Search

Passive acoustic monitoring (PAM) is an effective, non-intrusive method for studying ecosystems, but obtaining meaningful ecological information from its large number of audio files is challenging. In this study, we take advantage of the expected animal behavior at different times of the day (e.g., higher acoustic animal activity at dawn) and develop a novel approach to use these time-based patterns. We organize PAM data into 24-hour temporal blocks formed with sound features from a pretrained VGGish network. These features feed a 1D convolutional neural network with a class activation mapping technique that gives interpretability to its outcomes. As a result, these diel-cycle blocks offer more accurate and robust hour-by-hour information than using traditional ecological acoustic indices as features, effectively recognizing key ecosystem patterns. [ABSTRACT FROM AUTHOR]

Timo Maran has defined "biosemiotic criticism" as the study of human culture with an emphasis on the recognition that all forms of life are organized by sign processes. That approach guides this investigation of the sonic devices and practices that have been used in encounters between birds and humans in agricultural spaces. "Bird-scaring" has been a long-standing component of the semiotic relationship between humans and birds in what I am calling the agricultural semiosphere. The struggle between humans and "pest" species over the control of agricultural resources has spurred technological development, and an examination of the sonic tools that have been involved in bird-scaring practices, from wooden rattles to digital sound recordings, intersects with scholarship on sound technology and media studies. The field of ecoacoustics provides a conceptual framework for the sonic dynamics of the interspecies communication under examination. To the extent that the essay also explores the sign relations between humans and birds it has a place in the growing corpus of biosemiotic criticism. The historical account of bird-scaring practices is presented in three sections, demarcated by changes in technology as well as shifts in the mode of semiotic reference. Across the analysis, a historical approach to bird-scaring is interwoven with a discussion of biosemiotic and ecoacoustic themes and concepts. [ABSTRACT FROM AUTHOR]

In soundscape ecology analysis, the use of acoustic features is well established and offers important baselines to ecological analyses. However, in many cases, the problem is difficult due to high-class overlap in terms of time-frequency characteristics, as well as the presence of noise. Deep neural networks have become state-of-the-art for feature learning in many multi-class applications, but they often present issues such as over-fitting or achieve unbalanced performances for different classes, which can hamper the deployment of such models in realistic scenarios. In the context of counting the number of classes in observations, the quantification task is attracting attention and was shown to be effective in other applications. This paper investigates the use of quantification combined with classification loss in order to train a convolutional neural network to classify species of birds and anurans. Results indicate quantification has advantages over both acoustic features alone and the use of regular classification networks, in particular in terms of generalization and class recall making it a suitable choice for segregation tasks related to soundscape ecology. Moreover, we show that a more compact network can outperform a deeper one for fine-grained scenarios of birds and anurans species. [ABSTRACT FROM AUTHOR]

Biosemiotics to date has focused on the exchange of signals between organisms, in line with bioacoustics; consideration of the wider acoustic environment as a semiotic medium is under-developed. The nascent discipline of ecoacoustics, that investigates the role of environmental sound in ecological processes and dynamics, fills this gap. In this paper we introduce key ecoacoustic terminology and concepts in order to highlight the value of ecoacoustics as a discipline in which to conceptualise and study intra- and interspecies semiosis. We stress the inherently subjective nature of all sensory scapes (vivo-, land-, vibro- and soundscapes) and propose that they should always bear an organismic attribution. Key terms to describe the sources (geophony, biophony, anthropophony, technophony) and scales (sonotopes, soundtopes, sonotones) of soundscapes are described. We introduce epithets for soundscapes to point to the degree to which the global environment is implicated in semiosis (latent, sensed and interpreted soundscapes); terms for describing key ecological structures and processes (acoustic community, acoustic habitat, ecoacoustic events) and examples of ecoacoustic events (choruses and noise) are described. The acoustic eco-field is recognized as the semiotic model that enables soniferous species to intercept core resources like food, safety and roosting places. We note that whilst ecoacoustics to date has focused on the critical task of the development of metrics for application in conservation and biodiversity assessment, these can be enriched by advancing conceptual and theoretical foundations. Finally, the mutual value of integrating ecoacoustic and biosemiotics perspectives is considered. [ABSTRACT FROM AUTHOR]

Methods for surveying biodiversity range from rapid biodiversity assessments to full taxonomic inventories. Recently, surveying of biological sounds has been suggested as a means to rapidly quantify ecosystem health and biodiversity as it enables remote, non-intrusive and continuous surveillance of communities at a high temporal resolution. The present study explores the link between two acoustic features, i.e. the sound pressure level and acoustic complexity, of three reef sites on Bora Bora Island, French Polynesia and their biological and physical features, in terms of fish abundance and type of substrate. Sites with a high number of fish appeared to be noisier especially in the low frequencies while sites with more living coral cover presented a higher acoustic complexity. These indices may therefore provide valuable information on different aspects of the biology and ecology of coral reefs. However, some discrepancies in the daily periodicity of both indices remain between sites, highlighting the need to further strengthen the potential of using acoustic surveys in immediate state assessment of ocean ecosystems and better manage coral reef conservation in particular. [ABSTRACT FROM AUTHOR]

