studies generate thousands of hours of audio, which may be used to monitor specific animal populations, conduct broad biodiversity surveys, detect threats such as poachers, and more. Machine learning classifiers for species identification are increasingly being used to process the vast amount of audio generated by bioacoustic surveys, expediting analysis and increasing the utility of PAM as a management tool. In common practice, a threshold is applied to classifier output scores, and scores above the threshold are aggregated into a detection count. The choice of threshold produces biased counts of vocalizations, which are subject to false positive/negative rates that may vary across subsets of the dataset. In this work, we advocate for directly estimating call density: The proportion of detection windows containing the target vocalization, regardless of classifier score. We propose a validation scheme for estimating call density in a body of data and obtain, through Bayesian reasoning, probability distributions of confidence scores for both the positive and negative classes. We use these distributions to predict site-level call densities, which may be subject to distribution shifts (when the defining characteristics of the data distribution change). These methods may be applied to the outputs of any binary detection classifier operating on fixed-size audio input windows. We test our proposed methods on a real-world study of Hawaiian birds and provide simulation results leveraging existing fully annotated datasets, demonstrating robustness to variations in call density and classifier model quality.

Automatic Classification of Monosyllabic and Multisyllabic Birds Using PDHF

Bioacoustics plays an important role in the conservation of bird species. Bio-acoustic surveys based on autonomous audio recording are both cost-effective and time-efficient. However, there are many bird species with different patterns of vocalization, and it is a challenging task to deal with them. Previous studies have revealed that many authors focus on the segmentation of bird audio without considering specific patterns of bird vocalization. Based on the existing literature, currently there is no work on the segmentation of monosyllabic and multisyllabic birds, separately. Therefore, this research addresses the aforementioned concern and also proposes a collection of audio features named ‘Perceptual, Descriptive, and Harmonic Features (PDHFs)’ that gives promising results in the classification of bird vocalization. Moreover, the classification results improved when monosyllabic and multisyllabic birds were classified separately. To analyze the performance of PDHFs, different classifiers were used in which Artificial neural network (ANN) outperformed other classifiers and demonstrated an accuracy of 98%.

Ears in the Sky: Potential of Drones for the Bioacoustic Monitoring of Birds and Bats

In the context of global biodiversity loss, wildlife population monitoring is a major challenge. Some innovative techniques such as the use of drones—also called unmanned aerial vehicle/system (UAV/UAS)—offer promising opportunities. The potential of UAS-based wildlife census using high-resolution imagery is now well established for terrestrial mammals or birds that can be seen on images. Nevertheless, the ability of UASs to detect non-conspicuous species, such as small birds below the forest canopy, remains an open question. This issue can be solved with bioacoustics for acoustically active species such as bats and birds. In this context, UASs represent an interesting solution that could be deployed on a larger scale, at lower risk for the operator, and over hard-to-reach locations, such as forest canopies or complex topographies, when compared with traditional protocols (fixed location recorders placed or handled by human operators). In this context, this study proposes a methodological framework to assess the potential of UASs in bioacoustic surveys for birds and bats, using low-cost audible and ultrasound recorders mounted on a low-cost quadcopter UAS (DJI Phantom 3 Pro). The proposed methodological workflow can be straightforwardly replicated in other contexts to test the impact of other UAS bioacoustic recording platforms in relation to the targeted species and the specific UAS design. This protocol allows one to evaluate the sensitivity of UAS approaches through the estimate of the effective detection radius for the different species investigated at several flight heights. The results of this study suggest a strong potential for the bioacoustic monitoring of birds but are more contrasted for bat recordings, mainly due to quadcopter noise (i.e., electronic speed controller (ESC) noise) but also, in a certain manner, to the experimental design (use of a directional speaker with limited call intensity). Technical developments, such as the use of a winch to safely extent the distance between the UAS and the recorder during UAS sound recordings or the development of an innovative platform, such as a plane–blimp hybrid UAS, should make it possible to solve these issues.

Artificial light during the polar night disrupts Arctic fish and zooplankton behaviour down to 200\u2009m depth

For organisms that remain active in one of the last undisturbed and pristine dark environments on the planet—the Arctic Polar Night—the moon, stars and aurora borealis may provide important cues to guide distribution and behaviours, including predator-prey interactions. With a changing climate and increased human activities in the Arctic, such natural light sources will in many places be masked by the much stronger illumination from artificial light. Here we show that normal working-light from a ship may disrupt fish and zooplankton behaviour down to at least 200 m depth across an area of >0.125 km2 around the ship. Both the quantitative and qualitative nature of the disturbance differed between the examined regions. We conclude that biological surveys in the dark from illuminated ships may introduce biases on biological sampling, bioacoustic surveys, and possibly stock assessments of commercial and non-commercial species.

