Abstract
Reintroductions are an important feature in New Zealand conservation. With the rise of fenced sanctuaries on the mainland, taxa are being reintroduced to novel areas that are typically not encountered by wild populations, such as urban areas and highly fragmented environments. Understanding how animals use the space and resources in these landscapes, and identifying the threats they may face, is key for long-term conservation.
Kākā (Nestor meridionalis) are a large species of forest-dwelling parrot that are endemic to New Zealand and typically require large areas of native forest and a high level of pest control. In the last two decades, kākā have been reintroduced to ring-fenced sanctuaries in Wellington City (Zealandia Ecosanctuary) and Dunedin (Orokonui Ecosanctuary), heralding a new era of urbanised kākā. In Wellington, the population has expanded rapidly, resulting in a self-sustaining population, whereas in Dunedin, the population has been stable at approximately 50 individuals, despite successful breeding and yearly top-up translocations from captive facilities.
To investigate the space use and resource selection of kākā in the Orokonui population, ten kākā were tracked using GPS devices that use a SWIFT fix algorithm. An aim of this study was to test the suitability of these GPS devices with a stationary test, in vegetation conditions that represented kākā habitat. These devices performed comparably to standard-fixing GPS devices, resulting in similar estimates of fix success rate and median of the linear error, although they had more outlier locations.
The second aim of this study was to investigate space use of the Orokonui kākā, particularly their home range size and monthly variability, and to estimate risk based on the proportion of the home range that lies outside the fence of the sanctuary. The ten kākā were tracked for an average of 144 days, deriving locations every three hours. Utilisation distributions were estimated using dynamic Brownian Bridge Movement Modelling (dBBMM), and home range sizes varied widely between individuals (mean +/- SD = 6.6 +/- 5.4 km2), with significant age-related differences. Larger home ranges were observed in younger kākā, which may indicate more exploratory behaviour, and may also place them at higher risk to predation or human-based threats beyond the sanctuary fence. The estimate of risk based on proportion of an individual’s home range that was outside the sanctuary corresponded with mortality events prior to and during the study, with the observed mortality of seven kākā, six of which were three years or younger. There was a high level of monthly variation in home range sizes, which differed between sexes. Females increased their home range sizes prior to the nesting period, whereas males showed an opposite trend.
The third aim of the study was to estimate resource selection by Orokonui kākā to identify their key resources, and to assess their potential for population growth in the landscape beyond Orokonui Ecosanctuary. To provide robust population-level estimates of resource selection within a step selection function (SSF) framework, a population-level model using a Bayesian approach (integrated nested Laplace approximation) was fitted for all individuals combined, with individual as a random effect. Native forest (predominately Podocarpaceae and broadleaf spp.) was the only land cover type to be positively selected at a population level relative to the reference category, kanuka. Estimated probability of resource use was projected over the wider Dunedin area, suggesting a substantial amount of potential habitat is available. To explore individual variation of resource selection, a conditional logistic regression model was fitted for each kākā. Native forest was positively selected relative to kanuka by all but two kākā, and there was mixed selection and avoidance of exotic conifer and suburban areas. There was an increasing selection of native forest, and an increasing avoidance of exotic forest and suburban areas in older individuals, suggesting a higher adaptability in younger individuals.
The present study provides valuable fundamental knowledge of kākā spatial ecology, and highlights the variability of kākā behaviour at an individual level. It is expected that these findings will be valuable to the long-term management of the Orokonui Ecosanctuary kākā population, and will provide useful information for future kākā translocations and management throughout New Zealand.