New computer software can recognize individuals with 100% accuracyWolves are always controversial, but as top predators in many ecosystems, they play an important role that shouldn't be underestimated (like sharks in the sea, another controversial species that needs our help). There's an inuit saying that goes something like: "The wolf keeps the caribou strong."
Conservationists who study wolves to better understand and protect them need to follow wolves around, but that's not easy. The most reliable way is to capture a wolf and fit it with a GPS collar, but that's not easy to do and costs a lot of money. Or they can look for wolf tracks in the snow (when there's snow), but that's also definitely not ideal...
Holly Root-Gutteridge of Nottingham Trent University has developed sound analysis code that can tell which wolf is howling with 100% accuracy. Previously, pitch was used to tell wolves apart, but these only achieved an accuracy rate of 76%. Adding information about volume – or “amplitude” – to the code was key to increasing its accuracy, according to Root-Gutteridge. [...]
“Acoustic sensing is becoming much more popular in conservation, and it’s easy to see why,” Tallents added. Methods like radio collaring and ear tags require the wolves to be caught, which carries a small risk of injury and can cause disruption to the pack’s activities. “Being able to identify individuals aurally would be really useful to us, and cause much less disturbance to the wolves themselves.”
The method can never give a definitive count though. “We can only ever detect a minimum number of wolves in a pack: if they’re not howling, we won’t know they’re there,“ Root-Gutteridge said. "But the real revelation here is that we can now identify individual wolves.” (source)
Let's hope that this non-intrusive software will get in the hands of conservationists in the field quickly.