Molossid unlimited: extraordinary extension of range and unusual vocalization patterns of the bat, Promops centralis

AbstractThe big crested mastiff bat, Promops centralis, occurs in Central and South America, but knowledge of its ecology is limited due to its open space hunting strategy, making captures extremely challenging. Notwithstanding, members of the species produce echolocation calls that are easy to identify. After recording calls of P. centralis 1,500 km away from its known range in Brazil, we hypothesized that the distribution range of this species was probably greatly underestimated. To improve the accuracy of P. centralis’ real distribution, we employed acoustic surveys throughout parts of Brazil, conducted after a bibliographic review to gather additional records, and used MaxEnt to model the species’ potential distribution. We have found that P. centralis has a much wider distribution in South America than previously thought, adding more than 3.8 million km2 to its former known area. We also describe an unusual vocalization pattern of P. centralis, with individuals emitting at least three very distinct but highly variable calls. This study shows that bioacoustic surveys and species distribution models can complement traditional methodologies in studying species that are difficult to capture, such as P. centralis, potentially contributing to more effective conservation and management plans.

Acoustic Monitoring Reveals a Diverse Forest Owl Community, Illustrating Its Potential for Basic and Applied Ecology

Bioacoustic surveys are an increasingly common tool for wildlife research due to increasing software performance and decreasing hardware costs. We conducted landscape-scale passive acoustic surveys across the northern Sierra Nevada to assess the distribution and density of barred owls (Strix varia) at the leading edge of their ongoing range expansion. From the resulting audio, we derived occupancy data on barred owls and spotted owls (S. occidentalis). We also documented seven other owl species; because the audio was passively recorded, it will be possible to extract systematic monitoring data on those nine species and others. Our findings reveal the tremendous latent potential of passively recorded acoustic survey data for studying otherwise cryptic species and biological communities. Photo credit: Connor M. Wood. Photo credit: Connor M. Wood. These photographs illustrate the article “Acoustic monitoring reveals a diverse forest owl community, illustrating its potential for basic and applied ecology” by Connor M. Wood, R.J. Gutiérrez, and M. Zachariah Peery published in Ecology. https://doi.org/10.1002/ecy.2764

Acoustic profiling of Orthoptera: present state and future needs

Bioacoustic monitoring and classification of animal communication signals has developed into a powerful tool for measuring and monitoring species diversity within complex communities and habitats. The high number of stridulating species among Orthoptera allows their detection and classification in a non-invasive and economic way, particularly in habitats where visual observations are difficult or even impossible, such as tropical rainforests. Major sound archives were queried for Orthoptera songs, with special emphasis on usability as reference training libraries for computer algorithms. Orthoptera songs are highly stereotyped, reliable taxonomic features. However, exploitation of songs for acoustic profiling is limited by the small number of reference recordings: existing song libraries represent only about 1000 species, mainly from Europe and North America, covering less than 10% of extant stridulating Orthoptera species. Available databases are fragmented and lack tools for song annotation and efficient feature-based searching. Results from recent bioacoustic surveys illustrate the potential of the method, but also the challenges and bottlenecks impeding further progress. A major problem is time-consuming data analysis of recordings. Computer-aided identification software exists for classification and identification of cricket and grasshopper songs, but these tools are still far from practical for field application.A framework for acoustic profiling of Orthoptera should consist of the following components: (1) Protocols for standardized acoustic sampling, at species and community levels, using acoustic data loggers for autonomous long-term recordings; (2) Open access to and efficient management of song data and voucher specimens, involving the Orthoptera Species File (OSF) and Global Biodiversity Information Facility (GBIF); (3) An infrastructure for automatized analysis and song classification; and (4) Complementation and improvement of Orthoptera sound libraries using OSF as the taxonomic backbone and repository for representative song recordings. Taxonomists should be encouraged, or even obliged, to deposit original recordings, particularly if they form part of species descriptions or revisions.

Do bioacoustic conditions reflect species diversity? A case study from four tropical marine habitats

New tools, such as passive acoustic monitoring, can be helpful for measuring levels of biodiversity in habitats that are otherwise difficult to sample. Here, we tested the utility of acoustic measurements in shallow coastal waters by conducting simultaneous bioacoustic and biodiversity surveys in four habitat types in Panama: mangrove, reef, seagrass, and sand. We found that acoustic measurements in the “low band” (<1000 Hz) were positively correlated with cryptic fish species richness. However, our 24-h acoustic recordings revealed a clear toadfish chorus at dusk, which masked other fish sounds and confounded results from newer acoustic indices such as acoustic entropy and acoustic complexity. Band level in the “high band” (3,000-10,000 Hz) did not differ across habitat types and was not significantly correlated to biodiversity measurements. Our study demonstrates that bioacoustic surveys can help scientists to identify certain cryptic, soniferous species, and should be used in tandem with traditional biodiversity surveys. Additional research is needed in the marine environment to validate the utility of the newer acoustic indices.

