Gunshot recorded in the field.

Illegal logging and hunting severely threaten the tropical forests of the Yucatán Peninsula. These activities often remain undetected because of their cryptic nature and a lack of funding to hire personnel for patrolling the forests.

Effective monitoring will depend on developing efficient and low-cost technology, capable of providing sufficient ground-truthing information about human agents of threat to make a difference to forest conservation in underfunded reserves in the region. Listening devices are starting to be deployed in other countries (‘Elephant Listening Project’,‘Rainforest Connection’), using solar-powered devices connected to a web-cloud. The open tree structure of Yucatán dry forests, and the absence of a reliable wi-fi signal, demands smaller and less visible, intra-networked devices.

With this Small Rufford Grant I will record high-fidelity soundscapes of tropical forests in the Yucatán Peninsula, in order to develop a methodology for effectively categorising sounds emitted by human-agents of threat in relation to local wildlife activity. Controlled and ground-truthed recordings will prepare for subsequent development of cost-effective tools for classifying human exploitation and biodiversity in forests. My long-term vision is to develop an automated system for real-time detection and location, which lessens the demand on trained experts of forest disturbance to patrol forest tracts on foot, and consequently reduces stakeholder monitoring costs.

The grant will fund the purchase of three SM3+ acoustic devices, designed for monitoring wildlife in field conditions. These will be deployed in El Edén Ecological Reserve, Mexico, and in Cockscomb Basin Wildlife Sanctuary and the Belize Zoo, Belize. I will use them to pilot the collation of a soundscape of human disturbance and wildlife activities (e.g. gunshots, dogs, chainsaws and heavy machinery), experimenting with optimal positioning and concealment within the forest strata to maximise detection of anthropogenic disturbance in relation to wildlife activity (e.g. calls of felids and their prey). The simultaneous deployment of three devices will allow me to experiment with triangulating the source of a sound. The practicalities of locating sources will be tested by deployment across a range of geographies and separation distances between devices. The soundscape will be brought back to the acoustic laboratory at the University of Southampton for analysis of acoustic signals of disturbance and biodiversity, to develop analytical algorithms in the lab. These algorithms will be designed to isolate the sounds of disturbance activities from natural sounds of the forest environment. They will drive subsequent development of software for eventual application to low-cost listening devices capable of triggering recordings and alert signals.