What Is Causing the Decline of Little Penguins (Eudyptula Minor) on Granite Island, South Australia?

What is causing the decline of little penguins (Eudyptula minor) on Granite Island, South Australia?

Report to the South Australian Department for Environment and Heritage, Wildlife Conservation Fund and the Nature Foundation SA

Natalie M Bool1, Brad Page2 and Simon D Goldsworthy2

1 University of Adelaide, North Terrace Adelaide SA 5009 2 South Australian Research and Development Institute (SARDI), 2 Hamra Avenue, West Beach SA 5024

What is causing the decline of little penguins (Eudyptula minor) on Granite Island, South Australia?

South Australian Research and Development Institute SARDI Aquatic Sciences 2 Hamra Avenue West Beach SA 5024 Telephone: (08) 8207 5400 Facsimile: (08) 8207 5481 http://www.sardi.sa.gov.au/

Disclaimer Copyright South Australian Research and Development Institute 2007.

This work is copyright. Except as permitted under the Copyright Act 1968 (Commonwealth), no part of this publication may be reproduced by any process, electronic or otherwise, without the specific written permission of the copyright owners. Neither may information be stored electronically in any form whatsoever without such permission.

Printed in Adelaide, July 2007

SARDI Aquatic Sciences Publication Number F2007/000288-1 SARDI Research Report Series No. 217

Authors: Bool NM, Page B, and Goldsworthy SD

Reviewers: Dr Shane Roberts and Dr Adrian Linnane

Approved by: Dr. T. Ward

Signed:

Date: 12 July 2007

Circulation: Public Domain

Table of Contents 1 Executive Summary ...... 5 2 Background...... 6 3 Introduction ...... 8 Aims and approach of the project ...... 11 4 Methods ...... 12 Burrow monitoring ...... 12 Date of laying and fledging...... 13 Analysis of reproductive success ...... 13 Chick growth rate ...... 13 Census of the breeding population...... 14 Historical data...... 14 Foraging ecology...... 15 Data analyses...... 15 Spatial overlap in foraging area...... 16 Little penguin diet ...... 17 Diet of New Zealand fur seals ...... 18 5 Results ...... 21 Reproductive success ...... 21 Predation by rats ...... 21 Inter-annual breeding success at Granite Island...... 21 Inter-annual breeding success at West Island ...... 21 Chick growth...... 22 Population census during the breeding season ...... 23 Foraging ecology...... 24 Summary of foraging behaviour ...... 25 Foraging parameters...... 25 Oceanography...... 27 Overlap of foraging areas between Granite and West Island penguins...... 29 Penguin Diet...... 29 Size and weight of prey...... 31 Comparison of diet between colonies ...... 31 New Zealand fur seal diet...... 32 Number of New Zealand fur seals hauled-out...... 33 6 Discussion...... 34 Reproductive parameters...... 34 Impacts of tourism and predation by rats on little penguins ...... 36 Predation by New Zealand fur seals ...... 37 Foraging parameters...... 38 Similarity in prey diversity and meal sizes...... 40 Chick growth rates ...... 41 The future of the little penguin populations at Granite and West Islands...... 42 7 Conclusions...... 45 8 Recommendations for future research...... 47 9 References...... 48 10 Acknowledgements ...... 55

3 List of Figures Figure 1. The location of Granite and West Islands and other islands mentioned in the text...... 7 Figure 2. The number of active little penguin burrows during the breeding season reported at Granite Island from 2001-2006...... 24 Figure 3. Circular histogram of the mean direction of travel of breeding little penguins from Granite (blue) and West Islands (red) in 2006...... 25 Figure 4. Summary of the inferred foraging routes undertaken by little penguins at Granite Island (red) and West Island (blue) based on satellite tracking in 2006...... 26 Figure 5. The proportional time spent foraging in oceanic areas of 1 x 1 km grids by 10 little penguins from Granite Island. Proportional time spent in each grid is indicated by colour: where red represents regions where penguins spent more time followed by orange, yellow, green and finally blue areas where penguins spent relatively little time...... 26 Figure 6. The proportional time spent foraging in oceanic areas of 1 x 1 km grids by 8 little penguins at West Island. Proportional time spent in each grid is indicated by colour: where red represents regions where penguins spent more time followed by orange, yellow, green and finally blue areas where penguins spent relatively little time...... 27 Figure 7. The average number of New Zealand fur seals hauled-out at Seal Rocks, West Island and Granite Island between March-October 2006...33

