Passive acoustic monitoring and deep learning reveal spatiotemporal patterns in gibbon calling behaviour associated with habitat and climate variables
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
1. Understanding the basic ecology of endangered species is essential for effective conservation, yet this remains challenging for elusive species inhabiting tropical forests. For the endangered Bornean white-bearded gibbon (Hylobates albibarbis), basic ecological information remains limited. Most research on the species is restricted to peat swamp forests, while little is known from other forest types that make up a large part of its range. Passive acoustic monitoring provides an opportunity to study vocal behaviour to obtain such ecological insights, while enabling research across larger spatial and temporal scales than previously possible. 2. We deployed eight autonomous recording units across three forest types in Central Kalimantan, Indonesia, collecting 23,244 hours of acoustic data over 18 months. A pretrained deep learning automated detector was applied to identify great calls, performed by female gibbons as part of morning duets and a key indicator for comparing population density. We identified 83,956 great calls and examined how daily call rates varied across habitats and in response to seasonal rainfall (as an indicator of resource fluctuations). 3. Daily call rates did not differ significantly among forest types but showed significant temporal variation over the survey period. Higher call rates occurred during months with greater rainfall, consistent with seasonal resource availability driving vocal activity. To clarify this effect, we investigated the short-term effects of weather and found that rainfall on the day before observation reduced both call rates and the probability of calling, while rainfall two days prior increased calling activity, suggesting compensatory vocal behaviour. 4. Our findings highlight the need to account for variable vocalisation rates in acoustic monitoring, particularly when evaluating the additive effects of habitat loss and climate change on species behaviour and ecology. We emphasise the importance of considering phenological factors when interpreting calling activity and the value of incorporating spatial data to strengthen ecological inferences from acoustic datasets. Furthermore, this study demonstrates the power of deep learning for large-scale, long-term monitoring of species’ vocal behaviour, providing valuable ecological insights across increasingly broad spatiotemporal scales.