New Zealand Journal of Ecology (2014) 38(1): 86- 102

How can we detect introduced mammalian predators in non-forest habitats? A comparison of techniques

Research Article
Georgina A. Pickerell 1*
Colin F. J. O’Donnell 2
Deborah J. Wilson 3
Philip J. Seddon 1
  1. Department of Zoology, PO Box 56, University of Otago, Dunedin 9054, New Zealand
  2. Ecosystems and Species Unit, Department of Conservation, Private Bag 4715, Christchurch 8140, New Zealand
  3. Landcare Research, Private Bag 1930, Dunedin 9054, New Zealand
*  Corresponding author

Efficient detection techniques will confirm the presence of a species at a site where the species exists, and are essential for effective population monitoring and for assessing the outcome of management programmes. However, detection techniques vary in their ability to detect different species. A wide range of mammalian predator species, most introduced into New Zealand since the late 18th century, have had a detrimental impact on the native flora and fauna. To date, there has been little research to compare the efficiency of detection techniques for these species, especially in non-forest habitats. We used nine commonly-available techniques to survey for the presence of mammalian predators at 19 sites on the open, non-forested banks of the Rangitata River, a large braided river in the South Island, New Zealand. We compared the relative efficiency of the techniques using three metrics: raw detection rates, Kaplan–Meier survival analysis, and probability of detection. Techniques varied in their ability to detect eight species of mammalian predator. The most efficient detection techniques included large tracking tunnels and hair tubes for feral cats (Felis catus), large tracking tunnels for European hedgehogs (Erinaceus europaeus), and WaxTags® for brushtail possums (Trichosurus vulpecula). Using our data to simulate a reduction in survey effort, we found that detection rates would be significantly reduced only when devices were at very low densities. We show also that 3–71 nights of monitoring are needed for a 90% probability of detection by our most efficient techniques. Our findings emphasise the merit of using more than one technique to detect a species, and we recommend that detection devices are left open for at least 10 nights. Finally, we highlight the need for further research to develop standardised monitoring protocols for introduced mammalian predators in New Zealand’s non-forested habitats.