List of Tables Table 1. Measures of reproductive success for Granite and West Islands from 1990 – 2006. Data from 1990 - 2005 were obtained from the Granite Island Little Penguin Monitoring Group (N. Gilbert and R. Brandle unpubl. data)...... 22 Table 2. Growth parameters of 12 little penguin chicks at Granite Island in 2006...... 23 Table 3. Growth parameters of 11 little penguin chicks at West Island in 2006...... 23 Table 4. Summary of the foraging parameters of little penguins at Granite and West Islands in 2006...... 28 Table 5. Summary data of the oceanographic characteristics of the areas that the little penguins foraged in from Granite and West Islands in 2006. ....29 Table 6. The prey consumed and their relative numerical abundance (NA) and frequency of occurrence (FO) in stomach contents of little penguins from the Granite and West Island colonies in 2006. The number of samples examined, the number with prey remains and the minimum number of individuals consumed were 37, 37 and 191 respectively...... 30 Table 7. The mean, minimum and maximum lengths (mm) and weights (g) of anchovies, southern sea garfish and pilchards consumed at Granite and West Islands...... 31 Table 8 The percent numerical abundance (NA) and frequency of occurrence (FO) of prey types found in New Zealand fur seal scats at Granite and West Islands. The number of scats examined, the number with prey remains and the minimum number of individuals consumed are 29, 29 and 98, respectively...... 32

1 Executive Summary

The number of little penguins at Granite Island has been declining since the early 1990s. Large numbers of tourists visit Granite Island to view the penguins and the island is inhabited by non-native black rats (Rattus rattus) and water rats (Hydromys chrysogastes) whereas nearby West Island does not have any tourists nor land-based predators. Data on the little penguin population size, breeding success, diet composition, foraging behaviour and predation by both rats and New Zealand fur seals (Arctocephalus forsteri) were collected at Granite and West Islands to determine the likely onshore or offshore factors that may be causing the decline in the Granite Island population. In 2006, the percentage of chicks that fledged was significantly less at Granite Island (37%) compared to West Island (54%). Although there was little overlap (9%) in the foraging areas used by penguins at each site, there were no significant differences in prey species consumed or the mass of stomach contents at Granite (45.0 ± 33.1 g) (mean ± SD) and West Islands (46.0 ± 35.0 g). Both sites showed a similar delayed onset of breeding in 2006 and population surveys in 2006 indicated that both sites might be in decline, suggesting that tourism and predation by rats are not solely responsible for the decline in the Granite Island population. New Zealand fur seals consumed adult penguins at both sites, indicating that they may be causing part of the decline at Granite and West Islands and accordingly, ongoing monitoring of fur seal diet is required.

5 2 Background

The little penguin population at Granite Island, South Australia is accessible by a causeway from the township of Victor Harbour, and forms an important local tourism attraction, with 30,000 tourists undertaking nightly guided tours each year. Since monitoring began in 2,000, the number of little penguins coming ashore have decreased by between 8-16 % per annum, and the number of breeding pairs has declined from about 800 in 2001 to 300 in 2005 (N. Gilbert unpubl. data; R. Morcom unpubl. data). The cause(s) of these declines are currently unclear, but are significant enough to jeopardise the sustainability of the little penguin tourism business on the island. In addition to significant numbers of tourists, Granite Island has populations of water rats (Hydromys chrysogastes) and introduced black rats (Rattus rattus). Cats (Felis catus) and foxes (Vulpes vulpes) are also occasionally found on the island.

Little penguins also breed on West Island (6 km west of Granite Island), where the population was estimated at be about 4,000 penguins in 1990/91, with 2,029 being banded between 1990 and 1992 (M. Waterman unpubl. data). The island is relatively unmodified and does not have any introduced predators, and it is not visited by tourists. The current status of the little penguin population there is unknown.

This study was undertaken at Granite Island (35º 65’ E, 138º 62’ S) and West Island (35º 37’ E, 138º 35’ S) from March to November 2006. Granite Island is a South Australian Recreation Park connected to the mainland by a permanent causeway, which facilitates access to the island (Fig. 1). The park receives up to 70,000 visitors per year (Department for the Environment and Heritage unpubl. data). The waters around Granite Island are characterised by shallow sea grass beds around the northern end of the island, with water depths being a maximum of 16 m. (N. Gilbert unpubl. data).

6 West Island is a marine reserve 6 km west of Granite Island (Fig. 1). Access to the island by the general public has been prohibited for 30 years (Shepherd and Womersley 1970). The waters on the southern side of the island drop to 29 m, with a maximum depth of 5 m on the northern slopes. There are no introduced terrestrial mammals on the island. The current number of little penguin breeding pairs on West Island is unknown. West Island is a haul-out site for non-breeding New Zealand fur seals. New Zealand fur seals also reside on both the nearby island of Seal Rocks, which is located 2 km south of Granite Island, and on a breakwater, which extends to the east of Granite Island.

