Issue 1

December 2002

ISSN 1175 - 9844

The Young and the Restless: Dispersal errets and Surviv al of JJuv uv enile F Survival uvenile Ferrets

CONTENTS The Young and the Restless: Dispersal and Survival of Juvenile Ferrets

1

Editorial

3

F

errets are significant predators

indicates that effective ferret

of indigenous wildlife (birds,

control may be compromised by

eggs, lizards, and invertebrates) in

the rapid immigration of juvenile

New Zealand. They also carry

ferrets, and thus buffers designed

bovine tuberculosis (Tb), a disease

to prevent at least 50% of young

that threatens New Zealand’s beef,

ferrets reinvading a control area

dairy, and venison markets.

need to be at least 5 km in width.

In 1997–98, Andrea Byrom

Survival of the radio-collared

(working in braided riverbeds in

juvenile ferrets depended largely

the Mackenzie Basin) studied the

on the population density of

dispersal of young ferrets from

adult ferrets. In an area where

their place of birth and their

ferret density had been reduced

survival in their first 4 months of

to protect nesting areas of

life. Andrea radio-collared 52

endangered birds, juvenile ferrets

New Technology – A Proximity

juvenile ferrets when they emerged

had remarkably high survival:

Detector System

14

from their mothers’ dens, and then

86% in the first year of the study,

every few days from January to

and 100% in the second

Contacts and Addresses

15

April she radio-tracked each ferret.

year. However, in a

Ferrets or Stoats: Which are Worse for Kiwi?

4

What Limits House Mouse Population Irruptions in Beech Forests?

6

Sustained Feral Goat Control – Mt Egmont National Park

8

Are We-ta- Populations Affected by 1080?

10

Some Ecological Limitations of Predator Control

12

site with high

A Selection of Recent Vertebrate Pest-related Publications

16

Dispersal distances of juvenile ferrets ranged from less than half a kilometre to 45 km (see figure), with about 50% moving more than 5 km. This

Manaaki Whenua Landcare Research

densities

Kararehe Kino

December 2002

8 Females Males

7

No. of ferrets

6 5 4 3 2 1

8 -4

-4

4

44

-4

0

40

-3

6

36

-3

2

32

8

28

-2 24

20

-2

4

0

6

-2 16

-1

12

8

12 8-

4-

0-

4

0

Distance travelled (km)

Fig. Distances moved by radio-collared juvenile ferrets in their first 4 months of life.

Captured ferrets were sexed, weighed, ear-tagged, and radio-collared before release.

of resident adult ferrets and no

males to move several kilometres

concern to managers because of

history of recent control, juvenile

from their birthplace, contrary to

the risk female ferrets pose in

survival was lower: 19% in the

popular belief that only males

establishing a new population.

first year and 70% in the second

disperse long distances. In fact, the

year. People carrying out ferret

longest distance covered was 45

Some questions still remain

control should therefore be aware

km by a radio-collared female.

however. Andrea suggests

that survival of juvenile ferrets

Likewise, male and female ferrets

comparing the rate of recovery of

might actually be enhanced by

had similar survival rates. This is of

ferret populations trapped in

previous ferret control operations, and that the observed densitydependent survival of juvenile ferrets should be a key element in determining the frequency and seasonal timing of ferret control. The data also indicated that ferret control to conserve native wildlife and minimise Tb spread should be carried out in late autumn after juvenile dispersal, to provide a longer time lag before juveniles reinvade an area. If successful, this would reduce the need for annual control of ferrets. Another interesting finding was that female ferrets were just as likely as

2

Juvenile ferret with radio-collar.

Vertebrate Pest Research

December 2002

species like possums and pigs, thereby starting new cycles of infection of Tb in wildlife. Some of this research is currently being undertaken by Landcare Research. This work was funded by the Foundation for Research, Science and Technology.

Ferrets can move long distances during dispersal.

