Using detection dogs for surveillance of invasive dama wallaby (Notamacropus eugenii) in North Island, New Zealand

Abstract

Context Dama wallabies (Notamacropus eugenii) were introduced into the Bay of Plenty Region, New Zealand, in the early 1900s. They subsequently became an invasive pest, damaging pasture, silviculture, and native vegetation. One key management strategy is the eradication of isolated populations. Aims First, we empirically determined the detection probabilities of detection dogs and handlers searching for faecal pellet groups of dama wallaby in pasture and forested habitats. Second, we used detection probabilities to derive surveillance system sensitivity (SSe) and estimate the cost per hectare required to have a high confidence (95%) that a targeted wallaby population has been eradicated. Methods We collected and deployed faecal pellet groups in an area with no naturally occurring wallabies. Following faecal pellet surveys by detection dogs and handlers, we estimated the probability of a dog–handler team detecting a pellet group and derived SSe. We derived SSe for a single faecal pellet group and, using simulation, upscaled this estimate to multiple pellet groups voided by a single surviving individual. Key results The detection probability of dogs searching for a single faecal pellet group that was within the detection swathe was relatively high (minimum of 45%). Scaling this instantaneous probability of detection for a single pellet group to the multiple pellet groups voided by a single wallaby around its home range resulted in 30–33-fold increases in SSe compared to the sensitivity of searching for a single faecal pellet group. The costs of surveillance for confirming eradication using detection dogs and handlers are NZ$54–NZ$72 for a 100-ha area. Conclusions Detection dogs and handlers are an efficacious and cost-effective surveillance method for confirming eradication of dama wallaby in open and forested habitats. Implications Detection dogs and handlers are an important surveillance tool for the management of wallabies in New Zealand. The data from this study enable managers to determine the required surveillance effort to have high confidence (e.g. 0.95) that a wallaby population has been eradicated, or that a suspected new population is actually absent if no wallabies are detected. Additionally, it enables per hectare costs of surveillance to be estimated and compared to alternative methods.