Figure 1. The location of Granite and West Islands and other islands mentioned in the text.

The little penguin is a generalist predator, with clupeoid species being the preferred prey (Gales and Pemberton 1990; Cullen et al. 1992). During the breeding season, relatively short foraging trips of up to 20 km per day are undertaken (Klomp and Wooller 1988; Weavers 1992; Ropert-Coudert et al. 2006). In contrast, post breeding foraging trips in preparation for the annual moult can extend up to hundreds of kilometres from the breeding grounds

7 (Reilly and Cullen 1983; Weavers 1992). In South and Western Australia, the first clutch is typically laid around April, with second and replacement clutches occurring from August (N. Gilbert pers. comm.). Little penguins lay two eggs per clutch and double breeding is common (Johannesen et al. 2003). It typically takes 95 days from the lay date for chicks to fledge (Stahel and Gales 1987). Eggs are incubated for approximately 36 days and chicks fledge when they are around 59 days post hatching (Knight and Rogers 2004). Chicks are guarded by at least one adult for approximately three weeks following hatching, with foraging alternating between the two adults during the guard-stage. After this period both adults forage at the same time to provision their chicks.

3 Introduction

Seabirds are highly conspicuous components of the marine ecosystem and thus typically have a high conservation profile (Wanless et al. 1998). The decline of seabird populations has prompted numerous studies to examine the interactions between these apex predators and their environment in an effort to determine the cause of population decline (Bertram 1995; Catard et al. 2000; Rindorf et al. 2000). Investigations into seabird decline have highlighted that both natural and anthropogenic factors can affect seabird populations at sea and on land (Lewis 1981; Monaghan et al. 1992; Furness and Camphuysen 1997; Robinson et al. 2005). For example, populations of seabirds nesting in the Peruvian upwelling system have declined in response to guano harvesting and are further threatened because of reduced food availability due to physical oceanographic changes and pressures from commercial fisheries (Jahncke et al. 2004). Changes in food availability during the breeding season (Norman and Ward 1992), predation (Burger and Gochfield 1994), tourism (Giese 1996), habitat alteration, guano and egg harvesting (Reilly 1994), and oil pollution (Goldsworthy et al. 2000) can have short and long term impacts on seabird populations. Such pressures can lead to population decline, particularly if they reduce reproductive success, survival

8 and/or recruitment, as has been indicated by the decrease of some little penguin (Eudyptula minor) colonies in southern Australia (Dann 1992; Dann and Norman 2006).

Seabirds may be disproportionately susceptible to several threats because they have to cope with hazards on land and at sea (Edgar et al. 2005). Seabirds are marine animals whose foraging grounds are spatially separated from their breeding sites on land (Boersma 1998; Yorio 2000). This has led to the evolution of a number of life-history characteristics that are in tune with change in seabird’s marine environment (Yorio 2000). For example, seabirds are long lived, have delayed maturity, experience low fecundity and are often wide ranging (Yorio 2000), which means that populations are particularly sensitive to changes in ocean productivity and environmental perturbations such as the El- Nino and La Nina Southern Oscillation events (ENSO) (Lewis 1981; Barbraud and Weimerskirch 2001; Adams et al. 2004).

The marine environment is a dynamic system, in which food resources are patchily and widely distributed (Weimerskirch et al. 2005). However, the ability of seabirds to track their prey is restricted during the breeding season because of the need to return to land to provision young and defend nesting territories (Hunt et al. 1986; Ainley et al. 1998; Boersma 1998; Weimerskirch and Cherel 1998; Catard et al. 2000; Forero et al. 2002). The energetic costs of locating foraging grounds are critical elements to reproductive success (Numata et al. 2000). Therefore, it is possible to detect changes in food availability by monitoring changes in foraging habits using satellite telemetry and examining reproductive parameters (Chastel et al. 1993; Bryant et al. 1999; Rey and Schiavini 2005). Measures of breeding success therefore integrate information about the status of environmental conditions that affect the availability of prey near seabird breeding colonies (Safina et al. 1988; Diamond and Devlin 2003; Velarde et al. 2004; Rey and Schiavini 2005). However, interactions between predators and their prey are typically complex and tracking such relationships is difficult in a dynamic ecosystem (Botsford et al. 1997; Croxall and Wood 2002; McCann et al. 2003; Doherty et al. 2004).

