Background
Since early 2018, ZIP (with the support of DOC, NEXT Foundation and Predator Free 2050 Limited) has been carrying out a programme of work at a field site in the Perth River valley, South Westland. This project aims to completely remove possums, and potentially rats, and stoats, from 12,000 hectares of rugged back-country, and to then permanently prevent these predators from re-establishing.
One of the tools ZIP has developed to help protect the site against possums is the ZIP PosStop, a live-capture possum trap based on the PCR No. 1 leghold trap, presented in a raised platform to minimise risks to ground-dwelling birds. Each trap is attached to the platform with a chain, which is in turn fitted at the trap end with a short length of bungee cord to reduce possum escapes. The platforms are mounted on a tree or post, and lured with a white powder-coated aluminium visual lure.
Previous work by ZIP and others has shown lone possums can roam large areas (in the order of 50−100 ha) in search of other possums, so a relatively lean network of devices is believed to be sufficient to intercept them when they enter an otherwise possum-free area. In earlier trials by ZIP, a network of PosStop traps at a density of 1 per 50 hectare successfully prevented possums from re-establishing on the 400 hectare Bottle Rock peninsula in the Queen Charlotte Sound.
In the Perth River valley, we initially intended to deploy a network of PosStop traps, paired with the ZIP OutPost automated reporting system, at a density of 1 per 21 ha, within all forested areas of the project site (a total of 143 traps). However, the Perth River valley is home to an estimated population of 75-120 kea. While we anticipated that the removal of possums from the site would ultimately be beneficial to kea, they are curious and intelligent birds, with a reputation for thoroughly investigating novel objects in their environment. Therefore, before deploying a live network of ZIP PosStops in the Perth River valley, we needed to:
Identify and mitigate any impacts that curious kea may have on the performance of these devices at catching invading possums; and
mitigate any risks to kea from these devices
The ‘SafeLock’ function for daytime lockout
Early in the project, we identified a need to prevent kea from fiddling with components of the trap set up, which could reduce the effectiveness of the trap and possibly cause harm to kea.
To that end, we developed a version of the PosStop trap that included a ‘SafeLock’ mechanism to enable the traps to be locked during daylight hours, when kea are most active – while still enabling the trap to operate during darkness, when possums are most active. In the video below, our engineer John demonstrates how this mechanism works, using an early 3D-printed prototype.
Trial in the Perth River valley
Between December 2018 and July 2019, we ran a trial in the Perth River valley to assess the timing, frequency and nature of kea activity at ZIP SafeLock traps. Ten SafeLock traps were installed in the field site: five along an exposed alpine ridge, and five below the bush line. Each trap was monitored with trail cameras programmed to take 30 second videos upon triggering, with a 5 second delay between each video. To ensure the safety of kea during this trial, all 10 traps were permanently locked and unable to be sprung.
ABOVE: The exposed ridge where five SafeLock traps were deployed - three trap set ups are visible (circled) (L) and one of the five traps installed on the ridge, monitored with two trail cameras (R)
Over 213 days (equating to 1,838 trap nights), the trail cameras recorded 278 kea interactions with the SafeLock trap set ups. All of these occurred at the five ridge traps, with the exception of a single daytime interaction with one of the traps below the bush line.
From December through to the end of April, kea activity at the SafeLock traps was infrequent (34 records), and always occurred between the hours of sunrise and sunset – the time during which the traps would be locked in operational use. We observed a sharp increase in kea activity in late April 2019, which continued for the remainder of the trial. Kea activity increased at the trap sites as kea began to discover, and play with, the chain and bungee cord components of the SafeLock traps. The timing of increased behaviour coincides with the end of the fledging season, a time when kea chicks from the most recent breeding season reach independence and start to explore the world around them.
Our observations suggest that kea learning to play with the trap components may have provided an ‘entertainment’ reward which resulted in an unusually high level of kea activity at, and repeat visitation to, the SafeLock traps, and the occurrence of interactions outside of daylight hours.
In a ‘real world’ scenario, in which these traps were not locked during the night, 12 of the 278 interactions we recorded may have triggered the leghold trap, and been unsafe for kea. While it was difficult to distinguish individual kea in the trial footage, repeat visitation by kea is known to be common at sites where a reward is provided, so we believe these interactions are likely to represent a small number of repeat visitors. All nocturnal activity occurred after the traps had been in place for 6 months, and at a substantially lower rate (0. 01 per trap night) than daytime activity (0. 2 per trap night).
