Could a New “Voice Recognition” Software Program Help Protect Wolves?

Candice Gaukel Andrews August 20, 2013 11

A new computer program makes distinguishing individual wolves by their howls highly accurate. Could this help save their legal protections? ©Henry H. Holdsworth

Seeing a wolf in the wild is one of those lump-in-your-throat moments that you never forget. One of my first such experiences was in the Lamar Valley during a trip to Yellowstone National Park. One particular wolf I watched during that tour still stands out in my mind. On a late afternoon, scanning the valley with spotting scopes and binoculars, my fellow travelers and I caught sight of a wolf that had been collared earlier in the day. When we first saw him, he was lying down with his head up, still woozy from the tranquilizer. After a while, he slowly stood up, on wobbly legs. We watched him until it got dark, and we couldn’t see him anymore. But we knew he was still there, coming back into himself.

According to a recent report published in Bioacoustics: The International Journal of Animal Sound and Its Recording, soon wild wolves may not have to endure such stressful experiences. A team from Nottingham Trent University in the United Kingdom has just developed a computer program that can analyze the vocal signatures of individual gray wolves with 97.4 percent accuracy.

Could this advanced technology provide a new and less disruptive method for monitoring wild animals, and—better yet—help stop the loss of protections for our nation’s wolves?

I know that voice

Wolf tracking based on paw prints works—but only when it snows. ©Ben Forbes

Howling by wolves serves several functions: to form social bonds, to locate members of their own packs by voice recognition, to establish territorial positions and to discover dominance rankings.

While similar technology to the new computer program had been tested on captive wolves and their howls, this study is the first time such accurate results have been achieved from recordings of wolf howls taken in the wild, where varying conditions make recognition considerably more difficult. In the past, scientists using audio sampling to identify wild wolves had been able to achieve an accuracy rate of only 76 percent.

What’s different about this new computer program is that it not only analyzes the pitch of wolf howls but their amplitude (or acoustic energy). By adding in the detail about vocal intensity, it’s easier to distinguish individuals from one another. What’s more, the technology is able to scrutinize howl recordings and throw out extra, unneeded noises—such as that from wind and water—that might otherwise contaminate the data.

Scientists tested this new tool by studying dozens of archival recordings of howls taken from wild gray wolves that were living mainly in Algonquin Provincial Park in Ontario, Canada. Their success rate was 100 percent when recognizing individual wolves from their solo howls and 97 percent when identifying wolves calling together in a chorus howl.

Having that voice heard

Before European settlement, gray wolves were common throughout Alaska, Canada and the continental United States. By the early 1960s, however, wolves nearly went extinct, all but eradicated by humans. After the killing stopped, only about 300 gray wolves remained, keeping far away from us in the deep woods of upper Michigan and Minnesota.

Because of protection under the Endangered Species Act, gray wolves have come back from the brink and are considered one of the biggest conservation success stories in U.S. history. Though nowhere near the historical estimate of more than 400,000 in the United States, now as many as 5,000 gray wolves live in Michigan, Minnesota and Wisconsin, with another 7,000 in Alaska. Smaller numbers of reintroduced wolves live in Idaho, Montana and Wyoming.

Better monitoring may help keep gray wolves on the endangered species list. ©Henry H. Holdsworth

However, monitoring wolf populations, which remains a vital part of management, has always been a labor-intensive and inexact science. GPS collaring pinpoints where an individual is, but relates nothing about with whom an animal is communicating or spending its time. Plus, collars are expensive and require capturing a wolf first—a huge and stressful undertaking for all involved.

Now that this new computer program has been shown to be successful, researchers see it as a tool to help conserve wolves in their natural habitats. Better monitoring could help many conservationists make their case for keeping gray wolves on the endangered species list. The technology could also be put to use with other canids, such as African wild dogs and Ethiopian wolves, both of which are now endangered.

Do you think that this more efficient and accurate method of identifying and marking the whereabouts of individual wolves could help in preserving their federal protections? Or is it too little too late? In any event, is the new technology a better alternative than collaring animals?

Here’s to finding your true places and natural habitats,


The experience of watching the wolf in the Lamar Valley waking up is recounted in the Natural Habitat Adventures book, An Adventurous Nature: Tales from Natural Habitat Adventures, edited and written by Candice Gaukel Andrews.


  1. Patti Adkins August 26, 2013 at 8:29 am - Reply

    I sure hope so.

  2. Sophie Ståhlberg August 23, 2013 at 6:10 am - Reply

    I am doing my PhD programme in Italy where financial and practical obstacles has prevented capture and collaring of wolves, which would at the very least ease the mapping of the species dispersion northwards into the western Alps. This could be a wonderful opportunity for individual recognition in any rare and elusive predator research and conservation worldwide!

  3. Kirk Parker August 22, 2013 at 8:17 am - Reply

    This is amazing. I look forward to see where this technology takes the animal field, and how it can be implemented on so many different levels. It will be exciting to see how it will effect other research programs in the future.

  4. Dhona Lovick August 21, 2013 at 3:09 pm - Reply

    Interesting read. This certainly would be a less stressful monitoring method.

  5. Suzanne Stone August 21, 2013 at 1:41 pm - Reply

    Not sure how it may effect federal protections but it could offer a safer, cheaper means for monitoring wolves. Even when done correctly, trapping is hard on wolves and collars are expensive and burdensome.

  6. Pieter-Bas Stolte August 21, 2013 at 7:05 am - Reply

    Interesting development. Obvious other species where this could prove valuable in reducing human-animal conflict would be: Elephants, lions, leopards, etc. Also, this could be a great non-invasive way to gain more insight into the distribution and movement of any other species that uses long calls, like gibbons, orang-utans, chimpanzees, etc, etc.

  7. Donna Dannen August 21, 2013 at 7:04 am - Reply

    I would like to know more about this. I do fear this is way too little too late. Tracking wolves is one thing. Getting them back into the areas (National Parks) where they have total protection is quite another. I still maintain these animals should not be hunted. Yellowstone wolf populations are low at the moment. Last season’s hunt was the direct cause. We, as Americans, overwhelmingly supported a very expensive wolf reintroduction. It is very painful to see a handful of states profit while possibly undoing a very expensive endeavor. The hunting seasons never should have happened. Where is the democracy in this?

  8. Fatemah August 21, 2013 at 4:11 am - Reply

    This sounds like a positive step towards protecting the wolves, may God protect them and keep them safe.

  9. Kelvin August 20, 2013 at 7:36 pm - Reply

    That is really neat, I’m curious how they go about monitoring though. Collars for wolves are GPS oriented, but there are also Radio Telemetry ones that are not expensive. The benefit of these are that they can locate any one wolf at almost any point in time with a little man power. Where as I imagine the voice recognition devices would have to be placed along the borders of territories, and would only be useful when the wolves howl.

    The data, and lack of effort for a collar may be worth the cost of multiple units, data analysis, and survey hours required for these devices! However, for a set area like Yellowstone they might be very useful for monitoring residents. Dispersing individuals (like the ones recently found in Kentucky, and California would be a whole different story!!!)

    This would be cool to see more about!

  10. Phillip Tureck - FRGS August 20, 2013 at 10:37 am - Reply

    Have read about this in the last few days – seems a positive move forwards for wolves being analyzed. Let’s hope it saves many more of them.

  11. Travis August 20, 2013 at 10:09 am - Reply

    I wonder what’s measured? Maybe length, frequency (pitch), variation… Actually, probably just measuring waveform takes care of most of that?

Leave A Response »