9 Seabird populations are not only sensitive to fluctuations in prey but also to changes in predation rates (Long and Gilbert 1997; David et al. 2003). Seals are known predators of seabirds and their numbers are increasing around seabird colonies in Australia and Africa, following the end of exploitation by sealers (Wickens et al. 1992; Shaughnessy et al. 1995; Ainley et al. 2005; Mecenero et al. 2005; Page et al. 2005). It is possible that seabirds in these areas increased in number and distribution in the absence of large populations of seals (Crawford et al. 1989). However, now that many seal populations are recovering, predation on seabirds has become prevalent, such as in South Australia and in the Benguela ecosystem where New Zealand fur seals (Arctocephalus forsteri) and Cape fur seals (Arctocephalus pusillus pusillus) prey on penguins, gannets and cormorants (David et al. 2003; Page et al. 2005; Johnson et al. 2006). Although predators may have negative impacts on seabird populations, quantifying these impacts is challenging, because it is difficult to obtain information on the nature of the interactions between seals and their prey, which occur at sea (Croxall and Prince 1987; Descamps et al. 2005).

On land, seabirds often form large colonies because suitable nesting areas without large numbers of predators are sparse (Burger and Gochfield 1994). Seabirds are vulnerable on land to predators because they are less agile there due to their physical adaptation to foraging at sea (Crawford et al. 1989). Little penguins and other seabirds are particularly vulnerable to introduced predators, which may be able to enter burrows to eat eggs and chicks (Dann 1992). The introduction of exotic predators, in particular rats (Rattus spp), has led to the extinction of numerous animal species on islands (Lyver 2000; Martin et al. 2000; Stapp 2002; Votier et al. 2004; Descamps et al. 2005).

The current distribution of the little penguin, is smaller than indicated by historical records, possibly because of habitat alteration and predation (Stahel and Gales 1987; Gales and Pemberton 1990; Reilly 1994; Rogers et al. 1995). For example, little penguins recently became locally extinct at two sites in Victoria, probably because of predation by red foxes (Vulpes vulpes) and

10 domestic pets (Dann and Norman 2006). In addition, a breeding colony of little penguins on Kangaroo Island became extinct following the expansion of New Zealand fur seals into their breeding area between 1992 to 2000 (Shaughnessy and Dennis 2001; Page et al. 2005). Consequently, little penguin colonies in the vicinity of fur seal colonies and their haul-out sites and those little penguin colonies, which are accessible to introduced predators, are threatened by predation (Dann 1992; Weerheim et al. 2003).

At several breeding colonies throughout their range in southern Australia and New Zealand, little penguins attract large numbers of tourists, which are of significant importance to local economies (Dann 1992). However, disturbance caused by tourism can adversely affect penguin breeding success (Yorio et al. 2001; Carpenter et al. 2004) by causing adults to desert their nests (Bolduc and Guillemette 2003), or by causing chicks to receive fewer meals (Wilson et al. 1991). For example, little penguins breeding in areas disturbed by tourists had lower breeding success compared to those breeding in undisturbed areas on Penguin Island in Western Australia (Klomp et al. 1991).

Aims and approach of the project The aim of this study was to investigate factors that may be causing the decline in the little penguin population at Granite Island. The approach was to compare the breeding success and foraging ecology of little penguins breeding at Granite and West Islands, and determine if differences between sites could be explained by land-based factors including rat predation and tourism, or at sea factors including prey availability and predation by New Zealand fur seals. The numbers of breeding pairs at each site was also surveyed in order to determine if recent declines in numbers were restricted to the Granite Island population, or were part of a general decline in little penguin numbers in the region.

4 Methods

Burrow monitoring Data on little penguin breeding were collected on Granite and West Islands from May to November 2006. Active breeding burrows were identified by the presence of adult birds with eggs or chicks. Burrows were visited approximately every fortnight on Granite Island, but less frequently on West Island. In all, 73 burrows on Granite Island and 42 burrows on West Island were monitored. Each burrow was numbered and marked with flagging tape for individual identification.

The sex of adults was determined by measurement of bill length (vertical thickness of the bill at the nostrils) and bill width (lateral thickness of the bill at the nostrils) using callipers and equations developed by Gales (1988). Birds were weighed in a cloth bag using a Salter spring balance (2kg x 10 g). Small chicks were weighed in a smaller cloth bag using Salter 200 g x 1 g hanging scales. Individuals were tagged by inserting glass encapsulated TIRISTM (Texas, USA) transponders, which are individually encoded, subcutaneously midway down the back. Tiris transponders were read using an AllflexTM reader. Chicks were tagged once they had reached seven weeks of age and/or had lost ≥90% of their mesoptyle down (Johannesen et al. 2003). To determine whether rats were preying on chicks, all recently-killed chicks, which were found dead in or near a burrow, were collected and examined for signs of predator tooth marks.

