An investigation of systematic camera trap monitoring for kiwi (Apteryx spp.) : A thesis submitted in partial fulfillment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

Abstract

Introduction: Kiwi (Apteryx spp.) are one of New Zealand’s national taonga/taoka (treasures) that are our responsibility and privilege to protect. To inform protective management, we need to monitor kiwi population responses to management, for which a non-invasive method that can detect all age classes has been lacking. Camera traps are non-invasive monitoring devices, which are increasingly used with recently developed analyses to monitor cryptic terrestrial species around the world. To apply these methods to kiwi, standardized methods need to be developed and benchmarked against existing methods. Aims: The aims of this study were to 1) summarise previous monitoring work on kiwi to develop a draft systematic camera trap monitoring method for kiwi, 2) determine optimal site selection and camera trap set-up, 3) determine optimal survey length and camera trap spacing for use with spatial presence-absence (SPA) analysis and identify whether realistic population estimates are obtained, 4) compare camera traps with a current noninvasive method using acoustic recorders to determine if they give realistic and comparable estimates when used with SPA, 5) compare camera traps with dog survey and an observer listening survey to assess kiwi population health, 6) trial alternative analyses for use with camera traps in high density sites, 7) investigate stereo cameras for their potential to add value to camera trap surveys. Materials and methods: We summarised the current literature on monitoring kiwi and the use of camera trap surveys. We deployed 34 camera traps over six seasons in Orokonui Ecosanctuary. We deployed 29 acoustic recorders and carried out detector dog surveys to compare the number of juveniles detected. We deployed 17 camera traps in Rotokare Scenic Reserve and 18 camera traps in the Cape Sanctuary to examine their effectiveness in high kiwi density areas. We constructed a stereo camera by chaining two off-the-shelf trail cameras together to trigger from one PIR sensor and briefly trialled the stereo camera at Orokonui. Results: Cameras were able to detect kiwi of all age groups and to provide credible population densities and trends. Kiwi detections can be maximised by using a detector dog team to select camera sites and through camera orientation. Comparable population estimates were obtained using spatial presence-absence (SPA) analysis with an optimal survey length of four months, during peak incubation, and optimal camera spacing of 350 m. Cameras and acoustic recorders gave comparable population estimates using SPA. Estimates were realistic based on matrix population model projection. Camera traps and detector dog surveys found a similar number of juvenile kiwi. Estimates obtained using Royle-Nichols analysis likely underestimated population size but correctly indicated population trend direction and magnitude, while the index-manipuation-index method did not give a biologically possible estimate of population density. The stereo camera method using two trail cameras was capable of giving surprisingly accurate bill measurements, but further work in necessary to achieve repeatability. Discussion and conclusions: Systematic camera trapping is capable of monitoring the whole kiwi population, including female and young kiwi that are usually under-recorded by other methods. Systematic camera trapping paired with spatial presence-absence analysis performed well in a low density population. Camera monitoring shows much promise as another useful noninvasive tool in the kiwi monitoring toolbox.