Hear me roar! Each Leo has its own vocal pattern that people can use to recognize one another and help scientists track population movements
Each lion has its own roar, which allows the "kings of the jungle" to recognize one another and can be used to track population movements, according to a study.
Researchers from Oxford used machine learning to analyze the roar of different lions and selected the frequency with which they can be distinguished from one another.
According to the experts, lion calls are usually issued in a set – with a low moan or two, followed by several loud, full-throated roars and a grunt.
Previous research found that lions can distinguish the roar of their colleagues from one another to identify distant friends and enemy neighbors.
However, it was not previously clear what aspects of the call structure would enable them to distinguish between those made by different people.
Scroll down for video
Each lion has its own roar, which enables the "kings of the jungle" to recognize each other and can be used to track population movements, as a study has found (archive image).
"The numbers of African lions are declining and developing inexpensive tools to monitor and ultimately better protect the population. This is a conservation priority," said Oxford University paper writer and zoologist Andrew Loveridge.
"The ability to remotely assess the number of individual lions in a population based on their roar could revolutionize the way lion populations are assessed."
In their study, the researchers developed a wearable device they called a "biologger". When attached to a lion's GPS collar, both audio recordings and movement data can be recorded.
From this data, the team was able to link sounds and locations and associate each lion with its own characteristic roar.
Next, the team trained a pattern recognition algorithm to “learn” the characteristics of each lion's roar and then put it through its paces to analyze voice sequences that it had not “heard” before.
The researchers found that the algorithm can assign a record of a roar to a particular lion with an accuracy of 91.5 percent.
The software found that the key to distinguishing between each Lion's calls was to look at the changing shape of the fundamental frequency, or the lowest frequency, of one of the full-throat roars that are usually produced in the middle of a series of vocalizations.
These shapes, the team said, share traits in all lion calls – but they also differ so widely between individuals that lions can likely spot the subtle differences to differentiate themselves from one another.
The software found that the key to distinguishing between each Leo’s calls was looking at the shape of the fundamental frequency – or & # 39; ƒ0& # 39; highlighted on a spectrogram – – from one of the full-throated roars typically made in the middle of a series of vocalizations
"Distinguishing between individual roars (…) could facilitate the development of alternative techniques for assessing population density and tracking individual movements in the landscape," added paper author Andrew Markham.
After completing their first study, the researchers are now attempting to perform playback experiments with modified roar – with the aim of determining whether the frequency of the call alone contains enough information to enable speech recognition.
The full results of the study were published in the journal Bioacoustics.
(tagsToTranslate) dailymail (t) sciencetech