Listening to the Forest with Bioacoustics
Observing the health of ecosystems is important as it helps with species preservation, population monitoring, stress detection, and climate monitoring. In fact, “biodiversity loss is ranked as one of the top five global risks, both in impact and likelihood” (European Commission). Keeping an eye and or ear on an ecosystem over time can be hard for a person to do without any tools as there are many different sounds from animals coming from different directions, overlapping, etc.
Bioacoustics are one of the tools that allow scientists to get a comprehensive view of an ecosystem's health by capturing and analyzing the sounds produced by animals and environmental sources. So how do they work?
Bioacoustics utilizes a plethora of different technologies, including sound detection and recording, data processing, and sound analysis using AI algorithms and spectrograms (visual representation of sounds). With this, scientists can identify what species are in an area and monitor their activity. Monitoring activity doesn’t just mean making sure all is well with the birds; it can provide greater insight into how these species interact with each other, species outside of theirs, and how the environment is affecting them.
For the business guru, “bioacoustics may also be more efficient and cost-effective than traditional methods” (Teixeira et al. 2019). It’s argued that studying bioacoustics will allow us to gain a better understanding and insight into behavioral states and “contexts that provide insight into populations as it relates to their conservation” (Teixeira et. All).
While detecting and recording sound is self-explanatory as microphones are used, the software side isn’t so easy. Many research groups are working to produce their own way of recording, analyzing, and processing audio. For example, the Yang Center, which is a leader in the development of bioacoustics software, offers a wide range of technologies such as file management (“compressing, converting, and renaming audio files”), integration of “high-performance computing technologies and bioacoustics data-mining capabilities”, and “hundreds of years of recordings of the natural world” (The Cornell Lab).
Another group, OpenSoundscape produced “an open-source bioacoustics analysis package for Python” which aims to “provide a robust and open-source Python toolkit for detecting and localizing biological sounds in acoustic data” (Lapp et al. 2023). OpenSoundscape utilizes PyTorch, an open-source machine learning library, to provide machine learning tools for “locating biological sounds of interest in space and time” (Lapp et al. 2023).
Companies such as Microsoft are working on environmental projects such as “zero-shot transfer for wildlife bioacoustics, CLAP bioacoustics annotation tool, and using bioacoustics for multi-species classification” (Microsoft Research Forum, 2024).
Machine learning methods can efficiently and accurately distinguish between relevant biological sounds and background noise, detect different species, differences in a species' vocalization, frequency, amplitude, and temporal patterns, and more. In the case of OpenSoundscape, they provide “access to powerful acoustic detection, classification, and localization methods including both machine learning and signal processing algorithms” (Lapp et. All).
Convolutional Neural Networks is a machine learning technique used for image and sound pattern recognition. It assists with the analysis of spectrograms to identify species’ calls and classify different types of vocalizations. Other techniques include but are not limited to Support Vector Machines (for classification tasks), K-Means Clustering (classification of similar sounds), and Recurrent Neural Networks and Long Short-Term Memory Networks (for handling sequential data, making them ideal for analyzing temporal patterns in recordings). Machine learning has given scientists a way to understand what these sounds are telling about their direct ecosystem and the overall environment.
Although many may believe that Bioacoustics can only be applied to terrestrial activity, this isn’t the case. Bioacoustics can also be used in the ocean “to support terrestrial, aquatic, and marine passive-acoustic monitoring (PAM) efforts and acoustic studies of behavioral ecology, evolution, and conservation biology” (The Cornell Lab). In this case, recording sound is a difficult problem as you need to make sure the equipment stays dry while still being able to get recordings.
The Cornell Lab, for example, has created “Rockhopper acoustic recording units… to enable scientists to study any acoustically active marine species of interest” (The Cornell Lab). This device can sample at 394 kHz which allows the unit to record a wide range of marine species. These units are small enough “to be deployed by a single individual using a small vessel, which provides flexibility in field operation, cost saving, and efficiency in deployment/recovery process” with the first being deployed in the Gulf of Mexico in May 2018 (The Cornell lab).
Rockhopper unit(s) – The Cornell Lab
Continuing to monitor our ecosystems is something we should all participate in, even if it’s just watching what types of animals are in your backyard. Bioacoustics allows scientists to get a better understanding not just of which species are in an area, but how they interact with themselves, and others, and how factors such as climate are changing their lives. With machine learning and AI becoming more relevant by the day, technology will continue to be utilized to observe our planet and its complex systems.
Works Cited
Bioacoustic AI for wildlife protection. European Commision. (n.d.). o https://cordis.europa.eu/project/id/101116715
Bioacoustics. Microsoft Research. (2024, April 24). https://www.microsoft.com/en-us/research/project/bioacoustics/
Lapp, S., Rhinehart, T., Freeland-Haynes, L. F.-H., Khilnani, J., Syunkova, A., & Kitzes, J. (2023, August 9). Wiley Online Library | Scientific Research Articles, journals, books, and reference works. British Ecological Society. https://onlinelibrary.wiley.com/
The Cornell Lab, Our mission:. K Lisa Yang Center for Conservation Bioacoustics. (n.d.). https://www.birds.cornell.edu/ccb/
The Cornell Lab, Rockhopper – Marine Passive Acoustic Recording Unit. K Lisa Yang Center for Conservation Bioacoustics. (n.d.). https://www.birds.cornell.edu/ccb/rockhopper-unit/
Teixeira, D., Maron, M., & van Rensburg, B. J. (2019, June 11). Wiley Online Library | Scientific Research Articles, journals, books, and reference works. Society for Conservation Biology. https://onlinelibrary.wiley.com/
