Following the ‘lonely’ rat releases on Bottle Rock between July and September 2015, we began to reconsider our approach to detecting invaders in the Remove and Protect system.

These releases demonstrated to us that the ‘footprint’ of an individual ship rat in the landscape can be very small and therefore extremely difficult to detect using current methods.

We are now looking to determine whether the dispersal footprint of a first generation breeding event is both detectable and manageable using currently available tools.

Background – a ‘Generation One’ breeding event in action

By early September 2015, after a sustained effort by the ZIP field team, all resident rats had been removed from the Protected Area at Bottle Rock peninsula.

The field team monitors the Protected Area intensively on a 3 to 4 week cycle for both rat and possum incursion (which can happen through the ‘virtual barrier’ or via the surrounding ocean). Rats are detected using peanut butter baited chew cards, and removal is carried out using a 100 m x 60 m network of DOC150 traps, also baited with peanut butter, and bait stations containing first generation anticoagulants.

In late-September 2015 a pregnant female ship rat arrived in the protected area. At the time, we didn’t recognise this incursion as the beginning of a breeding event, but rather as a single invader. However, when our GIS analyst and modeller Nick analysed the data some time later, a pattern emerged. As you will see in the animated sequence below, it became clear that a breeding event had taken place.

In the animation below, the symbols flash on when the animal was discovered, dim shortly after, and disappear after 2 weeks.

What we observed

The first invader was detected in September 2015, with the last kill completed in December 2015. The entire Generation One event lasted a total of 90 days, and the invasion ‘footprint’ extended to a total of 66 ha.

A total of nine kills were made – it is likely that this was the size of the litter plus the mother.


If the typical footprint of a Generation One event is up to 100 ha (i.e. 1 km²), we believe that detection could be relatively light (say just one device every 50 ha) and predominantly coastal and waterway biased, stretching up to approximately 1 km inland.

In the breeding event we observed, the pregnant female appeared to have centred her home range near waterways, and her offspring also had a tendency to remain close to water. Therefore, there may be an opportunity to target detection into features such as streams.

Provided incursions are infrequent, an individual male or non-pregnant female rat would cause minimal ecological damage in an otherwise rat-free area, and will be functionally extinct as it is unable to breed.

We also hypothesise that we could have up to 100 days to detect and remove the first generation of invaders, before the juveniles reach sexual maturity and begin breeding themselves.

Additionally, the bigger ‘footprint’ of a breeding event (as opposed to an individual rat) means that detection may be scalable, although will require a bigger treatment area for removal.

It is likely that we would need to detect the breeding event within 60 days to give enough time to refine the invasion ‘footprint’ (i.e. by deploying more intensive detection) and remove the mother and offspring. Potentially, naïve juvenile behaviour could aid detection by making offspring more likely to interact with devices.

If this is successful, we could be in a position to design a very low-density detection system to manage rat incursions at a ‘Remove and Protect’ site, with potential cross application to the management of island and fenced sanctuary incursions.

So what's next?

To better understand the ‘footprint’ of a first generation breeding event and our ability to detect it, we are planning an experiment at Bottle Rock in which we will release a radio collared pregnant female rat into the protected area and observe the spatial distribution of her emerging offspring. We will begin this work as soon as possible, although may need to wait until spring 2016 for the best chance of breeding success.