Welcome to Stewart Island & Sails Ashore +64 3 219 1151 tait@sailsashore.co.nz

Ulva Island Rat Infestation

In the early 1990’s Department of Conservation successfully eradicated rats from Ulva Island. The strategy used then was hand poisoning, using Warfarin  . The Island was parallel tracked at 50 m centres with stations every 50 metres. The job took some 4 years, although staff involved told me they thought the rats were gone in around 12 months. I understand Ulva was the second Island in the world to be so cleared. Breaksea Island in Fiordland being the 1st. Post eradication Saddlebacks, Stewart Island Robins, Riflemen and Yellowhead were introduced. And did very well. A network of traps was established on the Island as a 1st line of defence, and to give early warning of re-infestation. Sadly, though perhaps not unexpectedly every year saw at least one rat caught. At lease one was known to have hitched a ride on a boat, with swimming also a path onto the Island. Rats can swim considerable distances, certainly more than the 800 odd metres to the closest part of the Stewart Island “mainland”


In late 2011 either a pregnant female or a pair managed to re-establish a population, and by early March 2012 they were well away. An emerging population.  The hand poisoning method used earlier was deemed unlikely to succeed ….. how do you persuade a rat living in an ice cream factory to walk 50 metres to your particular brand of icecream ??. The decision was made to airdrop 6 tonnes of a 2nd generation anti-coagulant toxin called Brodifacoum.  This was, from memory carried out in late July, with a drop of 4 tonnes. This resulted in a pellet every 4 or 5 sq metres. A second “follow up” drop of 2 tonnes was undertaken around a month later. The DoC staff told me rats were undetectable less than a week after the 1st drop.

The down side … we lost around 45% of our Robin population, and a similar % of our Saddlebacks.  I spent some time on Ulva immediately after the 1st drop, and watched birds in the near vicinity of the pellets, but never saw any interaction  at all. The mortality comes from secondary poisoning, when insects, which are un-affected by brodifacoum, ingests the toxin, and are then eaten by birds.

The robin population recovered in one breeding season …. by actual count by Otago University Students who had been studying robins since they were first introduced to Ulva. The Saddlebacks were not being counted, but by my observations the population seemed to have recovered in around 2 years.    A bit like the army general … Do I send my troops into the field, and lose half of them, but save the entire population, or do I keep them in their barracks and lose everybody.

This operation was apparently the first time anywhere in the world that an emerging population had been stopped in their tracks, and totally eradicated. DoC must have been over the moon as now they didn’t have to wait until the population peaked, and crashed though running out of food.

So from first detection the rats were gone 9 months,



As mentioned above, rats were getting back to the Island quite regularly … at least once a year, and in one bad year 5 got ashore … all swimming. These being picked up in the network of established traps. And then in late 2021 DoC realised that there was again a breeding pair established. They even had trail camera evidence of a pair busy making little rats. The decision was made to attempt to hand poison. Not scattered, but placed in discreet locations, which were marked by a red ribbon. Above I mentioned that the initial eradication was based around a 50 metre square grid….. this being around half the distance that rats will venture from their home burrow looking for food. A “nip it in the bud” approach was certainly worth attempting, and had they used a helo drop at full cover over the 40 or 50 acres the rats were thought to initially be it could well have succeeded. But the bait drop as indicated by the ribbons was sparse in the extreme, with probably well over 100 metres separation over the target area. I guess the only thing that was learnt was that it didn’t work.

So the rats carried on doing what rats always do, making little rats. And it’s exponential growth of course. The numbers are horrific. I’ve been told that in one year a pregnant female might theoretically have in excess of 15000 descendants.


Helo over Sydney Cove.
Before the drop starts the pilot and DoC calibrate the
spreader bucket delivery rate , altitude, and flight speed
over an open field so as to be sure the drop rate and pellet spread is within the specified parameters.
They aim for a 50% overlap each pass.

Pellet on the ground
The pellets are quite small and the amount of pure toxin for the whole project is around 6kg.
If you consider the top 200 mm is home to insects which birds might access then a very crude “guestimation” of the dilution rate would spread that 6 kg through 600.000 tonnes of top soil.
So rain breaking up the pellets and washing the resulting toxin into the soil, would to all intents and purposes be inconsequential. The predominant direction of soil water in a temperate rain forest is downward and eventually flushes into the ocean. And at that stage it would be undetectable.

Pellets on a track
This is about the density I noticed as we walked along a track. .After the helo is finished DoC staff will walk the trails and throw all pellets  into the forest.

GPS Plotting
The marvel of GPS enables the staff to oversee the accuracy of the drop. The helo is equipped with a recording GPS, which logs both track position and time duration of each “run”. This is downloaded onto a computer where the staff monitors the operation to ensure minimum drop per unit area, and that there are no missed areas. Weather is critical, as excessive wind will mean lack of accuracy and subsequent uneven drop.

Spreader Bucket being Lowered for filling.
The bucket has a motor driven spinner, which can be turned on and off by the pilot. It can be fitted with a baffle which allows for an accurate boundary drop, thus stopping over spray along the coast. Keeping bait out of the sea.

Helo landing to refuel.
The machine usually hovers while the spreader is being filled. 
The birds are cut outs stuck onto the window, not seagulls panicked by the big bird. 

Delayed Response

DoC failed to implement an immediate response modelled on the previous successful program, and so missed the opportunity to hit them in July 2022, which was quite wet… not ideal conditions. But that August would have been much better as it wasn’t too wet at all. I’ve discussed with several staff  this failure to do what they did so very well the first time. The responses have been many and varied, and range from funding, to planning to staff et al. All of which I find difficult to come to terms with, as even blind Harry had to know that with regular incursions some day the successful invasion of 2011 would happen again. In some ways it is perhaps not fair to compare the two events, as the last incursion coincided with a Rimu mast and a very heavy Coprosma mast as well. So lots of food. This would have had a major impact on rat breeding ….. lots of food, more litters surviving

In my world at least a competent manager would simply dust off the original plans and what they leant from them since, and swing into gear.   But no. It took until 2023 for the plan to get to the point that last Saturday, 5th August the helo arrive to do the job.

So what will the results of the delay be … 

All of us are hoping the project will be as successful as was the 2011 drop. And if so the impacted bird populations will recover. But not as quickly as was in the past due to that lost year. . Ulva will have lost significantly more of the robin and saddleback population than the previous 46%, as the emerging population of rats will have impacted on breeding success, as well as normal population mortality. It would have been very interesting had Otago University still been doing their Robin work, as then we would have had hard numbers rather than guestimates to lend weight to management responses if (when?) future events take place. As mentioned above the incursion coincided with a mast year, so poisoning would have been into an environment which was at the end of a major food availability period.  And certainly we caught huge numbers of rats in the traps we have around Sails Ashore and Kowhai Lane  approx 70 rats each property in a couple of months late winter. Post mast years Rimu does not fruit, and Coprosma has a very sparse fruiting as well. So a relatively hungry environment, but a rat population still expanding. 

So now we wait, and hope the job has been a success

My comments re DoC’s response to this whole event may seem harsh. But they must be taken in the context of an ex Forest Service Ranger in Charge of Stewart Island. And by ex …. 54 years ago. And also someone who has been self employed ever since. So the Public Service environment I was trained in is arguably a far different beast that the one DoC staff live in now. But sometimes new is not always best,

Photos in this blog were taken with the help and support of DoC staff on the day. Thank you all.