12 Date of laying and fledging The date of lay was determined by either recording the presence of the egg if the burrow was checked within two weeks of the egg being laid or was calculated by using the hatching or fledging date, following the methods of Knight and Rogers (2004). Hatching date was recorded if the chick was seen piping its egg or if the chick still had its eyes closed. If hatching was not recorded and the chick was observed to fledge, the date of lay was ascertained by back counting 95 days from the fledging date. A chick was determined to have fledged if it was greater than 7 weeks of age and had >90 % of its adult plumage.

Analysis of reproductive success Four measures of breeding productivity were calculated: 1) hatching success - the number of chicks hatched per egg laid; 2) fledging success - the number of chicks fledged per chick hatched; 3) egg success – the number of chicks fledged per egg laid, and 4) the number of chicks fledged per breeding pair (Rogers et al. 1995).

Chick growth rate Growth parameters were calculated for chicks that had either fledged or were very close to fledging (Rogers et al. 1995). Growth parameters were not calculated for chicks in previous years because neither laying, hatch nor fledge dates were available. The individual chick growth data were summarised by fitting a logistic curve, because this was found to be the best fit for penguin chick growth rates (Wienecke et al. 2000).

The logistic curve here is based on the one presented by O’Connor (1984) and Wienecke et al. (2000): A S = Equation 1 1− e−k(a−t ) where S is the parameter measurement, body mass (g) at age “a” days since hatching; A is the asymptote where growth ceases, e is the base of the natural logarithms, k is the rate at which the data drives the curve to asymptote, and t is the age at which maximum growth occurs.

Census of the breeding population To provide an indication of the size of the breeding population on both islands a population census was carried out.

Granite Island To coincide with the peak of nesting activity, the census was carried out on August 11 and August 18, four weeks after several nests with eggs were found (following the methods of Gilbert unpubl. data). Burrows were marked with talcum powder to avoid double counting. The status of each burrow was recorded as 1) active but empty (denoted by the presence fresh guano and/or, tracks, nesting materials, recent burrow excavations), 2) burrow active (with adult and or eggs and chicks present), 3) inactive, presence of cobwebs and the absence of the above criteria. The annual rate of decline was calculated by fitting an exponential curve to the census data.

West Island The census on West Island was carried out over a period of two consecutive days (September 20-21). Burrow status was determined using the same criteria as those employed on Granite Island.

Historical data To compare breeding success across locations and years, data were obtained from the Granite Island Penguin Monitoring Group (N. Gilbert and R. Brandle unpubl. data). Granite Island annual census data since 2001 are also used in this study to compare the number of active burrows during the breeding season across years (N. Gilbert unpubl. data).

14 Foraging ecology To determine the location of foraging areas used by penguins from each colony, satellite transmitters were deployed on adult birds that were rearing chicks. The Cricket KiwiSat 202 (Sirtrack, Havelock North, NZ) satellite transmitters are embedded in epoxy resin, are hydro-dynamically shaped and weigh 28 g, which is less than 5% of a penguin’s body mass. Transmitters have a saltwater switch that deactivates the transmitter when it is submerged. Transmitters were glued directly onto feathers on the centre of the penguins back, to coincide with its centre of gravity (Healy et al. 2004), using Loctite 401 glue, and were further secured with a single cable tie. Once the device was attached, the penguin was released back into its burrow. Penguins were recaptured on shore following a single foraging trip and the tracking device was removed. If the penguin was caught before it returned to its burrow, its catch was removed from its stomach using the water offloading technique (Wilson 1984; Gales 1987) (refer to diet section, below).

Data analyses Satellite derived location data from transmitters were provided through Service Argos Inc, France. Accuracy of the location data is categorised into six classes 1) Class 3: <150m, 2) Class 2: 150-350m, 3) Class 1: 350- < 1000m 4) Class A, and 5) B have no accuracy assigned, 6) Class Z: no location calculated (Sterling and Ream 2004). The inaccuracy of Z positions, resulted in these being discounted from analyses. Position data were filtered to exclude location points which exceeded the maximum horizontal speed of 8.02 km/h for little penguins, using the R statistical software (version 2.0.1, R Development Core Team, R Foundation for Statistical Computing, Vienna) and the Time-Track package (version 1.0-9, M. D. Sumner, University of Tasmania, Hobart). This maximum horizontal speed is based on travel speeds of little penguins in South Australia (A. Wiebkin unpublished data). Time-track also extrapolated new positions between the actual locations based on a constant swim speed between locations, so that there was a position at 15 minute intervals.

15 The foraging behaviour of each penguin was summarised to describe several behavioural parameters: (1) the total distance travelled, (2) the maximum distal point travelled in a straight-line from the colony (3) the mean bearing travelled from the colony, and (4) the average speed km/h travelled (the distance between interpolated locations, divided by the duration).

