They say you can identify the leopard by its spots, but as it turns out, you can also identify it through its unique roar. Leopards, notoriously difficult to monitor due to their elusive nature, could soon be tracked using passive acoustic recorders paired with AI-based analysis.
Leopard populations face significant challenges, with their ranges shrinking by 80% due to poaching, habitat loss, and urbanization. Classified as “vulnerable” by the International Union for Conservation of Nature (IUCN), leopards are in dire need of effective conservation strategies. Traditional tools like camera traps and GPS collars are useful but limited, particularly for a species known for its stealth and nocturnal behavior.
This is where passive acoustic monitoring comes in. The idea is to track individual leopards by using their unique roars.
Leopards don’t make spectacular roars like lions. They produce repeated low-frequency patterns, often audible for over a kilometer. These roars are primarily used to attract mates and to warn intruders that they are trespassing.
The setup was deployed across a 450-square-kilometer area in Nyerere National Park in Tanzania, combining camera traps with autonomous recording units (ARUs). The goal was to determine if individual leopards could be identified by their roars, similar to lions and tigers, whose vocalizations are known to carry unique signatures.
The setup included 50 paired stations, each equipped with ARUs that continuously recorded sound and camera traps that captured images when triggered. Over 62 days, the team collected over 191 leopard events and matched roars to individual leopards by comparing images to acoustic data. Researchers used software to visually analyze sound waves, isolating the second part of leopard roars, which carry the clearest individual markers. Then, using advanced analytical techniques, including Gaussian Hidden Markov Models, they achieved a remarkable 93.1% accuracy in distinguishing individual leopards by their roars.
“Discovering that leopards have unique roars is an important but fundamentally quite basic finding that shows how little we know about leopards, and large carnivores in general,” said lead author Jonathan Growcott, a PhD student at the University of Exeter.
“We hope it will allow leopards to become the focus of more acoustically complex science such as population density studies and open the door to more work on how large carnivores use vocalizations as a tool.
This could be scaled up, but it’s not easy
The implications of this research extend beyond leopards. The success of pairing ARUs with camera traps could transform monitoring for other elusive or nocturnal species. From tracking bird migrations to studying amphibian populations, the potential applications of bioacoustics are vast. The study also opens avenues for understanding how vocalizations vary with environmental factors or threats from competitors like lions and humans.
Despite its promise, the study also highlights difficulties in deploying this technology. ARUs require significant data storage and analysis, with the research team processing 72,000 hours of audio. Developing automated species classifiers could streamline this process, reducing reliance on manual analysis.
Another problem is the number of leopards used in the study: just 50 leopards from one area in Tanzania. Would the findings be carried out for more animals in different geographical areas? Many mammals are known to have different “accents” and “dialects”, and it’s unclear whether the same algorithmic approach would work in different scenarios.
Yet researchers are confident that the new technology could be used alongside the existing approaches.
“Even though additional equipment, data management and analytical expertise are required, paired sureveys are still a promising monitoring methodology which can exploit a wider variety of species traits, to monitor and inform species conservation more efficiently, than single technology studies alone,” the scientists note in the study.
The study “The secret acoustic world of leopards: A paired camera trap and bioacoustics survey facilitates the individual identification of leopards via their roars” was published in the journal Remote Sensing for Ecology and Conservation.
