Clever new traps that can instantly tell a kea from a possum are being added to New Zealand’s bold campaign to be predator-free – now exactly 10,000 days out from the mission’s 2050 deadline.
In rural Auckland and Coromandel, researchers have been testing out fully-automated, ultra-low power kill traps developed by Kiwi company Critter Solutions.
Able to be left in the wild for long periods without servicing, the traps trigger only when a lurking animal is identified by its artificial intelligence-powered imaging tech as a target pest.
“These traps are the first to be able to effectively ‘think for themselves’ and make a decision as to whether an interacting animal is a target pest species or not,” the project’s lead, Dr Helen Blackie, of Boffa Miskel, said.
“A key technical success has been developing the trap so that it is extremely fast at triggering when an animal is recognised - within a fraction of a second - after detecting the presence of a pest.”
The innovation also meant the traps could be designed to be far more open – something that’s posed a problem for accidentally killing other species in other traps.
“Traditional traps also require the pest to push, pull or stand on a trigger to activate the trap, which can further reduce catch rates. By using AI, we can do away with manual triggers completely.”
Further, the traps – expected to be ready for sale late next year – could remotely notify a user they had triggered, and send an image of the animal detected.
“This is the first time I’ve been able to run a trial from the comfort of my couch,” Blackie said.
“We use our AI monitoring cameras trained on the traps which show us the approaching species in real time, then we get a notification of the trap triggering, which species it was, and a further notification that the trap has reset and is ready and waiting for the next animal.”
The traps have already passed requirements for delivering a humane kill for rats, stoats and possums in independent trials.
As well as controlling pest species, the device can also be used in a “passive” mode to collect monitoring data on native species.
“We have enough data from the first field trials already to prove the trap’s effectiveness for possums, which is hugely exciting,” Critter Solutions director Kenji Irie said.
“Every time the fully automated trap has identified a possum and triggered has resulted in a successful kill.”
The project was boosted with funding from Crown-owned Predator Free 2050 Ltd, via the Government’s Jobs for Nature initiative.
“We want to ensure that achieving national eradication of rats, mustelids and possums - the Predator Free 2050 goal - is carried out in the best way possible,” Predator Free 2050 Ltd’s Professor Dan Tompkins said.
“Reducing impacts to non-target species is a critical part of that.”
An immense amount of work had gone into ensuring the device can reliably identify even small details of non-target species including kea, kaka, weka and kiwi, from thousands of animal interactions with the prototype units.
It comes as the 2050 campaign – announced just over seven years ago by former prime minister Sir John Key – today comes within 10,000 days of its stated mid-century goal.
Predator Free 2050 Ltd now funds 18 large projects targeting possums, rats and stoats over more than 750,000 hectares, in a mix of rural and urban landscapes.
Already, more than 50,000 hectares has been cleared and is being defended using a range of measures and technology.
The Government plans to review the 2050 strategy – with the possibility of adding feral cats to the hit-list – next year.
Despite recent progress, “interim” goals originally set for 2025 – including clearing all of our offshore islands of mammalian predators, driving possums or mustelids like stoats out of at least one city, and increasing by one million hectares the area where predators were suppressed – are far from being realised.
Successive studies and assessments have pointed to obvious roadblocks.
A recent review led by Manaaki Whenua Landcare Research set out several major regimes in which rodents, particularly, might be controlled: island eradications, fenced and unfenced eco-sanctuaries, and large-scale 1080 drops.
The bad news, they concluded, was that entirely eradicating them from the mainland was currently considered “impossible”, given their bounce-back couldn’t be stopped with most measures.
Another University of Auckland modelling exercise found that, if New Zealand’s rate of eradication implementation continued as is, the country wouldn’t be rat-free anytime in the foreseeable future.
In an earlier study – and the first strategic assessment of how various predator-busting tools could be deployed across the country at scale over the next 10 to 15 years - the same researchers found toxins could be potentially dropped over more of the country’s land area.
At the same time, predator-proof “exclusion” fences were found to be only suitable for about 500sq km – or just 0.2 per cent of the mainland – and some 29,000 sq km of pest-invaded land was likely unsuitable for any measures we currently had.
While some scientific reviews have singled out the potential for new genetic techniques, the most promising ones today might not be so a decade or two from now – and the Government currently has little appetite in using them.
All the while, our biodiversity crisis has worsened. The latest Stats NZ data shows more than three-quarters of native reptile, bird, bat and freshwater fish groups are threatened with extinction – or at risk of becoming so.
Predator Free 2050 chief executive Rob Forlong remains upbeat about progress.
“More and more we are seeing our earlier projects accomplishing significant goals that are showing us that with community support and innovation it can be done at wide scale on the mainland,” he said.
“There is no silver bullet. It will take a suite of new technologies, leading research and science, along with social and community engagement, but given how far we have come in such a small amount of time, I believe we will get there.”