How vocal organisms share acoustic space has primarily received attention in terrestrial environments. Comparable studies in marine environments, however, remain rare. By recording sounds on a coral reef in French Polynesia for 48 h and 24 h, this study provides first insights on how different sound types are distributed within the acoustic space and may create acoustic niches optimizing acoustic communication within a highly diverse community containing numerous soniferous fish species. Day-time was dominated by two to six sound types, while recordings performed at night revealed a more diverse vocal community made of up to nineteen sound types. Calling activity was distributed over time allowing each sound type to dominate the soundscape sequentially. Additionally, differences in the acoustic features of sounds occurring during the same period were observed. Such partitioning in time and acoustic spaces would reduce potential overlaps of sounds produced by vocal species living in sympatry in coral reefs. [ABSTRACT FROM AUTHOR]

Context: Roads are a widespread feature of landscapes worldwide, and road traffic sound potentially makes nearby habitat unsuitable for acoustically communicating organisms. It is important to understand the influence of roads at the soundscape level to mitigate negative impacts of road sound on individual species as well as subsequent effects on the surrounding landscape. Objectives: We seek to characterize the relationship between anthropogenic and biological sounds in western New York and assess the extent to which available traffic data explains variability in anthropogenic noise. Methods: Recordings were obtained in the spring of 2016 at 18 sites throughout western New York. We used the Welch Power Spectral Density (PSD) at low frequencies (0.5–2 kHz) to represent anthropogenic noise and PSD values at higher frequencies (2–11 kHz) to represent biological sound. Relationships were modeled using a novel two-stage hierarchical Bayesian model utilizing beta regression and basis splines. Results: Model results and map predictions illustrate that anthropogenic noise and biological sound have an inverse relationship, and anthropogenic noise is greatest in close proximity to high traffic volume roads. The predictions have large uncertainty, resulting from the temporal coarseness of public road data used as a proxy for traffic sound. Conclusions: Results suggest that finer temporal resolution traffic sound data, such as crowd-sourced time-indexed traffic data from geographic positioning systems, might better account for observed temporal changes in the soundscape. The use of such data, in combination with the proposed modeling framework, could have important implications for the development of sound management policies. [ABSTRACT FROM AUTHOR]

Rapid decreases in frog populations have been spotted worldwide, which are regarded as one of the most critical threats to the global biodiversity. Recent advances in acoustic sensors provide a novel way to assess frog vocalizations and further optimize the global protection policy. Specifically, frog populations can be reflected by detecting frog species using collected recordings. Previous studies have explored various acoustic features for classifying frog calls. However, few studies investigate visual features for frog call classification, which have been successfully used in acoustic event detection, speech/speaker recognition. In this study, various acoustic and visual features are proposed for frog call classification: MPEG-7 audio descriptor, syllable duration, oscillation rate, entropy related features, linear prediction codings, Mel-frequency Cepstral coefficients, local binary patterns, and histogram of oriented gradients. After segmenting continuous frog calls into individual syllables, different constructed feature sets are evaluated with a k-nearest neighbor classifier and support vector machines. Comprehensive results on 16 frog species demonstrate the effectiveness of both acoustic and visual features for classifying frog calls. [ABSTRACT FROM AUTHOR]

Invasive species are a major concern for the maintenance of ecosystem services and biodiversity but are difficult to mitigate. Upstream solutions to prevent their impact, including their detection, are needed. Wasmannia auropunctata, an invasive ant living in vagile supercolonies, is especially hard to track and is a major threat for tropical ecosystems and local animal communities. As part of such tropical communities, crickets are sensitive to ecological conditions, easy to collect, detectable and identifiable through their species-specific calls. Here, we evaluated the use of an acoustic community of crickets as an indicator of the presence of W. auropunctata in New Caledonia. We evaluated the dominance of the crickets in the soundscape, describe the cricket community structure and diversity along a shrubland to forest gradient, characterize these cricket communities structure and diversity in the light of ongoing invasion by W. auropunctata, and identify cricket species’ indicators of the invasion. Acoustic recordings collected on 24 sites were described using human-listening and spectrographic visualization. The results demonstrated a clear dominance of the cricket group in the New Caledonian nocturnal soundscapes. Each habitat harbored a specific acoustic cricket community related to specific environmental attributes including vegetation height, daily variation of humidity and temperature. The presence of W. auropunctata was significantly associated with a lower cricket acoustic activity and species richness at night. Of the 19 species detected, four nocturnal species were identified as indicator of non-invaded forests and preforests. This work supports the use of acoustic as an alternative method to detect invasion. [ABSTRACT FROM AUTHOR]

Sound is an inherent component of the environment that provides conditions and information necessary for many animal activities. Soniferous species require specific acoustic and physical conditions suitable for their signals to be transmitted, received, and effectively interpreted to successfully identify and utilize resources in their environment and interact with conspecifics and other heterospecific organisms. We propose the Acoustic Habitat Hypothesis to explain how the acoustic environment influences habitat selection of sound-dependent species. We postulate that sound-dependent species select and occupy habitats with unique acoustic characteristics that are essential to their functional needs and conducive to the threshold of sound frequency they produce and detect. These acoustic habitats are based on the composition of biophony, geophony, and technophony in the soundscape and on the biosemiotics mechanisms described in the eco-field hypothesis. The Acoustic Habitat Hypothesis initiates questions of habitat selection that go beyond the physical attributes of the environment by applying ecoacoustics theory. We outline the theoretical basis of the Acoustic Habitat Hypothesis and provide examples from the literature to support its assumptions. The concept of acoustic habitats has been documented in the literature for many years but here, we accurately and extensively define acoustic habitat and we put this concept into a unified theory. We also include perspectives on how the Acoustic Habitat Hypothesis can stimulate a paradigm shift in conservation strategies for threatened and endangered species. [ABSTRACT FROM AUTHOR]

Passive acoustic recording of marine noise has advanced considerably over recent years. For a long time, a lack of widely available technology limited the acquisition of long-term acoustic data sets to a small number of large, cabled installations mostly restricted to military use. For other users, recordings were limited by the available technology to short snapshots of minutes to possibly days of data at a time. As technology has improved, passive acoustic monitoring has shown marine soundscapes are filled with biotic and abiotic sounds that occur on a range of often unpredictable timescales. Thus, snapshot recordings can lead to biased data. In 1999, the Centre for Marine Science and Technology, together with Australia's Defence Science and Technology Organisation, began developing remote underwater sound recorders to increase the duration and quality of recordings. As time passed, the sound recorders were developed significantly, have been deployed over 600 times at a variety of Australian and international locations and have identified a plethora of biological, geophysical and anthropogenic sound sources. This paper presents a brief history of the recorders' development and characteristics, some examples of the information they have provided and future direction for their next generation. [ABSTRACT FROM AUTHOR]

Context: Winter soundscapes are likely different from soundscapes in other seasons considering wildlife vocalizations (biophony) decrease, wind events (geophony) increase, and winter vehicle noise (technophony) occurs. The temporal variation and spatial relationships of soundscape components to the landscape in winter have not been quantified and described until now. Objectives: Our objectives were to determine the temporal and spatial variation and acoustic-environmental relationships of a winter soundscape in south-central Alaska. Methods: We recorded ambient sounds at 62 locations throughout Kenai National Wildlife Refuge (December 2011-April 2012). We calculated the normalized power spectral density in 59,597 recordings and used machine learning to determine acoustic-environmental relationships and produce spatial models of soundscape components. Results: Geophony was the most prevalent component (84 %) followed by technophony (15 %), and biophony (1 %). Geophony occurred primarily at night, varied little by month, and was strongly associated with lakes. Technophony and biophony had similar temporal variation, peaking in April. Technophony occurred closer to urban areas and at locations with high snowmobile activity. Biophony occurred closer to rivers and was inversely related to snowmobile activity. Over 75 % of sample sites had >1 recordings of airplane or snowmobile noise, mainly in remote areas. Conclusions: The soundscape displayed distinct patterns across 24-h and monthly timeframes. These patterns were strongly associated with land cover variables which demonstrate discrete acoustic-environmental relationships exhibiting distinct spatial patterns in the landscape. Despite the predominance of geophony, the presence of technophony in this winter soundscape may have significant negative effects to wildlife and wilderness quality. [ABSTRACT FROM AUTHOR]

The sounds produced by animals have been a topic of research into animal behaviour for a very long time. If acoustic signals are undoubtedly a vehicle for exchanging information between individuals, environmental sounds embed as well a significant level of data related to the ecology of populations, communities and landscapes. The consideration of environmental sounds for ecological investigations opens up a field of research that we define with the term ecoacoustics. In this paper, we draw the contours of ecoacoustics by detailing: the main theories, concepts and methods used in ecoacoustic research, and the numerous outcomes that can be expected from the ecological approach to sound. Ecoacoustics has several theoretical and practical challenges, but we firmly believe that this new approach to investigating ecological processes will generate abundant and exciting research programs. [ABSTRACT FROM AUTHOR]