In the US, AI helps study wolf voices in Yellowstone — Mongabay
In Yellowstone National Park in the United States, linguist and computer engineer Jeff Reed uses artificial intelligence and bioacoustic devices to study wolf vocalizations. The AI programs he created identified more than 8,000 hours of continuous wolf vocalizations from recordings of the park’s sound environment. According to Mongabay, this is the world’s largest database of this type.
Reed launched the Cry Wolf project in 2023 as part of the Yellowstone Wolf Project, a long-term wolf research program. About 50 automated sound-recording devices have been placed in remote areas of the park. They record the sound environment around the clock and collect about 500,000 hours of audio data each year.
Analysis of wolf signals
The AI models trained by Reed make it possible to convert individual wolf vocalizations into spectrograms, or graphs of sound frequencies. The researcher compares these data with the behavior of specific animals at the moment of howling, barking, or whining.
Reed does not describe his work as translating “wolf language” into human language. He describes his approach as “telling the story of a howl”: to come closer to understanding the meaning of a signal, it is necessary to track the behavior of a specific animal and the circumstances in which it made the sound. Through repeated acoustic analysis, he has learned to distinguish the voices of individual wolves.
More current news is available on the UA.News Telegram channel Telegram.
According to Reed, wolves have more than 20 types of vocalizations, including barks, whines, howls, squeaks, moans, and teeth clicking. One form of their communication is collective howling, when a pack howls simultaneously.
Non-invasive monitoring
Yellowstone Wolf Project leader Dan Stahler noted that traditional monitoring involves the use of helicopters, sedating wolves, and fitting them with collars. Bioacoustic devices may become a non-invasive research method: they record sounds and location data, making it possible to remotely track pack activity and territories.
Reed’s technology is also being used in northern Italy. The Osservatorio Lupi group in Fidenza purchased 12 sound-recording devices to detect wolves approaching towns or farms and deter them without using weapons. Bear researchers, ecosystem researchers, and wildlife crossing researchers are also interested in the technology.
