Conservation dogs for controlling and monitoring invasive species
Since prehistoric times hunters have used dogs to help find and capture their quarry. Today specially trained dogs are also used in conservation, including monitoring rare or endangered species and detecting invasive plants and animals. Manaaki Whenua researchers have been studying the use of conservation dogs to manage invasive species.
In collaboration with the Department of Conservation (DOC), universities, and the New Zealand Conservation Dogs Programme, Al Glen and colleagues have tested the ability of trained dogs to detect invasive animals, compared the cost-effectiveness of trained dogs with that of motion-triggered cameras, and investigated how the use of trained dogs could be enhanced by search methods originally developed for military applications.
The Conservation Dogs Programme (www.doc.govt.nz/conservationdogs) is underpinned by dog teams trained to detect various native and introduced species. The teams assist in monitoring rare avifauna such as kiwi and whio (blue duck), and in tracking down invasive animals such as rodents, stoats and feral cats. When an incursion is reported into a pest-free area, such as an island or fenced sanctuary, conservation dog teams are often called in to help find and remove the invader. The teams also conduct routine surveillance of pest-free islands such as Rangitoto and Motutapu, and check vessels and cargo bound for these and other pest-free islands to prevent pests from invading them.
Scientists from Manaaki Whenua, DOC and Auckland University recently took advantage of a routine surveillance operation in the Hauraki Gulf to estimate the probability that trained dog teams could detect rats at distances of up to 100 m. The trial used dead laboratory rats, and the island’s rodent-free status ensured there was no scent of wild rats to interfere with the results. The dogs detected 100% of rats placed directly along their search path, but the probability of detection declined sharply with increasing distance from it. For rats placed 10 m from such search paths, there was a 33% probability they would be detected. However, in practice dogs probably detect rats from greater distances by following scent trails left behind by them as they move.
In a separate trial, Al and his colleagues used motion-triggered cameras to monitor feral cats on farmland in Hawke’s Bay. After the cameras had been left in place for 3 weeks, the monitored area was also searched by a specialist cat detection dog team. The two methods had a very similar probability of detecting a cat, if present, and the cost of each method was also similar. However, the choice of using cameras or dogs depends on several factors. For example, cameras are only useful for detection, whereas dogs can assist their handlers to locate and capture or kill animals. On the other hand, cameras can operate in most weather conditions, whereas dogs require fine weather. That said, both methods will have an ongoing role in managing invasive species in New Zealand.
Al and his colleagues’ research so far shows that conservation dogs are undoubtedly a valuable tool in wildlife management. However, the search methods used by dog teams are generally ad hoc and based on the personal judgement and experience of the handler. For example, one dog team may choose to search along parallel transects, whereas another may focus on areas of habitat likely to be preferred by the target species. If nothing is detected, what does that mean? Is the target species absent from the search area, or do teams just need to keep searching?
Search theory − a branch of mathematics originally developed for naval applications – can help answer these questions. Armed with estimates of detection probability and the range of distances at which a target animal is likely to be detected, optimal search patterns can be designed for different areas. Search methods can also be standardised so that results can be directly compared between different times and places. A review paper soon to be published in the journal Wildlife Biology describes how and why these approaches could further enhance the usefulness of conservation dogs.
The review paper also outlines priorities for further research. First among these is to estimate the detection probability for a range of non-rodent target species at various distances, as estimated for rats. Detection probability may vary depending on conditions such as temperature, wind, habitat, and the dog team’s level of skill and experience. As more is learnt about these variables, it will be possible to use software to design optimal strategies for specific search scenarios, including different areas, habitats, target species, and available resources. Such software already exists for applications such as search and rescue. Similar tools could help pest managers more effectively harness the formidable abilities of conservation dogs.
This work was funded by the Department of Conservation.
Karen Vincent (Department of Conservation)
James Russell (The University of Auckland)