Publications

A call to scale up biodiversity monitoring from idiosyncratic, small-scale programmes to coordinated, comprehensive and continuous monitoring across large scales

Summary: Conservation managers can't protect what they don’t know about, but our current ways of tracking biodiversity are random and small-scale. Australia has committed to create a national biodiversity monitoring programme. This has not yet occurred despite the urgent need to monitor common and threatened species, as highlighted by the challenges of determining the biodiversity impacts of the Black Summer fires of 2019/20. With new advancements in automation, smaller devices, and better power sources, the world needs to quickly expand biodiversity monitoring to be organized, thorough, and continuous on a large scale. We propose the BIOMON project to achieve this by using individual sensor devices equipped with machine learning models to identify biodiversity through sound and/or photos. Devices would be set up in networks that send the results back to researchers, who analyze the data and make it available to the public for free. These networks could cover entire continents to measure changes in biodiversity. No one has achieved this yet, and there are still big challenges like training the algorithms, having enough cellular network coverage, balancing sensor power and memory, and deciding where to place the sensors. There's a lot of work to do, but in the 21st century, we can't achieve big goals unless we start working toward them. Hayward et al., 2022. Australian Zoologist

Harnessing edge computing and citizen science: A new prototype design for continental-scale acoustic monitoring in Australia and beyond

Summary: Inadequate monitoring of biodiversity is a characteristic of conservation the world over. The potential of acoustic monitoring is compelling, although the challenges remain substantial. Effective solutions require transdisciplinary collaboration among stakeholders, a focus on open-source development, and flexible, multi-pronged technical approaches. The potential to harness the power of citizen science is immense. Here, we present the first open-source, modular, expandable multitaxon recording unit and workflow pipeline that integrates recent advances in edge processing, network connectivity, and citizen science into a single system, called BioMon. The edge computing unit uses multiple on-board artificial intelligence (AI) models to identify environmental sounds before using the mobile phone network to send the detection audio clips to both a central data repository where summary statistics are presented and to a citizen science platform for validation. Field testing revealed that the system classified bird and frog calls in real time and stimulated high levels of citizen engagement in the verification process, providing a proof-of-concept prototype. Technical design features and observations relating to field deployment and levels of citizen engagement are provided. With potential to be expanded to other detection modalities (e.g. images, pollen, bushfire smoke), BioMon provides a highly effective platform for improving AI species classifiers, as well as a viable model for large-scale, real-time acoustic monitoring of biodiversity. We discuss several existing cooperative initiatives that could be emulated. Griffin et al., 2025. Methods in Ecology and Evolution