To describe the differences between penguin foraging in relation to physical oceanographic parameters the: (1) mean and (2) median bathymetry, (3) mean bathymetric gradient (change in depth in metres for each horizontal kilometre), (4) median directional bearing of the bathymetric gradient (slope) were calculated using the MapInfo program (MapInfo Corporation, Troy, NY, USA) and the Vertical Mapper (GIS software) extension. The bathymetry values were calculated based on the distance to the nearest nodes and were assigned to each 15 minute interval. Bathymetry values were interpolated based on the distance to the nearest bathymetry reading, taken from 1 km by 1 km depth readings (GeoScience Australia).

Spatial overlap in foraging area To examine the degree of spatial overlap in the foraging areas used by birds from Granite versus West Islands, the proportions of time spent in each 1 km grid cell was explored using the niche overlap index (O), developed by

n Schoener (1968): Op=−10.5 ×∑ 12j −pj Equation 2 j=1

Where p1j and p2j are the proportion of time spent in the jth 1 x 1 km grid cell by birds from each colony. If the 2 groups foraged in the same areas in identical proportions, the index would equal 1. If entirely different areas were used, the index would equal 0. One penguin which did not return to the colony, was both included and excluded in this analysis, to examine whether its foraging trip influenced the extent of spatial overlap between birds from the two colonies.

16 Little penguin diet To obtain diet samples, birds were caught as they returned to the colony. Penguins were selected at random and as a result their breeding status (breeding or non-breeding bird) was not known. Birds were held in an enclosure until they were stomach flushed following the techniques outlined by Wilson (1984) and Gales (1987). This was done by inserting a duodenal tube Unomedical TM 6 mm diameter into their stomach to apply water using a hand-held pump. After being lavaged the penguin was fed with 40 ml of a blended pilchard mix and was given 40 ml of Vy-Trate TM diluted to 10% to prevent dehydration. Birds were kept in the enclosure for 30 minutes to ensure they were not harmed by the procedure. Each bird was weighed, a TIRIS encapsulated transponder was inserted and flipper band numbers (if present) were noted.

Samples were drained and weighed to the nearest 0.5 g. Large fleshy remains were sieved to allow extraction of sagittal otoliths and cephalopod beaks. All otoliths and cephalopod beaks were removed from the sample and stored in vials. Fish species were identified by comparison to an otolith reference collection (S. D. Goldsworthy and B. Page unpubl. data). Cephalopods were identified from upper and lower beaks following standard techniques (Lu and Ickeringill 2002; Chiaradia et al. 2003). Otoliths were paired to estimate the minimum number of fish consumed and any odd otoliths were assumed to represent additional individual fish. The minimum number of cephalopods was estimated by counting pairs of lower and upper beaks. Percentage of larval fish was estimated when present in a sample. Otoliths in this case were too small to be extracted for identification.

To estimate the length and mass of fish consumed, sagittal otolith length was measured using a binocular microscope, which was fitted with a graticule, to the nearest 0.1 mm, using the Image- Pro Plus TM (Media Cybernetics, Inc) image analysis program. Digestion of stomach contents causes degradation of otoliths (Gales 1988), and hence can reduce the size of the otoliths and may result in underestimates of fish length and mass. Therefore, only otoliths that had not been affected by digestion were measured. The small number of

17 otoliths and cephalopod beaks collected from most prey species and the unavailability of regressions for some prey species meant that analyses of prey size were restricted to anchovies (Engraulis australis), southern sea garfish (Hyporhamphus melanochir) and pilchards (Sardinops sagax).

The diet composition was summarised in 3 ways 1) the frequency of occurrence (FO) (the proportion of samples collected that contained a specific taxon), 2) percentage numerical abundance (NA) (percentage of the total prey items made up by each prey taxa) and 3) wet mass. Estimated fish mass (EFM) and fish length were calculated using otolith length versus fish length regressions, which were calculated from the equations in Gales and Pemberton (1990).

The niche overlap index was used to identify the extent to which the same prey species were consumed by penguins from West and Granite Islands

(Equation 2) Schoener (1968), where p1j and p2j are either the percent biomass or percent numerical abundance of the jth prey taxa for each colony. If the 2 groups consumed the same prey in identical proportions, the index would equal 1. If entirely different prey were taken, the index would equal 0.

Diet of New Zealand fur seals Scats of New Zealand fur seals were collected from the non-breeding haul out sites on the Granite Island break-water and from West Island from March to November 2006. Each scat was placed separately into a plastic bag and was taken back to the laboratory and stored at –20°C until examined. Scats were soaked in hot soapy fresh water for at least one day to allow extraction of prey remains. Individual scats were sieved using water in 1.0 and 0.5 mm sieves. Otoliths, feathers and crustacean carapaces were stored dry and cephalopod beaks were kept in 70% ethanol.

