Acoustic technology helps researchers hear kiwi in the dark

It’s a rare sight to see kiwi in the wild. Nocturnal, shy, and perfectly camouflaged, they remain difficult to monitor even as North Island brown kiwi are now considered non-threatened.

Conservationists can map kiwi territories from their calls, but turning those calls into reliable population estimates remains out of reach.

That gap has prompted Associate Professor Yusuke Hioka, head of the University’s Communication Acoustics Laboratory and the director of the Acoustics and Vibration Research Centre, to rethink how we listen to kiwi.

Instead of trying to see kiwi, he is learning to hear them more precisely.

Hioka is developing a small, low-cost device that can identify a bird call and determine the direction that it came from. His prototype uses four microphones arranged in a tetrahedral shape, allowing it to estimate the direction of a call in a three-dimensional space.

"Knowing where a bird is calling from unlocks powerful ecological insights," says Hioka.

"If one kiwi calls 20 times, the system can recognise it’s from the same location, so it may be the same bird, not 20 different birds."

Hioka’s expertise lies in audio and acoustic signal processing – collecting analogue sounds via microphone arrays and converting them into digital data.

The device captures sound continuously, filters out irrelevant noises like wind, insects and other species, and stores the recording safely. It’s also weather-proof and designed for long-term outdoor use, with timestamping to track behaviour patterns.

"Kiwi only call after dark, so the device has an automatic on-and-off function," says Hioka.

"It only operates at the times it needs to, optimising battery and storage space."

His research is part of AviaNZ, a national bioacoustics collaboration monitoring ecological systems through sound.

The collaboration is led by Professor Stephen Marsland, a mathematician, conservationist and machine learning expert at Victoria University of Wellington.

Marsland says the technology fills a major data gap, allowing conservationists to monitor endangered species without ever touching or disturbing them.

"We can find out where the burrows are and actually look at that and make sure that the eggs are developing well, they’re fertile, and set up cameras to see if the chicks leave the nest," he says.

"Acoustics lets us answer really cool questions and it lets us do it without ever having to handle the birds."

Hioka says the goal of his research is to create a reliable, field-ready tool that can be deployed across Aotearoa and enable everybody from community groups to researchers to monitor their local ecosystems.

"We are in a race against time to protect Aotearoa’s bird life. Conservationists grapple with the challenge of collecting comprehensive monitoring data. We have an innovative solution in the making to help close this data gap," he says.

The project was funded by the Science for Technological Innovation National Science Challenge in 2020 and is currently stalled at the prototype stage.

With additional funding, Hioka says the technology could be refined, strengthened against environmental interference, and eventually deployed across several hundred to a thousand forests nationwide.