Abstract
Within cities, areas of green space, such as patches of forest, play a vital role as habitat for wildlife, and can be especially important for supporting native biodiversity. As such, urban restoration efforts often aim to enhance the quality of existing green spaces to provide better habitat for native wildlife and increase biodiversity as a whole. In New Zealand, the number of urban restoration efforts have been steadily growing in recent years, with emphasis often placed on supporting bringing native bird species back to the city. One of the greatest threats to these conservation efforts are introduced mammalian predators (IMPs). IMPs have substantially impacted native biodiversity throughout NZ, causing the extinction of many endemic species. While the impacts and ecology IMPs have been thoroughly studied outside of urban areas in NZ, their urban ecology is not well understood. Such information is essential if successful restoration and reintegration of native wildlife is to be successful. The research undertaken in this thesis aimed to identify what IMPs are present in the urban green spaces of NZ cities, and what habitat characteristics influence their distributions, for the purpose of improving the effectiveness and efficiency of urban predator management. To investigate these aspects of IMP ecology, transects of predator detection devices were deployed during spring and autumn in three types of prominent urban green space: forest fragments, amenity parks, and residential gardens, and within three NZ cities: Dunedin, Hamilton, and Wellington.
First a broad assessment of predator detection rates across season, city and habitat type was conducted, followed by a comparison of detection methods. Lastly, detection data was paired with complex microhabitat information to generate models used to assess small scale and broadscale habitat preferences. There were five predator species/groups that were regularly detected: mice (Mus musculus), rats (Rattus spp.), hedgehogs (Erinaceus europaeus), possums (Trichosurus vulpecula), and domestic cats (Felis catus), while detection of mustelids (Mustela spp.) were very rare. Possums exhibited a clear affinity for forest patches, with distance to nearest forest patch a strong determinant of detection probability, and were very abundant in Dunedin. Possums were rarely detected in residential areas and were almost absent in Wellington due to a history of robust possum control. Rodent detections were substantially higher in autumn than in spring and were generally detected least in residential areas. Rats and mice were significantly associated with abundant low vegetation cover. Detection probability of rats were also significantly higher in residential properties with compost bins. Rats were more abundant in Wellington, likely in part because of the low possum abundance. Hedgehogs displayed a strong seasonal pattern, displaying much higher detections in spring than in autumn. Hedgehogs had no clear preference for any broad habitat but were linked with forest patch proximity and were very abundant in Dunedin. Detection rate of cats generally increased in autumn, especially in residential areas. This may have resulted in behavioural changes by residential rodents.
Comparison of predator detection methods (chew cards, tracking tunnels, and cameras) found that while some methods are better for specific species, using a multidevice approach is beneficial. Although, cameras were the most effective at detecting all species.
The body of work within this thesis will help towards the effective and efficient control of urban IMPs, which is essential if restoration and biodiversity goals are to be successful in New Zealand cities.