Scats from both islands were pooled to give an indication of seal diet in the region, because the sample sizes were not large enough to describe patterns at each island. Prey remains such as bird feathers, fish otoliths, crustacean

18 carapaces and cephalopod beaks were identified to the lowest taxonomic group possible using a reference collection (S. D. Goldsworthy & B. Page unpubl. data). To determine the importance of prey taxa in the diet, two standardised measures were assessed 1) percentage numerical abundance: NA) and 2) percentage scats with a particular prey item or frequency of occurrence: FO. Estimation of the minimum number of individuals in each sample was carried out by counting the: upper and lower cephalopod beaks, left and right otoliths and eye lenses. Scats that contained penguin feathers were assumed to contain one individual only. Where otoliths or bones were too eroded to enable identification they were assigned to the unidentified prey class. Octopods were identified to genus and squid beaks were identified to Family level.

To provide an estimate of the number of New Zealand fur seals hauled-out in the Granite and West Island region, counts were made on each visit to both Granite (March – October) and West Islands (May – October). Because Seal Rock is approximately 2 km from Granite Island, the number of seals present was counted using binoculars (June – October). Where multiple counts were conducted in a month the average number of seals counted is presented.

To determine whether there were any significant differences in reproductive success, foraging behaviour and diet of little penguins between colonies ANOVA were used if the data were normally distributed. When examining the difference in prey size and mass between colonies a Levene’s test for equality of variances was used, prior to analysis. If the data were not normally distributed Mann-Whitney tests were used, for which Z approximations are presented. Means are presented as ± standard deviation and all t tests are two-tailed, unless stated, the α level of statistical significance was accepted at 0.05. Initial analyses indicated that data collected from male and females, in all cases, did not differ significantly, so these data were pooled for analyses.

To explore correlations between foraging and oceanographic variables a Spearman’s correlation coefficient r was used. Differences in the mean direction of foraging trips among the two groups were tested using the circular

19 statistics package Oriana (V 2.02c, Kovach Computing Service, Pentraeth, Wales, UK). A two- tailed Pearson parametric correlation coefficient was used to examine the relationships between dietary variables.

To examine differences in diet composition and foraging behaviour of little penguins at Granite and West Islands, Analysis of Similarities (ANOSIM), using PRIMER® (Plymouth Marine Laboratory, UK) was used. ANOSIM ranks the dissimilarities of the average of all groups using the Bray-Curtis dissimilarity testing procedure (Quinn and Keogh 2002). To compare the rank between-group and within-group variation a global test statistic (r) and pair- wise R values were calculated. R is scaled from +1 to -1. Groups are considered significantly different when R values are significantly greater than 0 indicating that within group variations are less than those between groups. Negative R values indicate that within group dissimilarities are greater than those between groups. The null hypothesis cannot be rejected when the R value does not differ significantly from 0 indicating no significant difference between groups.

20 5 Results

Reproductive success In all, 62 and 28 breeding attempts by little penguin pairs at Granite and West Islands respectively were included in the final analyses. The onset of breeding was approximately 2 and 3 months later than average at both colonies. The breeding season was also asynchronous at Granite Island with the first eggs laid in June, with a peak hatching period in August. At West Island the onset of the breeding season was later than at Granite Island. At West Island, the peak of breeding activity (approximately 40 breeding burrows) occurred in September, whereas in early August there were only 10 burrows with eggs and two burrows with chicks.

Predation by rats It was possible to confirm that rats killed 4 chicks at Granite Island. An additional 15 dead chicks were found near their burrows with wounds inflicted by rats, but in these cases it was not possible to determine if the chick had been scavenged by the rat, because they had been dead for >1 night.

Inter-annual breeding success at Granite Island Breeding success was variable across years with 0.3 to 1 chicks fledged per pair (Table 1). On average 0.58 ± 0.23 chicks fledged per pair across all years (n=9). The lowest number of chicks fledged per pair (0.30) occurred in 1991 and 2001. Breeding success from 2002 to the present (0.60-0.70 chicks fledged per pair) did not fluctuate as much as in previous years (0.30-1.00 chicks fledged per pair).

Inter-annual breeding success at West Island At West Island breeding success varied from 0.4 to 0.9 chicks fledged per pair with an average of 0.78 ± 0.20 chicks fledged per pair across all years (n=5) (Table 1). In years where comparative data were available in 1991, 2002, 2003 and 2006, breeding success was greater at West Island than at Granite

21 Island. This difference was most apparent in 1991 when only 0.3 compared to 0.8 chicks were fledged per pair at Granite and West Islands respectively. Hatching success (F = 0.39, df = 1, P > 0.05), egg success (F = 1.80, df = 1, P > 0.05), and the number of chicks fledged per pair (F = 1.2, df = 0.1, P > 0.05) did not vary significantly between colonies. However, fledging success did vary significantly (37.0% at Granite Island, 54.3% at West Island; F = 7.43, df = 1, P < 0.05) (Table 1).

