Predicting Possums: Modelling reinvasion of the common brushtail possum (Trichosurus vulpecula) within a New Zealand urban centre

Abstract

Invasive mammalian pests threaten biodiversity globally across a diverse range of habitats. Recent interest in urban biodiversity enhancement has increased the need for effective urban pest control, however efforts are hampered by a lack of understanding of the unique ecological processes occurring in cities. Projects seeking to eradicate mammals from mainland urban sites face the ongoing threat of reinvasion, and as such, often represent a significant long-term investment. There is a need for fundamental research of the invasive mammals that occupy urban habitats, and for this research to be integrated into evidence-based management.

The introduced common brushtail possum (Trichosurus vulpecula) has invaded the majority of New Zealand landscapes, and is a nationally significant pest species. Decades of control and research in New Zealand have been focused on non-urban habitats, but the lens has shifted with the initiation of a national pest control initiative, which aims to eradicate T. vulpecula and several other mammalian pest species from the whole of New Zealand by 2050. New urbanbased control projects have arisen in response, creating a need for urban-based research of possums. The overall aim of this thesis was to inform the management of possums in a New Zealand city, Ōtepoti/ Dunedin, by assessing the density of possums across urban habitat types, and applying this information to a spatially explicit, individual-based model of possum reinvasion.

The first aim of this study was to estimate the density of possums across three habitat types within Dunedin, highlighting the capacity for cities to harbour possums. Density was assessed at three sites, representing an urban forest fragment, and two residential areas of varying vegetation quality. Possums were live-trapped and camera-trapped over 8 days. Spatially explicit capture-recapture methods were applied at the forest fragment site to estimate density, while Minimum Number Alive estimates were calculated at the residential sites. The forest fragment was found to support possums at a density capable of inflicting harm on resident native wildlife, however the density was at the lower range of what might be predicted for its habitat type, suggesting an influence of anthropogenic disturbance. Possums were at similarly low numbers in the two residential sites, however evidence pointed to the possibility of behavioural trap avoidance, which should be pursued as a future avenue of research.

The second aim of this study was to incorporate these density estimates into the application of a spatially explicit, individual-based model of possums, to simulate trapping scenarios a community group can implement to manage the reinvasion of possums onto the Otago Peninsula following eradication. Patterns of possum reinvasion were examined, and the relative efficiencies of trapping efforts and layouts were assessed through simulations. Model results highlighted the importance of complete eradication of possums, targeting possum dispersal corridors, and increasing trapping effort.

The present study outlines the importance of urban areas as habitat for possums, and demonstrates the potential for a spatial modelling tool to inform urban mammalian pest management in Dunedin, and elsewhere in New Zealand.