Andrea Byrom works on the spring and autumn, to be sure that

If so, ferrets are capable of creating

population ecology and

autumn trapping is more cost-

new foci of infection far from the

management of mustelids.

effective. She also believes it would

original source of Tb. A third area

be extremely useful to find out

needing investigation is the

For advice on ferret control visit

whether juvenile ferrets are

potential for Tb-infected ferret

www.landcareresearch.co.nz/

infected with Tb before dispersal.

carcasses to transmit Tb to other

research/biosecurity/ferrets

his is our first issue of a

We will seek to address key questions

newsletter every 6-months or so.

revamped Landcare Research

Editorial

T

that land managers and the public

We hope Kararehe Kino–

newsletter that expands the focus of He Ko- rero Paihama – Possum

want answers to about protecting

Vertebrate Pest Research is as

New Zealand’s indigenous biodiversity

favourably received as its

Research News to cover all our

from assault by vertebrate pests

predecessor, and we encourage

research on vertebrate pests. The

and improving vertebrate pest

you to take up any issues of

change is in recognition of Landcare

management. Articles will cover

concern to you raised in our pages

Research’s ongoing work with a

recent and ongoing research

with the authors involved.

wide range of vertebrate pests in

funded largely by FRST, the

addition to possums, and of the wide

Department of Conservation, and

suite of vertebrate pests managed

the Animal Health Board, on the

by many of our readers. The

ecology and management of major

Foundation for Research, Science,

vertebrate pests in New Zealand.

and Technology (FRST) is providing funding support for us to broaden the focus of this and future issues

We will continue to send to those of you who received He Ko- rero

of our newsletter.

Paihama, free copies of this

Jim Coleman

3

Kararehe Kino

December 2002

Ferrets or Stoats: Which are Worse ffor or Kiwi? Pete & Judy Morrin

I

n the next few months, the Biosecurity Amendment Bill will

go before Parliament for its second reading. If it is eventually passed, it will prohibit the keeping of ferrets in captivity. This legislation was called for, in part, by conservation groups concerned at the impact ferrets were having on kiwi and other birds. They argued that the liberation of unwanted ferrets and escapees were boosting populations in the wild, and allowing ferrets to establish in areas not yet colonised. So what effect is the legislation likely to have? Will it really benefit kiwi, or will it merely deny ferret lovers the chance to keep the animals as pets? John McLennan is in no doubt that ferrets kill kiwi, weka, and waterfowl. Ferret kills are generally unmistakable. The animals often leave distinctive strong–smelling faeces near the remains of their prey and their canine puncture marks are generally easily recognised. Ferrets also often eat the bill, skull, neck-vertebrate and skin of birds, something that cats and stoats seldom do (Fig. 1). At

Fig. 1. The remains of an adult male kiwi, killed by a ferret and eaten over 1–2 days. An adult male kiwi weighs about 2 kg, possibly twice as much as the ferret that killed it.

Lake Waikaremoana, ferrets killed four radio-tagged kiwi (three

pets (e.g. dogs, cats) and could

John also believes the decline of

adults and one sub-adult) over a

ferrets significantly accelerate kiwi

kiwi is being accelerated by ferrets.

10-year period. In Northland forest

decline? John’s answer to the first

However, kiwi are likely to

remnants, an adult male ferret is

question is a qualified yes. In the

continue to decline even if ferrets

believed to have killed six adult

three localities mentioned above,

are eliminated, because stoats are

kiwi over a 5-month period, while

two other adult or sub-adult kiwi

the main threat driving kiwi

in northern Hawke’s Bay, ferrets

were lost, one to a cat and the

populations towards extinction in

killed five of 18 sub-adults

other to a dog. In a more

mainland forests. In most

released into a reserve.

extensive 8-year study in

localities, stoats kill about 60% of

Northland, eight radio-tagged

chicks in their first 20 weeks of

So are these predation losses any

adult kiwi were lost to dogs and

life. Natural mortality also plays a

worse than those caused by other

nine to ferrets.

part, and up to 95% of chicks fail

4

Vertebrate Pest Research

December 2002

to reach adulthood. The few that

longevity was 26.8 years. If the

populations. This target is seldom

do make it are insufficient to

three ferret kills had not occurred,

reached in the presence of

replace the adults, which inevitably

mortality would have been 2.1%

uncontrolled stoat populations.

die of old age. It is this shortfall in

per annum, and the average

recruitment resulting from predation

longevity 46.8 years. Incredibly, the

Clearly, several different predators

that is causing kiwi populations

loss of just three adults nearly

are collectively responsible for the

throughout the North Island to

halved average adult longevity and

decline of kiwi in mainland forests.

decline at about 6% per year.

lifetime productivity.

