The lions roar, and so far nothing strange: it is a proverbial verse, one of the most famous nature verses in the world. Until now, however, we thought that their roar was monotonous, that there was only one version of it. A study by the University of Exeter published in Ecology and Evolution However, it now shows that we underestimated the variety of lions’ roars: there are two different versions, with functions that are, in all likelihood, equally different.
The discovery is not just a curiosity, but could revolutionize the way we approach their conservation.
A helping hand from AI. Artificial Intelligence, in this case, lends a hand to science: the study on lions is a demonstration of this. The roars of these large cats, in fact, are the equivalent of our fingerprints: they allow us to identify single individuals, but also to estimate the size of a population by studying the variety of roars they emit. The problem is that until now the identification of roars was entrusted to the ears of experts.
Bionic ear. Instead, the University of Exeter team trained an AI machine learning algorithm to distinguish individual roars. The results are more than promising: the AI correctly identified individuals with an accuracy of 95.4%, much higher than what a human can achieve. And this is the first advantage of using AI: it is possible to monitor a population simply by recording its roars and, so to speak, feeding them to the machine.
Two different roars, but why? However, the study also made it possible to identify the components of the lions’ roar more precisely. Which is not just the classic “full-throated roar”: it also includes moans and grunts, and above all a new category of roar, which has been named “intermediate roar”. It is a weakened version of the full roar: it lasts less, covers fewer frequencies and… well, its function is not yet clear, given that the study focused only on identification and classification.
Listen to nature. Knowing how to distinguish not only full roars but also intermediate ones will allow individual lions to be identified with even greater precision: according to the authors of the study, bioacoustics and sound and noise monitoring techniques should become central in conservation activities, because they are more effective methods than classic camera traps or the analysis of tracks and excrement.
In short, to save nature we must learn to listen to it better.
