Acoustic Monitoring for Biodiversity Assessment

Acoustic monitoring harnesses environmental sound recordings to survey wildlife communities at scales and resolutions unattainable by traditional field methods. Deploying networks of autonomous recording units across terrestrial and marine habitats enables continuous, non-invasive sampling of vocalising taxa, from birds and mammals to insects and amphibians. Raw audio is transformed into interpretable metrics through soundscape analysis, acoustic indices and machine-learning algorithms, which together estimate species presence, community composition and temporal dynamics. Advances in low-cost open-source hardware and cloud-based processing have democratised access to passive acoustic monitoring, facilitating long-term, large-scale studies of ecosystem change, restoration success and conservation interventions. Despite these strengths, challenges persist in standardising survey protocols, building comprehensive reference libraries, mitigating ambient and anthropogenic noise and ensuring transferability of analytical models across biomes. Integrating acoustic data with complementary methods—such as visual surveys or molecular assays—yields robust, multidimensional biodiversity appraisals. Globally, acoustic monitoring informs policy and management by tracking habitat degradation, assessing protected-area effectiveness and revealing phenological shifts under climate change.

Research from Nature Portfolio

Recent studies have deployed soundscape analysis coupled with deep learning to quantify biodiversity recovery in tropical forests, demonstrating that automated acoustic indices and convolutional neural networks can track community composition along restoration gradients, including elusive nocturnal insects. Another line of research has critically examined the generalisability of soundscape metrics across ecosystems. Large-scale comparisons of manually annotated bird counts and passive audio recordings reveal that common acoustic indices and machine-learning models often fail to predict species richness consistently, yet reliably capture community turnover, emphasising the need for integrated survey approaches that combine acoustic monitoring with traditional field methods.

Acoustic Monitoring for Biodiversity Assessment publication trend

The graph below shows the total number of articles in acoustic monitoring for biodiversity assessment across all publications each year (not limited to Nature Index journals).

Technical terms

Passive acoustic monitoring (PAM): Recording wildlife sounds continuously or at preset intervals to study biodiversity non-invasively.

Soundscape: The composite of all biotic, abiotic and anthropogenic sounds in a given environment.

Acoustic index: A quantitative metric summarising aspects of audio recordings, such as complexity or diversity of sounds.

Convolutional neural network (CNN): A deep-learning architecture effective in recognising patterns within spectrogram images of sound.

Autonomous recording unit (ARU): A standalone device programmed to capture environmental audio without direct human operation.

Spectrogram: A time-frequency representation of an audio signal, displaying amplitude across frequencies over time.

References

1. Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests. Nature Communications (2023).
2. Limits to the accurate and generalizable use of soundscapes to monitor biodiversity. Nature Ecology & Evolution (2023).
3. Long-Range Bird Species Identification Using Directional Microphones and CNNs. Machine Learning and Knowledge Extraction (2024).
4. Large-scale avian vocalization detection delivers reliable global biodiversity insights. Proceedings of the National Academy of Sciences of the United States of America (2024).
5. BirdNET: A deep learning solution for avian diversity monitoring. Ecological Informatics (2021).
6. AudioMoth: Evaluation of a smart open acoustic device for monitoring biodiversity and the environment. Methods in Ecology and Evolution (2018).