Nevertheless, John believes that

Given the extraordinary impact of

In Northland, ferrets increase adult

decline. If their impact was

stoats on kiwi chicks, it is easy to

death rates by as much as 5% per

eliminated for 2 consecutive years,

dismiss the loss of a few adults to

year. When this happens, stoat

kiwi populations would double.

ferrets as inconsequential.

predation on chicks becomes

Two years of reprieve from ferrets

However, predation on adults is

critically important. If females

(with stoats still present) would

much more significant than

produce just 12 chicks over a

hardly make any difference.

predation on chicks. In the absence

(much shortened) 12-year adult

of predators, natural mortality

lifespan, then nearly 20% have to

Kiwi are in trouble, so any effort

rates of adult North Island brown

survive to adulthood to maintain

that increases their survival rates

stoats are the primary driver of kiwi

must be viewed

and they live for 20–50

favourably. The legislation

Pete & Judy Morrin

kiwi are 2–5% per year, years. During this time, pairs fledge about one chick each year, and two of these chicks must reach

before Parliament will benefit wildlife if it prevents ferrets from establishing in uncolonised

adulthood to replace their

areas. However, nothing

parents. Any additional

much is going to change

survivors contribute to

throughout the greater

population growth.

part of New Zealand

Clearly, populations can

where self-sustaining

withstand chick mortality

populations of ferrets

rates of 90–96%, and still

already exist. In these

persist, provided adults

places, ferret lovers may

have normal life spans.

justifiably feel aggrieved if the feral ferret

Small changes in adult

populations are not

survival profoundly

controlled.

influence adult longevity, and thus the chick survival

This work was funded by

rates required for

the Foundation for

population stability. At

Research, Science and

Lake Waikaremoana, the

Technology, Department

annual mortality of 74

of Conservation, and Bank

radio-tagged adults

of New Zealand.

monitored by John over a 10-year-period was 3.7%, and the estimated average

John McLennan studies John McLennan with a sub-adult male kiwi.

kiwi in the wild.

5

Kararehe Kino

December 2002

What Limits House Mouse P opulation Irruptions in Beech Population Forests?

B

eech forests in New Zealand seed heavily at irregular

intervals, providing periodic increases in food for native birds, insects and introduced rodents. Within 3 months of a heavy beech seedfall, house mouse and ship rat populations increase, followed by an increase in stoat populations. This cascade of pest irruptions is a major conservation concern in New Zealand because mice in beech forest prey on beech seed and native invertebrates, while ship rats and stoats prey on both invertebrates and also grounddwelling and hole-nesting native birds. The Hollyford Valley research site.

Wendy Ruscoe, Ivor Yockney and Richard Heyward have been

valleys in Fiordland National Park.

standard live-trapping techniques

studying the factors limiting rodent

In each valley, rodent populations

on two grids. Beech seedfall was

and stoat populations in beech

were monitored quarterly from

also scored. In the Eglinton Valley,

forest in the Eglinton and Hollyford

May 1999 to February 2002 using

792 stoats were destroyed by the

2000

40

1000

20 May Aug Nov Feb May Aug Nov Feb May Aug Nov

2000

1200

25 20 15 10

400

5 May Aug Nov Feb May Aug Nov Feb May Aug Nov

1999

2000

2001

40

May Aug Nov Feb May Aug Nov Feb May Aug Nov

1999

0

2000

0

2001

d) Hollyford MR2

30

800

0

2000

0

0

35

Beech seedfall Mice

80

3000

2001

c) Hollyford MR1 1600

120

1000

Seeds /m2

1999

Beech seedfall Mice

1800 1600 1400 1200 1000 800 600 400 200 0

80

Beech seedfall Mice

60 40 20

May Aug Nov Feb May Aug Nov Feb May Aug Nov

1999

2000

mice MNA

60

Seeds /m2

4000

3000

0

Seeds /m2

80 mice MNA

Beech seedfall Mice

5000

mice MNA

Seeds /m2

4000

b) Eglinton MR2

100

mice MNA

a) Eglinton MR1 5000

0

2001

Fig. Cumulative within-year beech seedfall and mouse population size on each of the grids in the Eglinton and Hollyford valleys from May 1999 to November 2001 (no rodent data was collected in May and August 1999 in MR1).

6

Vertebrate Pest Research

December 2002

Department of Conservation as

due to the presence of other

part of an ongoing native bird

seed-bearing plants not

protection programme, and stoat

found in the Eglinton Valley.

numbers there were very low. In

Mouse populations crashed

the nearby Hollyford Valley, no

in 2001, with none being

stoats were killed.

caught in either valley, when no beech seedfall occurred at

Overall, beech seedfall varied

either site.

sharply between years: in the Eglinton Valley, seedfall from the

Ship rats were present on all

red beech was high in both 1999

four grids in most quarterly

and 2000, although the amount and

trapping sessions, with up to

its timing varied between the grids

six caught over 5 nights of

(see figure). In the Hollyford Valley,

trapping in Eglinton Valley, and up

decline in mouse populations

the seedfall from the silver beech

to 17 ship rats (and/or kiore) in the

indirectly, by influencing their

was high only in the second year

Hollyford Valley. Although the

foraging efficiency, although this

and was markedly more abundant

study was not designed to index

remains unproven. In contrast to

on one (MR2) of the two grids.

their abundance, stoats were also

stoats, rat diet studies have yet to

trapped in both valleys.

show rats actively prey on mice.

followed the pattern of seedfall

Overall, Wendy and her team

The study was funded by the

(see figure). In the Eglinton Valley,

believe the differences in mouse

Foundation for Research, Science

mouse numbers were highest in the

numbers between the Eglinton

and Technology.

months and years of highest

and Hollyford valleys are most

seedfall at each site. In the Hollyford

likely related to beech seedfall

Valley, however, mouse numbers

and not to differences in rat or

were less predictable. They were

stoat numbers. Following a high

highest on the grid with greatest

seedfall event and high mouse

seedfall (MR2), despite both grids

numbers in winter and spring,

being only 2 km apart, and were

mouse populations decline to

present in moderate numbers on

reach low levels in late summer–

both grids in 1999 when there was

autumn. At this time, stoat numbers

virtually no seedfall, presumably

are at their highest, with the

Wendy Ruscoe processing live-captured rodents in the Hollyford Valley.

Mouse population changes largely

increase due to their young being weaned in JanuaryFebruary, and entering the population during the mouse population decline. High house mouse

An ear tagged mouse about to be released.

numbers in the previous

Wendy Ruscoe works on the

year gave rise to pregnant

ecology of small mammals in

stoats in good condition,

beech forests; Ivor Yockney and

and high birth and juvenile

Richard Heyward work on the

recruitment rates. Stoats

ecology and management of a

may, however, exacerbate a

wide range of vertebrate pests.

7

Kararehe Kino

December 2002

Sust ain ed F eral Goat Control – Mt Egmont Nation al Park Sustain ained Feral National

F

eral goats are a widespread, abundant conservation pest in

New Zealand. The idea of eradicating them is intuitively appealing, although it is likely to be more difficult and more expensive than their sustained control, at least in the short term. A team led by Dave Forsyth has been analysing annual hunting effort and kills of goats in Egmont National Park to establish whether sustained control to low densities could be changed to eradication. Goats have been present in the park

A feral goat photographed foraging in an introduced grassland.

since about 1910 and controlled 5000

annually since 1925 in one of the longest sustained vertebrate pest control programmes in the world. park records between 1961 and 1999. During the winters of 1961– 1964, the New Zealand Forest Service (NZFS) employed four hunters with dogs and rifles to control goats. Subsequently, NZFS hunters undertook control over

4000

Number of goats killed

Dave and his colleagues analysed

3000

2000

1000 9000 goats

summer from 1964 to 1987. This effort has been continued by hunters working for the Department of Conservation (DOC). Helicopter-based control

Park Board NZFS/DOC Recreational Honorary Rangers

0

1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 Year

Fig. 1. Annual kills of feral goats at Egmont National Park, 1944–1999.

has been used only occasionally since 1971 and has taken about

Annual hunting effort in the park

per hunter- per day; Fig. 2b). Kills

150 animals. Honorary rangers and

was initially low, but steadily

declined to less than one kill per

recreational hunters have had

increased to a maximum of over

hunter per day in 1986, and have

unrestricted hunting access to

1400 hunter days in both 1983

remained at low levels since then.

goats in the park since 1955. They

and 1984 (Fig. 2a). Hunting effort

These figures indicate a large

contributed a large proportion of

then declined to a low in 1994 and

decline in the abundance of goats

the kill until 1965, but almost

stabilised thereafter. Goat kills also

in the park from 1961 to 1999.

none since 1970. Including the

declined with time (Fig. 1), with the

9000 goats known to be shot

number shot per hunter per day (a

Dave and his team argue that three

during 1925–1943, at least

crude index of the total number

conditions are essential before goats

96,900 goats have been killed in

present) highest when NZFS

can be eradicated from the park.

Mt Egmont National Park (Fig. 1).

hunters started in 1961 (seven kills

Firstly, all immigration must be

8

Vertebrate Pest Research

December 2002

1500

prevented. Goats are currently farmed on several properties around the park. One expensive management option is erecting goat-proof fences around the park. A considerably cheaper and probably more effective option would be to

1250 Number of hunter days

Mt Egmont, and some escape into

restrict the farming of goats within

1960

hunters, but it is not clear whether

Kills per hunter per day

targeted by helicopter-based

1980

1985

1990

1995

2000 (b)

4 3 2 1 0

exceed their rate of increase from

1975

5

living in the subalpine shrubland.

Thirdly, the harvest of goats must

1970

6

this could effectively reach all goats Further research is needed.

1965

7

likelihood of being killed. The team

the alpine grasslands could be

500

8

so that all animals face a strong

hunting with dogs. Goats living in

750

0

Secondly, all goats must be targeted

forest are at risk from ground-based

1000

250

a ‘buffer’ zone around the park.

believes that goats living in the tall

(a)

1960

1965

1970

1975

1980

1985

1990

1995

2000

Year Fig. 2. Trends in (a) hunting effort, and (b) the number of feral goats killed per hunter per day, during 1961–1999.

breeding and immigation. A simple mathematical model suggested

Dave and his team consider 12

is now in a position to consider

that the current population of

years the maximum time frame for

the eradication of goats from

goats in the park is about 1050,

an eradication attempt. Provided

Mt Egmont National Park and an

and that removing 50% of them

the above conditions can be met,

end to over 75 years of sustained

annually would achieve eradication

culling more than 90% of the goat

goat control.

in >50 years while removing 90%

population annually is quite a

annually would achieve eradication

sobering challenge for even the

This work was funded by the

in 12 years.

most stout hearted. That said, DOC

Department of Conservation.

Dave Forsyth and John Parkes work on the ecology and management of mammalian pests; Jim Hone is Associate Professor at the University of Canberra and researches the management of vertebrate pests; Garry Reid and Dean Stronge work for the Department of Conservation.

9

Kararehe Kino

December 2002

Are We- t a- P opulations Aff ected By 1080? Populations Affected

W

e-ta- are potentially at risk

after bait application, and then

from 1080 poisoning for

again at monthly intervals for the

possum control as several species have been observed eating toxic baits. Also some we-ta- collected alive after 1080-poisoning operations have

next 4 months. Eric and Peter found the Wellington tree we-ta- and a species of cave we-ta-

contained residues of 1080. To date, we-ta- populations that have

in the refuges, as well as a wide

been monitored in poison baited

the numbers of most of these

areas have not been affected by

invertebrates using the refuges

the toxin. These results are open to

increased steadily over 15 months

challenge, however, because the

of monitoring in both the poisoned

methods used to monitor impacts

and non-poisoned plots (Fig. 2).

have not included individually marked we-ta-. For example, one

The bait application had no impact

study recorded we-ta- calls heard at night and another the number of we-ta- caught in pitfall traps.

range of other invertebrates, and

on the numbers of either species of we-ta- or on the numbers of slugs, spiders, and cockroaches

Fig. 1. Artificial refuge used for monitoring we-ta- populations (lid ajar to expose galleries).

(the main other invertebrates occupying the refuges).

To confirm the risk to we-ta- (or not), Eric Spurr and Peter Berben

8

monitored individually marked we-ta- occupying artificial refuges 1080-poisoning (i.e. baits spread by hand). To do this, they set up 10 randomly located artificial refuges (Fig. 1) in each of 20 plots spaced at

Tree we-ta- per plot

before and after simulated aerial

1080- poisoning

Poisoned plots

7 6

Non-poisoned plots

5 4 3 2

least 50 m apart on a north-facing

1

ridge in Tararua Forest Park in

0

August 1999. From October onwards, the refuges were checked monthly for occupancy by we-ta- and

8

other invertebrates, and any tree we-ta- present were individually

6

August 2000, 10 of the plots, chosen at random, were sown by hand with 1080 bait at 5 kg/ha.

Cave we-ta- per plot

marked with coloured paint. In

4 3 2 1

lured, Wanganui No.7 cereal-based

0

1080. The remaining 10 plots were not baited. The artificial refuges were checked for occupancy by we-taand other invertebrates a week

10

1080- poisoning

Non-poisoned plots

5

The bait was green-dyed, cinnamonbait containing 1500 ppm (0.15%)

Poisoned plots

7

A

S

O N 1999

D

J

F

M

A

M J 2000

J

A

S

O

N

D

Fig. 2. Number of we-ta- occupying artificial refuges in poisoned and non-poisoned plots, before and after the experimental 1080-poisoning operation; (a) tree we-ta-, (b) cave we-ta- .

Vertebrate Pest Research

December 2002

% marked tree we-ta- re-sighted

90 80

Poisoned plots

70

Non-poisoned plots

60 50 40 30 20 10 0 1

4

8

12

16

Weeks after experimental 1080-poisoning operation Fig. 3. Percentage of individually marked tree we-ta- resighted in poisoned and nonpoisoned plots. Tree we-ta- in gallery of refuge.

One week after spreading bait, 80% of the 56 marked tree we-ta- were

mortality, predation, loss of paint

resighted alive in the poisoned

artificial refuges into natural refuges.

to affect the population numbers

plots and 72% of the 46 marked tree we-ta- in the non-poisoned plots. The number of marked tree we-ta-

There was no evidence that it

of the slug, spider, and cockroach

resulted from 1080-poisoning.

species recorded in the artificial

resighted alive declined over the

The study indicates that aerial 1080-

restricted to one area in Tararua

next 4 months, but the rate of

poisoning for possum control is

Forest Park, there is no reason to

decline was similar in both the

unlikely to affect the population numbers of Wellington tree we-ta- or

believe that the results would be

poisoned and non-poisoned plots (Fig. 3). Eric and Peter presume the

of one species of cave we-ta-. Aerial

of invertebrates exposed to 1080

decline was a result of natural

1080-poisoning also appears unlikely

baits elsewhere in New Zealand.

markings, and movement out of the

refuges. While the study was

different for these or related species

This work was funded by the Foundation for Research, Science and Technology, and is modified from an article submitted for publication in ConScience.

Eric Spurr and Peter Berben work on the effects of 1080 baits laid for possum control on vertebrate Cave we-ta- in gallery of refuge.

and invertebrate non-target species.

11

Kararehe Kino

December 2002

Some Ecological Limit ations of Predator Con trol Limitations Control threatened populations of

some native prey species from predators is often blamed on the technical limitations of control methods to kill enough predators over the required area at the critical time. However, ecological factors can also limit the effectiveness of predator control and Grant Norbury and John Innes have been examining some of these. Firstly, the relationship between a

% nesting attempts fledging young

T

100

he failure to protect

80

60

40

20

0 0

10

20

30

40

50

60

70

Possum abundance

predator’s abundance and its impact on prey populations (the ‘damage function’) is seldom linear

Fig. The nesting success of ko-kako and ku-kupa in relation to possum abundance (% trap catch).

and can be strongly curved, e.g. nesting success of ko-kako and ku-kupa in the face of predation by possums (see figure). For these birds, possum control provides little protection for eggs and nestlings unless the possum population is reduced to levels indicated by trap catches of