Bat Echolocation Call Identification for Biodiversity Monitoring: A Probabilistic Approach

SummaryBat echolocation call identification methods are important in developing efficient cost-effective methods for large-scale bioacoustic surveys for global biodiversity monitoring and conservation planning. Such methods need to provide interpretable probabilistic predictions of species since they will be applied across many different taxa in a diverse set of applications and environments. We develop such a method using a multinomial probit likelihood with independent Gaussian process priors and study its feasibility on a data set from an on-going study of 21 species, five families and 1800 bat echolocation calls collected from Mexico, a hotspot of bat biodiversity. We propose an efficient approximate inference scheme based on the expectation propagation algorithm and observe that the overall methodology significantly improves on currently adopted approaches to bat call classification by providing an approach which can be easily generalized across different species and call types and is fully probabilistic. Implementation of this method has the potential to provide robust species identification tools for biodiversity acoustic bat monitoring programmes across a range of taxa and spatial scales.

Influence of Reservoirs Created by Small Dams on the Activity of Bats

Freshwater systems have been profoundly changed by the construction of dams, and the influence of dams on bat activity is poorly understood. In this study, we analyzed the effects of small dams along mountain streams on local bat communities. This work was carried out in five small streams located in the NE of Portugal using bioacoustic surveys during the summer of 2011. The present study confirms that, in the northeast areas of Portugal, the majority of bat species use artificial bodies of water for either drinking or foraging, but species differed in terms of their activity levels over the reservoirs when compared with intact stream habitats. As predicted, small dams in the study area were important centers of overall bat activity relative to other sampling areas. At the same time, feeding activity was also higher in the flooded areas. We also found that Pipistrellus pipistrellus, Myotis daubentonii, Pipistrellus kuhlii, Pipistrellus pygmaeus/Miniopterus schreibersii, Tadarida teniotis and Nyctalus leis...

Is hearing believing? Patterns of bird voice misidentification in an online quiz

This study aims to uncover patterns of species identification error in bioacoustic surveys of central Amazon birds. To quantify errors, we developed an on-line quiz based on vocalizations of an undisclosed set of 41 antbird (Thamnophilidae) and woodcreeper (Dendrocolaptinae) species. We invited experts to answer the quiz and obtained 820 answers from 20 participants. The answers were compared to the results of a binomial experiment with a success probability of 0.5; i.e. we examined whether participants identified species correctly more often than expected by the toss of a coin with a 50% chance of producing the right identification. We also examined whether species were correctly identified more often than expected under a similar coin toss experiment. Quiz answers were compiled in a triangular matrix showing species ranked by taxonomic order on both axes. From the triangular matrix we can ask whether closely-related species were mistaken for each other, i.e. confused, more often than distantly- related species. We tested this hypothesis with a null model approach that compared the mean taxonomic distance between confused species in the observed matrix to the distribution of mean taxonomic distances between confused species in 10,000 randomized matrices. Finally, we drew a dendrogram to represent the similarity between species with regard to the distribution of identification errors. The 20 participants who took the quiz showed substantial variation in their ability to identify species correctly. Fourteen species were correctly identified more often than expected at random, while only one was misidentified more often than expected at random. The observed mean distance between confused species was smaller than all of the mean distances from the randomized, null-model matrices, indicating that confusions are more frequent between closely related species than between distant ones.

Acoustic identification of Mexican bats based on taxonomic and ecological constraints on call design

Summary Monitoring global biodiversity is critical for understanding responses to anthropogenic change, but biodiversity monitoring is often biased away from tropical, megadiverse areas that are experiencing more rapid environmental change. Acoustic surveys are increasingly used to monitor biodiversity change, especially for bats as they are important indicator species and most use sound to detect, localise and classify objects. However, using bat acoustic surveys for monitoring poses several challenges, particularly in megadiverse regions. Many species lack reference recordings, some species have high call similarity or differ in call detectability, and quantitative classification tools, such as machine learning algorithms, have rarely been applied to data from these areas. Here, we collate a reference call library for bat species that occur in a megadiverse country, Mexico. We use 4685 search‐phase calls from 1378 individual sequences of 59 bat species to create automatic species identification tools generated by machine learning algorithms (Random Forest). We evaluate the improvement in species‐level classification rates gained by using hierarchical classifications, reflecting either taxonomic or ecological constraints (guilds) on call design, and examine how classification rate accuracy changes at different hierarchical levels (family, genus and guild). Species‐level classification of calls had a mean accuracy of 66%, and the use of hierarchies improved mean species‐level classification accuracy by up to 6% (species within families 72%, species within genera 71·2% and species within guilds 69·1%). Classification accuracy to family, genus and guild‐level was 91·7%, 77·8% and 82·5%, respectively. The bioacoustic identification tools we have developed are accurate for rapid biodiversity assessments in a megadiverse region and can also be used effectively to classify species at broader taxonomic or ecological levels. This flexibility increases their usefulness when there are incomplete species reference recordings and also offers the opportunity to characterise and track changes in bat community structure. Our results show that bat bioacoustic surveys in megadiverse countries have more potential than previously thought to monitor biodiversity changes and can be used to direct further developments of bioacoustic monitoring programs in Mexico.