All five of the ridgeline traps were disarmed or had their bungees damaged as a result of kea activity during this period. Left unattended, these traps would not have been able to catch invading possums – and, in fact, may have led in some cases to trap-shyness, by partially catching or simply frightening any possums that triggered them.
Attempts to remove the entertainment ‘reward’ for kea from the trap
Based on these results, we initiated work on a solution to protect the chain and bungee from kea access, and reduce the incentive for kea to interact or play with the traps.
We designed a stainless-steel prototype ‘protection plate’ to fit between the set SafeLock trap and the chain bucket, and tested the effectiveness of the plate in the Willowbank Wildlife Reserve kea enclosure.
To our disappointment, the steel plate was found to have little effect on restricting kea access to the chain and bungee. Kea rapidly learned to hook their beaks around the steel plate, pull the bungee out, and chew on the bungee, within hours of installation (see below). As we observed in the field trial, they also flipped one trap jaw up, which in a field situation would render the trap ineffective at catching possums. A second design, with smaller corner gaps, also failed to prevent kea interfering with the chain and bungee.
ABOVE: A SafeLock trap with jaw flipped up and chain pulled out due to kea interaction
We determined that further prototypes of this type were unlikely to be effective, unless the trap was seriously modified so that the front jaw could not be flipped up, or the whole platform was redesigned with a deeper chain bucket and in-built protection plate. This was not a feasible option for the Perth River valley field site, as retooling, testing and production would likely take months to complete, when we needed a solution much sooner.
Conclusion
We can draw several conclusions from the data gathered from the trials in the Perth River valley and at Willowbank Wildlife Reserve:
The ZIP SafeLock system works to prevent unwanted trigger events. Despite 216 total ‘virtual trigger’ events in the field (and multiple at Willowbank), while traps were locked, none were able to be triggered by kea.
Kea in the Perth River valley are significantly more likely to interact with traps in exposed alpine locations, than with traps below the bush line. Of the five bush traps in this trial, only a single kea interaction was recorded over 895 trap nights in this environment.
Over time, kea can learn to interfere with the bungee and chain components of the traps in a way that regularly renders them ineffective for catching possums.
Sub-adult or juvenile kea are more likely to interact with traps and render them ineffective.
While there were night interactions on the exposed SafeLock traps on the ridge, these occurred at an extremely low rate, and did not occur until after the traps had been in place for 6 months, when kea activity increased (most likely as a result of fledging kea learning to play with the bungee and chain components of the traps).
Overall, risks to kea from the SafeLock trap are very low, and can be further reduced by minimizing kea exposure to the trap.
Where to from here?
Based on the conclusions described above, we plan to minimise opportunities for kea to learn to play with the bungee and chain components of the SafeLock trap, by minimising kea exposure to traps in the Perth River valley field site.
Instead of deploying an extensive, permanent network of automated reporting SafeLock traps as the primary tool to detect and respond to invading possums (as we originally intended), we will use lured trail cameras to detect invading possums. SafeLock traps will then be used to respond to detections in a temporary, targeted network, along with other devices and hunting by our predator dog Pepper and her handler Mike. These traps will then be removed once the possum is caught. Furthermore, SafeLock traps will only be deployed below the bush line, where the probability of kea interacting with traps is substantially lower.
The network of lured trail cameras used in the Perth River valley has proven to be a highly sensitive method for detecting possums, rats and stoats. In addition, the use of trail cameras as the primary detection method for possums is expected to significantly reduce the infrastructure costs associated with protecting the site long term, relative to the cost of a permanent network of automated reporting traps.
A single interaction may be the only opportunity we have to capture an invading possum in the Perth River valley. These measures will ensure that the opportunity is not compromised, while ensuring that kea exposure to the ZIP SafeLock trap is carefully managed.
The damage possums do to native forests is well documented: they are the major cause of the decline of many tree species, including rātā, kāmahi, kōtukutuku and pōhutukawa. But they are also opportunistic predators, and have been recorded preying on the eggs and chicks of native birds, including kea. We are already beginning to see the benefits of removing possums and other invasive predators from the Perth River valley, and our observations suggest that kea are thriving in the field site as we work to maintain predator freedom.
Technical report pending.