TIRITIRI MATANGI, NEW ZEALAND (FILE) (ZSL/WILDSIDE PRODUCTIONS UK) – ‘Eavesdropping’ technology and artificial intelligence has helped scientists listen in on an endangered bird population in a remote part of New Zealand.
The technique, hailed a breakthrough for conservation, involved deploying recording devices in a remote rainforest to listen in on the endemic hihi bird, allowing the team to assess the success of a reintroduction programme with minimal impact on the birds.
“We’ve put autonomous acoustic recorders all the way around the reserve so we can monitor how these birds behave when we’re not there by listening for their calls, identifying which areas the birds are calling from the most and if that is changing across time,” Oliver Metcalf, a former Conservation Science Masters Student at the ZSL (Zoological Society of London), told Reuters.
A hihi call sounds like two marbles clanging together in what is known as the ‘stitch’ call.
Scientists saw the calls change from an initial random distribution to a more settled home range – marking the hihi reintroduction and the new method a success.
“We found that after a month they settled down into areas that they really preferred being in which were areas close to water which is a really good indication that the hihis are happy with their new site,” Metcalf said.
The study, published this month in Methods in Ecology and Evolution, was carried out in the Taranaki region of North Island, where 40 juvenile birds were released in April 2017.
The first time hihi have been seen in the region since their regional extinction over 130 years ago.
“In Maori hihi actually means first ray of light, first ray of sunshine in the morning. So they have that lovely cultural significant but they’re also really important to us as scientists because they’re excellent bellwether indicator species of forest quality,” Metcalf said.
“If hihi survive there then that that forest is in good condition,” he said.
Combining the remote audio recordings and artificial intelligence to identify the birds distinctive call allowed the scientists to analyse huge amounts of data.
“We developed an algorithm which recognised the hihi call amongst all of that data. Then we worked out where they were calling more often and then we produced a model which hadn’t been used in this way before to work out how there they were was changing over time,” Metcalf said.
Metcalf says he hopes his new technique can be applied to reintroduction programmes for other species.
“You can get very similar, very accurate information as if you were having people following around or as if you were radio tracking the species but without that intrusiveness for the birds,” he said.
Now a PhD at Manchester Metropolitan University, Metcalf is using AI to analyse how bird and frog populations are affected by different levels of forest degradation in the Amazon.