Tuesday 23 July 2013
Scientists unlock secret to identifying wild wolves from their howls
Experts have been able to identify individual wolves from just their howls
Scientists have for the first time been able to identify individual wolves from just their howls with an accuracy of 100%. A Nottingham Trent University team has developed a new code for improved sound analysis – making it possible to distinguish between vocal signatures of wild Eastern grey wolves by studying both the volume and pitch of their howls.
The research, being published in the journal Bioacoustics, could have wide-reaching implications for surveying and conserving wild wolves. Wolves have enormous home ranges of up to 1,000 square miles and can hunt as far as 30 miles in a single day, making it very difficult for scientists to track them.
They maintain their territorial boundaries through scent-marking, scratching and howling – with howling used for territorial defence, social bonding and contacting members of their own pack.
But now the researchers – based in the University's School of Animal, Rural and Environmental Sciences and School of Science and Technology – have developed a bespoke high-level computational code, specifically for the extraction of howls from recordings of wolves.
They have been able to substantially improve the unlocking of the vocal signatures of wolves by adding amplitude – or volume – information for the first time in their analysis, in addition to the traditional frequency, or pitch, features used in previous studies.
When analysing 67 archive recordings of solo howls from ten individual wild wolves the team was able to identify the correct individual every time. Furthermore, 112 chorus howls – whereby wild wolves howl in unity – involving 109 individuals, achieved an accuracy of more than 97%.
The previous accuracy rate using acoustic systems to identify individual wild wolves was just 76%.
The future identification of individuals using this method would give scientists the potential to produce accurate counts, which is seen as crucial in terms of surveying and conservation efforts. Current survey methods can involve researchers standing and howling and judging the numbers in a pack from the returned howls, which often results in miscounting, or counting individuals twice.
"We already knew that wolves, like humans, had distinct voices, and now we are able to identify them with 100% accuracy without ever even seeing them," said researcher Holly Root-Gutteridge, who is based in Nottingham Trent University's School of Animal, Rural and Environmental Sciences.
She said: "In scientific terms this is really exciting, because it means that if we hear a howl on night one we can tell if it is or isn't the same wolf that you hear on subsequent nights.
"For the first time we will be able to be sure which wolf is howling to us and track individuals across their territories using just their howls. This is much easier and cost-effective to do than other tracking methods such as GPS-collaring."
The research team also involved Dr Richard Yarnell, Dr Martin Bencsik, Dr Louise Gentle, Dr Chris Terrell-Nield, Manfred Chebli and Alexandra Bourit.
Notes to editors:
View the journal paper.
Press enquiries please contact Dave Rogers, Senior Press Officer, on telephone +44 (0)115 848 8782 or via email, or Therese Easom, Press and Internal Communications Manager, on telephone +44 (0)115 848 8774 or via email.