Using passive detection devices to monitor occupancy of ship rats (Rattus rattus) in New Zealand temperate rainforest

Reliable estimates of invasive pest mammal abundance and distribution in New Zealand are vital for effective conservation management of endangered native species. In this study, passive detection devices were used to monitor site occupancy by ship rats (Rattus rattus) in temperate rainforest in the Eglinton Valley, Fiordland, New Zealand. Ship rat occupancy was monitored on three grids of c. 100 ha each, containing 50 tracking tunnels spaced at 150-m intervals, for seven nights each in November 2004, January 2005 and March 2005. Site occupancy estimates were obtainable for only one of these grids, Walker Creek, where estimates increased by a factor of 2.5 over the total sampling period. Detection rates were highest within forest and forest-edge habitats, and on nights with rainfall. In March 2005, 48 ship rats were caught in an effective trapping area of 132 ha. Removal trapping gave an estimated density of 0.38 rats ha–1 (0.36–0.48 rats ha–1, 95% confidence interval). Given the linear relationship between ship rat tracking and trapping rates, we are confident the reported trends in occupancy are realistic, and more accurate than tracking rate estimates. Improving the current monitoring methods so that the probability of detection can be estimated would be a good first step towards more accurate estimates of ship rat distribution.