London:
Scientists have identified a new type of lion roar, which they have named the “intermediate roar.” This newly documented vocalisation is shorter, lower in pitch, and less intense compared to the lion’s well-known full-throated roar. According to researchers, this intermediate roar always appears immediately after a full-strength roar, revealing that lion communication is far more complex than earlier understood.

Jonathan Grobekote, a PhD researcher in mathematics and statistics at the University of Exeter, and the lead author of the study, explained that a full roar is “a burst of sound—loud, complex, with rising and falling tones.” In contrast, the intermediate roar is “lower-pitched and shows far less variation.”

The findings, published in the journal Ecology and Evolution, challenge the long-held belief that lions have only a single recognisable type of roar. Lion roars serve as signals for communication within the pride and for marking territorial boundaries.

Lion Roars Contain a ‘Signature’

Previous studies have shown that a lion’s full-strength roar contains a unique acoustic signature that can reveal the animal’s sex, age, and other characteristics. These personalised roars allow individual lions to be identified based solely on vocal patterns.

Grobekote says the newly collected audio data could help scientists estimate lion populations more accurately, using sound analysis rather than visual surveys.

Massive AI-Driven Acoustic Study

For the study, researchers installed 50 microphones across Tanzania’s Nyerere National Park and fitted acoustic collars on five lions in Zimbabwe’s Bubi Valley Conservancy.
In total, the team recorded 3,149 lion vocalisations. Artificial intelligence was then used to analyse and classify the sounds, leading to the discovery of the intermediate roar.

The AI model successfully classified lion calls with over 95% accuracy, significantly reducing human bias. Grobekote noted that distinguishing between full roars and identifying individual lions was previously “highly subjective,” a problem now solved by machine learning.

Aside from roars, lions also produce grunts, snarls, chuffs, moans, and mews, but only the full-strength roars have been proven to contain identifiable signatures. The new research now adds the intermediate roar to the scientific understanding of lion communication.

Researchers believe this discovery will help scientists better differentiate full roars in the future and improve wildlife monitoring across Africa.