Search for: All records

Passive acoustic monitoring (PAM) is a powerful tool for ecological research, but recordings can be compromised by background noise such as wind. Addressing wind noise (e.g., clipping and masking) in bioacoustic data remains a challenge, especially as climate change is predicted to increase wind speeds, particularly near the poles. Adélie penguins (Pygoscelis adeliae), key indicators of the Antarctic ecosystem, are well-suited for PAM, where large-scale monitoring could assess climate-driven population changes—if wind noise is managed effectively. In this study, the convolutional neural network, BirdNET, inversely identifies unwanted sounds in Adélie penguin colony recordings. Multiple custom models were developed in which the background nontarget noise was Adélie vocalizations, and wind conditions (low, medium, and high) were the target classes. The best-performing model achieved an F-score of 0.43 and accuracy of 0.53. The high wind class within this model had a precision of 0.76 and recall of 0.94. A six-step workflow is presented for creating custom BirdNET models, evaluating their performance and determining an optimal confidence threshold prior to model application on an entire dataset. By automating unwanted sound detection, this approach enables researchers to efficiently identify and remove affected files, streamline data cleaning, and focus on recordings of interest for further analysis. 

Passive acoustic monitoring for bats has become a common method to determine species presence and activity levels. However, current acoustic methods are ineffective for monitoring species abundance at large summer colonies. We used synchronized acoustic and thermal-imaging data collected at 6 colonies of Myotis grisescens (Gray Bats) and found a significant positive relationship between acoustic energy and number of emerging bats. Our findings reinforce that acoustics have the potential to estimate population sizes of summer bat colonies. Additionally, we examined ultrasonic amplitude variance across 19 AudioMoth devices at 5 different gain settings and found significant differences among devices and settings. Further exploration into device variability and bat behavior are necessary to develop a robust model of population estimates using acoustic energy. 

Adélie penguins (Pygoscelis adeliae) are bioindicators for the rapidly changing Antarctic environment, making understanding their population dynamics and behavior of high research priority. However, collecting detailed population data throughout the breeding season on many colonies is difficult due to Antarctica’s harsh conditions and remote location. The colonial breeding ecology of Adélie penguins has led to the evolution of a highly vocal species with individualized calls, making them well-suited for passive acoustic monitoring (PAM) with autonomous recording. PAM units can potentially provide an easily deployable and scalable way to collect fine-scale data on population estimates and breeding phenology. Here I present a framework for using acoustic indices to monitor phenology of dense penguin colonies even under high wind conditions. I evaluate the relationship between acoustic indices such as RMS amplitude and penguin colony size between distinct breeding stages (incubation, guard, crèche, and fledge) on Torgersen and Humble Islands in the West Antarctic Peninsula with an automated pipeline implemented in R. Using PAM to interpret penguin vocalizations for population size and breeding phenology estimates could lead to the development of a real-time remote monitoring system over a large spatial footprint, revealing Adélie penguin responses to climate change. 

As a species that lives at the land/water interface, the American bullfrog (Rana catesbeianus) serve as a bioindicator in many habitats, yet also invasive in many locations. Due to challenges with traditional monitoring approaches, there is a lack of fine-scale population and phenological data for bullfrogs. Passive acoustic monitoring (PAM) can provide a low-cost alternative with high-resolution data for monitoring vocal animals. Sexually mature male bullfrogs attract mates by calling from exclusive territories. These vocalizations can be used to explore bullfrog behavior, population size, and phenology. We describe the analysis framework and initial results from an project monitoring the vocal behavior of frogs in 25 ponds in southeastern New Hampshire during the reproductive season using acoustic arrays. By using an acoustic energy index (RMS amplitude), we can estimate numbers of frogs in ponds, determine timing of reproduction, and even document anthropogenic disturbance. Our results can lead to future uses of PAM to monitor population size and phenology and develop reliable long-term management and conservation strategies. 

Migratory seabirds are vulnerable to decline due to climate change and anthropogenic disturbances. Common terns (Sterna hirundo) are highly vocal colonial seabirds that serve as bioindicators of their foraging grounds throughout their migratory range. Historically, monitoring colonial seabirds is invasive and time-consuming, and traditional acoustic approaches are complicated by high amounts of call overlap. Monitoring the behavioral ramifications of disturbance, as well as overall colony size and health, is crucial to implementing effective management decisions. However, methods are needed to do so efficiently and with minimal disturbance. In this study, we demonstrate that population size, demographics, and behavior can be assessed acoustically through changes in acoustic energy across varying temporal scales. To do this, we compared acoustic energy to in-person observations of nest density, chick-hatching, and investigator disturbance. We found that trends in acoustic energy align with observations of nest density, and the distribution of acoustic energy across frequency bands is indicative of colony demographics. Furthermore, we found a significant relationship between acoustic energy and investigator disturbance within 20 meters of an acoustic recorder. Overall, our findings suggest that colony-wide trends in population size, demographics, and behavior can be monitored via acoustic energy without the time-consuming analysis of individual calls. 

Acoustic indices are an efficient method for monitoring dense aggregations of vocal animals but require understanding the acoustic ecology of the species under examination. The present understanding of avian behavior and vocal development is primarily derived from the research of songbirds (Passeriformes). However, given that behavior and environment can differ greatly among bird orders, passerine birdsong may be insufficient to define the vocal ontogeny of non-passerine birds. Like many colonial nesting seabirds, the Adélie penguin (Pygoscelis adeliae) is adapted to loud and congested environments with limited cues to identify kinship within aggregations of conspecifics. In addition to physical or geographical cues to identify offspring, adult P. adeliae rely on vocal modulation. Numerous studies have been conducted on mutual vocal modulations in mature P. adeliae, but limited research has explored the vocal repertoire of the chicks and how their vocalizations evolve over time. Using the deep learning-based system, DeepSqueak, this study characterized the vocal ontogeny of P. adeliae chicks in the West Antarctic Peninsula to aid in autonomously tracking their age. Understanding the phenological communication patterns of vocal-dependent seabirds can help measure the impact of climate change on this indicator species through non-invasive methods. 

Accurate population counts are essential for understanding the status of species and for researchers studying various phenomena including monitoring the relationship between environmental stresses and the spread of disease within populations. Both small roosts and large colonies of bats provide challenges when attempting to determine an accurate population count. Recently, there have been a number of new video analysis software applications, that are available on the internet, which can be used to provide population counts. When software-based counts are compared with manual counts, the software provides counts that are substantially less labor intensive, determined substantially more quickly, and have the potential to be more accurate. This short paper discusses the use of neural networks to determine the number of bats that there are in a region when multiple bats may overlap. The work discussed in this manuscript demonstrates that the counts of multiple overlapping bats can be improved using trained neural networks. This is a critical improvement for providing accurate counts in high density videos. This manuscript contains the biological motivations, and a brief overview of how artificial intelligence is being implemented. The results discussed compare the accuracy values of neural networks for a few case studies including cross-comparisons of data trained on different video types and for different animals which can have accuracy values above 90 % for comparable video types. Finally, the generation and use of synthetic images, to increase the amount of data in a training set, is also discussed, which resulted in a trained neural network that produced an accuracy value of 80% on 12 unbiased categories. 

One of the biggest challenges with species conservation is collecting accurate and efficient information on population sizes, especially from species that are difficult to count. Bats worldwide are declining due to disease, habitat destruction, and climate change, and many species lack reliable population information to guide management decisions. Current approaches for estimating population sizes of bats in densely occupied colonies are time-intensive, may negatively impact the population due to disturbance, and/or have low accuracy. Research-based video tracking options are rarely used by conservation or management agencies for animal counting due to the perceived training and elevated operating costs. In this paper, we present BatCount, a free software program created in direct consultation with end-users designed to automatically count bats emerging from cave roosts (historical populations 20,000–250,000) with a streamlined and user-friendly interface. We report on the software package and provide performance metrics for different recording habitat conditions. Our analysis demonstrates that BatCount is an efficient and reliable option for counting bats in flight, with performance hundreds of times faster than manual counting, and has important implications for range- and species-wide population monitoring. Furthermore, this software can be extended to count any organisms moving across a camera including birds, mammals, fish or insects.