For decades, the roar of a lion was considered one familiar sound. But hidden within that iconic call is a second, long-overlooked “in-between roar” that reveals a more complex vocal landscape than anticipated.
Results published in Ecology and evolutionThis is the first time artificial intelligence has been used to automatically classify types of roar, a task that previously relied on skilled listeners.
“Lion roars aren't just iconic—they are unique traits that can be used to estimate population sizes and monitor individual animals. Until now, the identification of this roar has relied heavily on expert judgment, introducing potential human bias. Our new AI approach promises more accurate and less subjective monitoring, which is critical for conservationists working to protect dwindling lion populations,” lead author Jonathan Growcott said in the paper. press release.
Why the lion's roar matters more now
Lion Africa's population has plummeted, with an estimated 20,000 to 25,000 people remaining – about half the population of a quarter-century ago. The International Union for Conservation of Nature now lists the species as vulnerable to extinction, reflecting ongoing habitat loss, conflict with humans and shrinking ranges across the continent.
Lions roar not just to be heard—their calls help defend territory and maintain communication between pride members across the landscape.
As passive acoustic monitoring expands into large protected areas, this far-reaching roar becomes an increasingly valuable source of information. Because each call can be transmitted over long distances and contains unique acoustic features. individual lionsResearchers see vocalizations as a promising, low-cost way to detect and count animals—as long as the calls can be accurately classified.
Read more: Lion breaks record for swimming in waters infested with hippos and crocodiles
Revealing the second lion's roar
To limit the bias that can arise from human interpretation, the team created a statistical model to separate loud roars from other sounds during a lion's roar. The model confirmed that lions produce two different types of roars and sort them correctly about 85 percent of the time, supporting the idea that the intermediate roar is a distinct sound and not just a variation of a familiar one.
The second approach simplified the process even more. By focusing on just two key characteristics of a lion's call—the length and pitch of its roar—the system achieved 95.4 percent accuracy in sorting out the two types of roars. This simplified method also made it easier to distinguish individual lions, surpassing the accuracy of experienced listeners.
The results show that automated tools can outperform expert judgment while significantly reducing effort and bias. And because the workflow is intentionally simple, the authors say it can be easily implemented in conservation settings where resources are limited. This approach also reflects advances in bioacoustics for other large carnivores, including spotted hyenas.
A new way to monitor lions
Taken together, the findings indicate a major shift in how lions can be observed in the wild. By combining simple acoustic measurements with automated classification, the approach could ultimately change how conservation teams track individual lions and estimate population sizes across vast landscapes.
“We believe there is a need for a paradigm shift in wildlife monitoring and a large-scale shift to the use of passive acoustic techniques. As bioacoustics improves, they will be vital for effective conservation of lions and other endangered species,” Growcott said in a press release.
Read more: Neanderthals killed the cave lion 48,000 years ago, and scientists know how
Article sources
Our authors in discovermagazine.com use peer-reviewed research and high-quality sources for our articles, and our editors review scientific accuracy and editorial standards. Review the sources used below for this article:
