Fiordland tokoeka: Walk-through survey of kiwi in the Clinton Valley

Full text of publication

Summary

This report summarizes the results of a walk-through survey of the kiwi population in the Clinton Valley, on the Milford track, funded by Air New Zealand. Walk-through surveys are usually conducted to provide site records in an area where kiwi density is low or their presence is unknown. The method is used in the Clinton Valley to assess whether the known kiwi population in a designated area is stable, increasing or declining every 5 years.

This is the second walk-through survey conducted in the valley, the first was in 2010, where 43 birds were detected.  A total of 49 kiwi were heard, seen or caught along 17km of the track in the Clinton Valley during March/April 2015. The results indicate the population is stable and may have increased, the survey effort was a lot less in 2015, and two juvenile kiwi were caught, approximately 6 months old.

Introduction 

The Milford track is one of four Great Walks where biodiversity enhancement projects have been funded through a partnership between Air New Zealand and the Department of Conservation. The project’s primary threatened species foci are whio (Hymenolaimus malacorhynchos), and pateke (Anas chlorotis) supporting efforts to re-introduce the species primarily through stoat control. Fiordland tokoeka (Apteryx australis australis) also inhabit the area. Studies have shown that juvenile mortality in kiwi is as high as 94%, with stoats causing the majority of these losses (McLennan et al. 1996).

Stoat control on the Milford track was initiated in the Clinton Valley in 1999 and the Arthur Valley in 2003, primarily to protect whio. With additional funding from the Air New Zealand Great Walks Biodiversity Project there are now 833 stoat trap tunnels along the Clinton and Arthur valley including the side branches that are also trapped such as the Clinton North Branch, Neale Burn, and Joe’s Valley.  Aerial 1080 operations to target possums were carried out over 2 years in the Clinton North Branch, Neale Burn and true left of the lower Clinton valley (2005); and the upper Clinton Valley and true right of the lower Clinton Valley (2006). An aerial 1080 operation targeting rodents was undertaken in the main Clinton valley in 2014. There may be another 1080 operation in the valley in spring 2016 if seedfall and rodent tracking levels are reached. 

A kiwi productivity and survival study was conducted in the Clinton Valley from 2001 to 2005, using telemetry and 24hr video recording at nests. At the end of the project, data suggest that one more adult death in the population would have resulted in a prediction of a declining population. Adult survivorship was found to be high with a lifespan estimate of 63 years, from 63 transmitter-years and 1 death (Hugh Robertson pers.comm) Larger data sets suggest an average lifespan of 50 years.

Subsequent work in the Murchison Mountains found a much lower adult survivorship from more data (lifespan estimate of 36 years from 146 transmitter-years and 4 deaths) (Tansell, unpublished data 2009). A territory map was compiled from all known birds at the end of the Clinton study and bands were left on individuals. In order to check the population status, a decision was made to conduct a walk-through survey every 5 years. This method was deemed most appropriate as there are a high proportion of known birds in known territories in the area, therefore can be used as a presence/absence marker, also the method of call count monitoring was found to be difficult in the area due to river noise. 

Survey objectives

The objectives of this research are to:

• estimate the number of kiwi in the study area
• confirm known birds still occupy known territories
• where possible catch kiwi to remove metal bands and insert transponders.

Study area 

The Clinton Valley is situated at the head of Lake Te Anau in Fiordland National Park. The Clinton River flows through this glacial U-shaped valley into Lake Te Anau. Temperate rainforest in the Clinton Valley is predominately silver beech (Nothofagus menziesii) with red beech (N. fusca) in the lower valley and mountain beech (N. solandri var. cliffortoides) at higher altitudes. The bird life is relatively intact with notable species including kiwi, whio, kea (Nestor notabilis), kaka (Nestor meridionalis), weka (Gallirallus australis) and käkäriki (Cyanoramphus spp). The study area is approximately 2520 ha.

Methods

Walk-through survey

The walk-through survey method is based on best practice (Robertson & Colbourne, 2003). Two teams of two people spend the night time walking along the track and at every second trap box play a male kiwi call from a Foxpro caller, wait for 1 minute then play a female call and wait for a further 5 minutes. When a kiwi was heard a GPS waypoint was taken and a compass bearing toward the sound recorded, as well as an estimate of distance.

An attempt was made to capture all birds that closely approached. Any birds captured that had been previously banded had the band read then removed and an Allflex transponder inserted. Un-banded birds had transponders inserted also. Bill measurements and weights were taken from most kiwi caught. 

Since specific details about the previous survey were unavailable we split the survey area into four sites. On the first night one team surveyed along the track from trap number C072 to C090, while the other team surveyed from C072 to C054. On the second night teams swapped sites, however there was a significant reduction in calls therefore we decided to survey new areas the third night and not to swap areas the following night, cutting the survey short by one night. The other two sites were from trap number C054 to C038, then from C038 to C021. We did not survey the lower North Branch.  The total area covered was approximately 17km along the track. 

We assumed that all birds within 200m were heard and that there was no movement of previously banded territorial adults out of their territory .We assumed no band loss had occurred.

Transponders

Passive Integrated Transponders (PIT tags) were inserted into the flank under the right wing, as recommended by best practice (Robertson & Colbourne, 2003). We used Allflex transponders and read them using a hand held automatic Allflex transponder reader.

Kiwi records

A total of 49 individual kiwi were heard or seen during a total of 37 hours in approximately 2520 hectare area surveyed from 30 March to 2 April 2015. (see appendix 1 for transcription of field notes). They were recorded as 14 male-female pairs, 29 males, 18 females, and two juveniles (Table 1). Of these, 36 records were from calls, 4 were seen, and 9 captured.

The teams consisted of 4 experienced kiwi practitioners: Hannah Edmonds, Jane Tansell, Pete Kirkman and Pete McMurtrie.

Table 1. Kiwi recorded in the core study area of the Clinton Valley in the three study periods, note September 2005 was the telemetry study and all 40 birds were marked. April 2010 and 2015 were walk-through surveys

	September 2005	April 2010	April 2015
Total number of individuals	40	42	49
Total number of males	22	27	29
Total number of females	18	15	18
Number of pairs	16	13	14
Number of single males	6	14	15
Number of single females	2	2	4
Juveniles	0	0	2

Kiwi captures

Table 2. Details of the kiwi caught in the Clinton Valley, Fiordland in March/April 2015

Date	ID	Age/ Sex	Weight (kg)	Bill (mm)	Band combo	Transponder	Location	Notes
30/03/15	New	Ad/Fem	*	118.5	-	982 000363410362	1189540 5021439	Juv in territory
30/03/15	Jess	Ad/Fem	*	125.3	RA-0827	982 000363438768	1189055 5021849
30/01/15	Bruce	Ad/Male	*	-	R-31561	982 000363437978	1188700 5022229	Juv in territory
30/03/15	New	Ad/Fem	*	129	-	982 000363438094	1188570 5022399
30/03/15	New	Juv	*	62.5	-	982 000363438455	1188720 5022199	In Bruce’s territory
30/01/15	New	Juv	*	60.0	-	982 000363424752	1189430 5021489	In first female’s territory
30/01/15	Hannah	Ad/Fem	3kg	133.4	RA0604	982 000363387458	1191360 5019299
30/01/15	Spaz	Ad/Male	2.5kg	103	-	982 000146311151	1191380 5019319	Had transponder already
01/04/15	Murray	Ad/Male	2.5kg	103.5	-	982 000145488783	1196395 5016819	Had transponder already

*weights were not taken as second pair of scales accidently ended up in lower valley

Kiwi locations

The locations of individual kiwi were based on sightings and bearings taken to calls and subsequently marked onto a map (appendix 2). An attempt was made to distinguish between individuals heard/seen by both teams to ensure the same birds were not recorded twice.

Discussion

The walk-through survey methods used this year differed slightly to the methods used in 2010.  The sections of track surveyed each night in 2010 were not defined in the report, therefore for the 2015 survey we pre-determined sections using trap box numbers and planned to survey 4 sections with each team covering each section once, so each section was covered twice by the two different teams.

During the second night of survey we realized that the second team did not hear or see many new kiwi, hence we covered the two remaining sections with only one team from then on. This led to a reduction in hours of night survey, from 51 hours in 2010 to 37 hours in 2015. Even with these reduced hours we detected more kiwi, (42 in 2010, 49 in 2015).

We did not survey the North Branch because only one pair “Mooch” and “Smole” were heard there in 2010, and the other birds “Lockers” and mate and “Kapai” can be heard from the main track on the true right. We did not use transmitters to find mates during this survey as none of the four transmitters used in 2010 led us to find more birds. We did not use kiwi dogs to find kiwi during the day as we discovered in 2010 it was too taxing to work both day and night. We did one less night of survey than in 2010.

The assumption that all birds within 200m were heard is true for the majority of the sections searched, however several areas above and below the Clinton Forks have loud river noise making it difficult to hear kiwi. We had initially thought we could hear birds 700m away, and although we heard some kiwi 400-500m away, an assumption of 200m is more accurate. Kiwi may be far from the track when observers are listening, they may not respond to calls, especially females which typically do not call as often and are more difficult to hear over large distances. Several known kiwi from known territories were not heard during this survey. As with the 2010 survey, the total number of kiwi detected will be an underestimate of the population in the area due to the fact not all kiwi will call during the survey.

One of the birds caught, “Jess” was originally known from a territory on the true left of the river, however she was caught on the true right in what was known as “Elvis” and “Pricilla’s” territory, confounding the second assumption.

Catching two juvenile kiwi was very promising. Bill measurements suggest the two birds are approximately 6 months old and therefore close to reaching stoat safe weight of over 1kg. During the productivity and chick survival study in the Clinton chick survival to sub-adult was only 17.6%. The increase in trapping effort and several 1080 operations since the study was conducted may be enhancing chick survival, and may well be indicative of other areas in Fiordland which incorporates the aforementioned predator control regimes. We could expect to see an increase in kiwi during the 2020 walk-through survey if the current predator control remains the same or is increased. The results of this walk-through survey are an indication of the population trend of a sample of kiwi in the valley. It is not as robust a method as a long-term telemetry study investigating productivity, chick and adult survival.

Recommendations

The walk-through survey method should be undertaken in April 2020, during fine weather, surveying each of the four sections specified in the methods section, using experienced kiwi practitioners. Any kiwi close enough to be able to be caught should be, any bands removed and transponders inserted. A map should be taken so that teams can identify where they heard/saw kiwi to make it easier to distinguish between double ups.

Current predator control regimes should be maintained or enhanced to benefit kiwi and other endangered species in the area.

References

Edmonds, H. 2005. Fiordland Tokoeka: Productivity and Chick Survival Study, Clinton Valley, Fiordland National Park, 2004/2005. September 2005.  Unpublished report, Department of Conservation, Te Anau

McLennan, J.A, Potter, M.A, Robertson, H.A, Wake, G.C, Colbourne, R, Dew, L,

Joyce, L, McCann, A.J, Miles, J, Miller P.J, Reid, J. 1996, Role of predation in the decline of kiwi Apteryx SPP. In New Zealand. New Zealand Journal of Ecology (1996) 20(1): 27-35

Robertson, H; Colbourne R. 2003. Kiwi Best Practice Manual DOC Internal Report

Tansell, J. 2009. Tokoeka Monitoring Murchison Mountains 2009/2010. Unpublished report, Department of Conservation, Te Anau

Acknowledgements

This survey was made possible through funding from Air New Zealand. Thank you to Erina Loe for logistical support, and the team Pete Mc Murtrie, Pete Kirkman, and Jane Tansell. Thank you to Em Oyston for review of this report . 

Appendices

Appendix 1. Field trip notes transcribed

30 March 2015

Hannah and Pete Mc:

• 8.35 Male 60° 400m
• 9.40 Male 50° 500m
• 9.50 Male 145° 400m
• 10.10 same male 130° 300m
• 10.20 Female Hannah caught. Band RA 0604, Wgt 3.kg, Bill 133.4 Transponder 982 000363387458
• 10.15 Male heard while catching female 160° 80-100m – then caught Spaz. No band. Wgt 2.5kg, Bill 103. Transponder 982 000146311151
• 11.33 Male 160° 80-100m. Female heard 20m off track
• 11.37 Male and female 40° 400m
• 1.10 Male 170° 50-80m
• 1.45 Male 170° 20-30m Just upstream from Big Slip
• 1.25 Male 210° 100m High up on down- stream side of Big Slip
• End 3am

Jane and Pete K:

• 20.59 Male 10° 150m
• 21.18 Male 317° 500m
• 21.18 Male 243° 200m
• 21.32 Male 300° 400m
• 22.00 Male 1 bar heard only 100m
• 22.12 Male 100° 150m
• 22.12 Male 328° 100m mate caught
• 22.14 Male and female 160° 300m
• 22.35 Female 355° 50m caught No band or transponder. Bill 118.5. Transponder 982 000363410362. One bar of strange sounding juv heard several times but couldn’t pin direction
• 20.50 Male 075° 400m
• 23.55 Male 234° 150m. Bus stop
• 12.57 Male 247° 200m Female 232° 150m Caught 01.08am Band RA 0827 New transponder 982 000363438768 Bill 125.3. Male seen but wouldn’t approach
• 01.47 Female 298° 150m
• 02.05 Male approached caught Band R-31561 New transponder 982 000363437978
• 02.25 Female 005° 200m
• 027? Male 220° 300m
• Female “Honker” caught Bill 129.0 New transponder 982 000363438094
• Juvenile caught Bill 62.5. New transponder 982 000363438455
• 03.57 Female 299° 150m
• 04.55 Male 027° 400m
• End 5am

31 March 2015

Hannah and Pete Mc:

• 20.30 Male 180° 500m on cliff? 
• 21.01 Male and Female 20° 400-500m Heard from Pomp hill back toward Prairie
• 23.39 Male up from Marlene’s Dave? 230° 400-500m
• 23.41 Male 130° 50m middle of Marlene’s
• 23.58 Male 30° 200m heard from Bus stop
• 00.20 Male and Female 240° 100m
• 00.28 Male 80° 200m
• 00.54 Male and Female 280° 150-200m Back towards Pomp hill
•  End 2.30am

Jane and Pete K:

• 20.10 Male 217° 300m
• 21.33 Female Hannah? 177 ° 150m
• 21.41 Male 350° 300m Hannah again
• 22.18 Male 150° 150m
• 22.35 Male 150m
• 23.13 Male 036° 300m Female 0.15° 350m
• 23.17 Male 305° 200m
• 23.34 Female 100° 300m
• 23.52 Female 225° 300m
• 00.15 Male 215° 50m
• 00.35 Male 003° 250m
• 01.33 Male 50° 400m
• End 3am

1 April 2015 

Hannah and Pete Mc

• 20.56 Male 210° 300m 
• 21.05 Female seen on track 
• 21.30 Female from 21.05 120° 50m – male also 
• 22.55 Male 150° 400-500m 
• 23.07 Male 250° 50m Tried to catch 
• 00.31 Male 22° 400m toward Neale Burn 
• 2.00? Male caught Murray 982 000145488783. Bill 103.5. Weight 2.5kg. Opposite North Branch crossing 
• End 2.30am 

Jane and Pete K

• 21.37 Male 350° 200m 
• 21.55 Female 192° 250m 
• 22.07 Male 020° 50m
• 22.38 Male 306° 100m
• 00.55 Male 140° 20m
• 01.15 Female 327° 150m
• 01.25 Male 350° 100m
• 01.53 Male 350° 500m
• 02.03 Male 322° 150m
• End 2.30am

Appendix 2. Map of kiwi locations, Clinton Valley April 2015 showing trap locations where survey starts and ends.

Kiwi locations, Clinton Valley (JPG, 369K)