Did poisoning the predators bring back New Zealand's forest birds?

Find your issue, and check it is measured

An environmental issue: something happening, to something, because of something, with proof that somebody measures it.

New Zealand's forest birds evolved with no mammals hunting them. Introduced possums, rats and stoats now eat their eggs, chicks and food, and the country's main answer over large, rugged forests is a poison dropped from the air.

Proof it is measured

Possums, rats and forest birds before and after a 1080 drop, opened: 2,006 five-minute bird counts in four forests, 2012 to 2015, from Manaaki Whenua Landcare Research, with chew cards that measure the possums and rats.

In the news

RNZ, 21 April 2026: a threatened shorebird, the southern New Zealand dotterel, rose from 105 to 160 after a 1080 drop on Stewart Island, and opinion on the island was split over the poison.

The system

Store: the native birds of each forest. Input: chicks raised. Outputs: deaths, many of them eggs, chicks and nesting adults eaten by introduced mammals. The pressure: possums and rats, and the stoats that eat rats, whose numbers boom after a heavy seeding of the beech trees. The intervention: 1080 cuts the predator store. Feedback: negative. The rats that survive a drop have more food and fewer rivals, so they breed fast and their numbers climb back towards where they were.

Write your research question

One sentence naming the indicator and its units, the years and the places. No strategy in it.

A comparison of native birds per five-minute count between forests treated with aerial 1080 poison in 2012 and untreated forests, on the East and West Coasts of New Zealand's South Island, in the summer of 2014/15.

Independent variable

Treatment: whether the forest was dropped with aerial 1080 in 2012 (Wilberforce on the East Coast, Mt Alexander on the West) or left untreated (Harper and Taipo). Human.

Dependent variable

Native birds per five-minute count: every bird native to New Zealand seen or heard at a station in five minutes, added up, then averaged over that station's counts in the summer. Natural.

Why the third summer

It is the last season in the file, and the furthest from the drop. If the birds' gain lasts, it shows here; if the predators have come back, this is where it fades.

Not obvious

Fewer possums and rats should mean more eggs and chicks surviving. But rats breed far faster than forest birds, and three summers is only a few bird generations. The birds may have recovered, or the predators may have got there first.

Find a real-world strategy and its tensions

One real-world strategy people are genuinely arguing about, two named positions, and what the tension does.

Aerial 1080 poison drops: cereal baits spread from helicopters over forest to kill possums, rats and, through the rats, stoats.

The strategy

The Department of Conservation (DOC) runs a national programme of drops, its National Predator Control Programme, which began in 2014 as Battle for our Birds and is part of the Predator Free 2050 goal. In 2023 to 2024 DOC ran 28 aerial 1080 operations over 735,000 hectares. The drops in this file were earlier, one-off operations against possums.

Position 1 · environmental, scientific

Dr Jan Wright, Parliamentary Commissioner for the Environment, in Evaluating the use of 1080 (June 2011): "not only should the use of 1080 continue (including in aerial operations) to protect our forests, but that we should use more of it." Other methods, she found, can complement it but cannot replace it.

Position 2 · cultural, economic

The New Zealand Deerstalkers' Association, which speaks for recreational hunters, in its policy on poisons: it is "unequivocal in opposition to poison as a non-selective killer of game animals" and rejects the argument that limited "by-kill" is an acceptable cost of poisoning. It wants other ways of controlling the pests developed and used.

The tension

Saving native birds across forests too big and steep to trap, against killing animals the poison was not aimed at, some of which people value.

What it does

Resistance: hunters campaign against large drops, as the Deerstalkers' Association did over Stewart Island's in 2025. Uneven costs: hunters lose deer, while the gain in birds is shared by everyone. And a drop suppresses the predators rather than removing them, so the same forests are dropped again every few years.

For the environment

Fewer predators for the birds, lizards and insects of the forest, but a poison that also reaches animals it was not meant for.

Link to the question

The case for aerial 1080 rests on the birds coming back. The question asks whether they had, in two pairs of forests, three summers after a single drop.

Select and document your data

A dataset you can defend: counted, documented, and cut to a rule fixed before you looked.

Source

Manaaki Whenua Landcare Research DataStore, Data from "The effects of one-off aerial 1080 possum-control operations on common forest birds" (Burge, Van Vianen, MacFarlane and Kelly): the bird counts are birdsrelabund.txt, which ends .txt but is comma-separated. 76,228 rows.

What a row is

One species in one five-minute count. Every count lists all 38 species, so most rows read 0: that is a real zero, no bird of that kind heard. One count is one Site, Transect, Station, Date, Time and Observer together.

Filters

Season is Summer 2014/15, and Winter 2012 for the check in step 5. Keep species, count, Site, treatment and the six columns that identify a count.

Inclusion rule

Native means Endemic or Native on NZ Birds Online: 22 of the file's 38 names. That includes silvereye, the most counted bird in the file, which arrived from Australia by itself in the 1850s. The 14 introduced species are left out, and so are unknown and finch, which nobody could identify. Decided before looking; check the answer without silvereye too.

Derived quantities

For each count, the native birds added up. Then for each station, the mean of its counts that summer: one number per station, 20 per forest.

Why stations, not counts

Each station was visited on three or four days, with two counts back to back on most visits, so its counts are not independent of each other. Averaging to one value per station stops a busy station counting several times over.

How the numbers were made

An observer stands at a marked station for five minutes and records every bird seen or heard. It is an index of how many birds there are, not a census: the weather and the noise, recorded in the file, change how many you hear.

Controls

Pair: each poisoned forest is compared with the untreated forest on its own coast, never across coasts. Restrict: one season, so every forest is compared in the same summer. Normalise: birds per count, because the forests had different numbers of counts.

Credit

Burge, Van Vianen, MacFarlane and Kelly, Manaaki Whenua Landcare Research DataStore, doi:10.7931/J2ZW1J3R, CC BY-SA 4.0, with the date downloaded.

Treat your data

Run a statistical test, leading to figures you built and a statistical result.

Mann-Whitney U, two-tailed, at p = 0.05, once for each coast.

Why this test

A difference between two independent groups of stations, 20 and 20. Bird numbers per count cannot go below zero and a few stations are very busy, so the groups need not be bell-shaped: a rank test rather than a t-test.

Decided before running it

Critical value of U for 20 and 20: 127 (check it in the table you cite). A U at or below it is significant. Run it on each coast separately, and report both: if the two coasts disagree, that is a finding. Free calculator: Social Science Statistics.

The check

Run the same test on Winter 2012, before the drop (20 against 20 stations on the East Coast, 10 against 19 on the West). If the forests already differed then, a difference in 2014/15 cannot all be put down to the poison.