Spatial patterns of pollock and zooplankton distribution in the Pribilof Islands, Alaska nursery area and their relationship to pollock recruitment

Data from six years, September 1994–1999, of bio-acoustic surveys near the Pribilof Islands, Alaska, (AK), from which age-0 walleye pollock school and zooplankton patch locations have been extracted, were analysed using image-processing methods. Multiple passes along four transects in this major pollock nursery area were examined. The data showed high year-to-year variability in overall abundance of both fish and plankton, but consistent abundance differences between the transects. Juvenile pollock abundance was generally highest in the shallow shelf region to the north of the Pribilof Islands and lowest in the mouth of the Pribilof Canyon to the south. Plankton biomass patterns tended to be the reverse. We identified fronts and regions within the transects based on changes in hydrograpy (e.g. vertical stratification) and bathymetry. Diel migration patterns of pollock and zooplankton within these regions appear to depend on the degree of stratification, the depth, the size of the pollock and the relative abundance of the pollock and zooplankton. Several hypotheses are also discussed concerning the relationship of pollock recruitment year-class strength to large year classes including differences in the environmental conditions, the pervasiveness of the pollock, the size of juvenile pollock and the density of predators.

Ascending and descending particle flux from hydrothermal plumes at Endeavour Segment, Juan de Fuca Ridge

Bio-acoustic surveys and associated zooplankton net tows have documented anomalously high concentrations of zooplankton within a 100 m layer above the hydrothermal plumes at Endeavour Segment, Juan de Fuca Ridge. These and other data suggest that congregating epi-plume zooplankton are exploiting a food substrate associated with the hydrothermal plume. Ascending, organic-rich particles could provide a connection. Consequently, two paired sequentially sampling ascending and descending particle flux traps and a current meter were deployed on each of three moorings from July 1994 to May 1995. Mooring sites included an on-axis site (OAS; 47°57.0′N, 129°05.7′W) near the main Endeavour vent field, a “down-current” site 3 km west of the main vent field (WS), and a third background station 43 km northeast of the vent field (ES). Significant ascending and descending particle fluxes were measured at all sites and depths. Lipid analyses indicated that ascending POC was derived from mid-depth and deep zooplankton whereas descending POC also contained a component of photosynthetically derived products from the sea surface. Highest ascending POC fluxes were found at the hydrothermal plume-swept sites (OAS and WS). The limited data available, however, precludes an unequivocal conclusion that hydrothermal processes contribute to the ascending flux of organic carbon at each site. Highest ascending to descending POC flux ratios were also found at WS. Observed trends in POC, PMn/PTi, and PFe/PTi clearly support a hydrothermal component to the descending flux at the plume-swept WS site (no descending data was recovered at OAS) but not at the background ES site. Alternative explanations for ascending particle data are discussed. First-order calculations for the organic carbon input (5–22 mg C m −2 d −1) required to sustain observed epi-plume zooplankton anomalies at Endeavour are comparable both to measured total POC flux to epi-plume depths (2–5 mg C m −2 d −1: combined hydrothermal and surface derived organic carbon) and to estimates of the total potential in situ organic carbon production (2–9 mg C m −2 d −1) from microbial oxidation of hydrothermal plume H 2, CH 4 and NH 4 +.

Distribution and Abundance of Phytoplankton, Zooplankton, Ichthyoplankton, and Micronekton in the Deepwater Gulf of Mexico

Expeditions in the 1960s and 1970s are the basis for the general paradigm that standing stocks and productivity of phytoplankton are both low ( <0.1 mg chl·m-3; <150 mg C.m-2·d-1) seaward of the shelf-slope break in the Gulf of Mexico. The present review supports this description of the mean (stable) state but also shows "hot spots" in primary production (>2 g C·m-2·d-1) occur when/where nutrient availability is locally enhanced seaward of the shelf-slope break. Recent collections with Bongo and MOCNESS nets, midwater trawls, and bioacoustic surveys of the Loop Current and associated cyclonic and anticyclonic eddies in the Gulf of Mexico show that these deepwater "hot spots" have higher stocks of zooplankton and micronekton as well. The local aggregations ranged in size from coarse- to meso- spatial scales (10s to 100s of kilometers) though locations of such "oases" were spatially variable along the continental margin.