Table 1. Measures of reproductive success for Granite and West Islands from 1990 – 2006. Data from 1990 - 2005 were obtained from the Granite Island Little Penguin Monitoring Group (N. Gilbert and R. Brandle unpubl. data). Island 1990 1991 1995 1996 1999 2000 2001 2002 2003 2004 2005 2006 Mean Granite Island Hatching success 76.6 49.0 85.5 - - - 72.2 54.3 85.0 66.0 78.8 83.0 71.7 ± 13.3 Fledging success 67.3 27.6 52.1 - 48.1 39.1 17.9 52.6 41.2 43.5 39.0 37.0 39.8 ± 13.1 Egg success 51.6 13.5 44.6 - - - 13.0 28.6 35.0 28.7 30.8 30.0 28.0 ± 12.6 No. of fledglings per pair 1.0 0.3 0.9 - - - 0.3 0.6 0.7 0.6 0.6 0.6 0.6 ± 0.2 West Island Hatching success - 62.8 - 57.0 - - - 71.8 72.9 - - 76.1 68.1 ± 7.9 Fledging success - 66.7 - 38.8 - 73.0 - 64.3 62.8 - - 54.3 60.0 ± 12.0 Egg success - 41.9 - 22.1 - - - 46.2 45.8 - - 41.3 39.4 ± 9.9 No. of fledglings per pair - 0.8 - 0.4 - - - 0.9 0.9 - - 0.8 0.8 ± 0.2

Chick growth Chicks at Granite and West Islands reached their maximum growth rate at a similar age and the rate of growth was similar for chicks at both sites in 2006 (Tables 2 & 3). The age at which maximum growth occurred t (Z = -1.01, P > 0.05), the maximum amount of growth per day for a chick (Z = -1.81, P > 0.05), the rate of growth since hatching k (F = 1.15, df = 1, P > 0.05) and the asymptote weight (Z = 1.63 P > 0.05) did not differ significantly between colonies.

22 Table 2. Growth parameters of 12 little penguin chicks at Granite Island in 2006. Asymptotic Chick mass r k t Max growth rate g/d

1 812 0.97 9.7 11 36 2 837 0.99 11.8 11 27 3 1091 0.98 35.7 18 180 4 862 1.00 24.4 29 28 5 980 0.99 33.4 36 34 6 884 0.99 9.5 14 34 7 870 0.99 26.4 22 35 8 1230 0.98 21.8 21 90 9 1165 0.99 12.0 13 32 10 819 1.00 46.4 5 21 11 1075 0.99 43.0 20 29 12 1430 0.95 1.2 5 135 Average 1005 0.98 22.9 17 57

Table 3. Growth parameters of 11 little penguin chicks at West Island in 2006.

Asymptotic Chick mass r k t Max growth rate g/d

1 639 0.99 0.3 17 90 2 870 0.97 31.7 15 220 3 992 0.99 1.4 17 47 4 440 0.99 12.5 21 24 5 614 0.99 15.7 13 14 6 771 0.99 16.1 19 23 7 722 0.99 13.9 17 20 8 895 0.99 17.0 9 16 9 764 0.99 14.7 12 29 10 1190 0.99 36.6 28 24 11 981 0.99 27.9 23 16 Average 807 0.99 17.1 17 48

Population census during the breeding season On Granite Island, the number of burrows occupied by adult little penguins has been declining since 2001 (Fig. 2). There were 774 occupied burrows in 2001, which subsequently dropped to 414 and 294 active burrows in 2002 and 2003, respectively. However, in 2004 the number of active burrows increased to 542. The mean rate of decline in the number of active burrows at Granite Island was 15.6 %/year. No data on the number of active burrows are available for West Island, prior to 2006. In 2006, 120 active burrows were counted.

800

y = 679.89e-0.1453x R2 = 0.4743

s 600

400

No. of active burrow 200

0 2001 2002 2003 2004 2005 2006 Year

Figure 2. The number of active little penguin burrows during the breeding season reported at Granite Island from 2001-2006.

Foraging ecology Satellite transmitters were deployed on 6 male and 4 female penguins on Granite Island during July, and 5 male and 3 female penguins were tracked from West Island during September 2006. The average direction of travel, foraging routes and proportion of time spent in area are presented in Figure 3, 4 and 5, respectively. One of the penguins from West Island failed to return to the colony after undertaking a foraging trip. The transmitter failed after four days at sea. Therefore, part of the data on its foraging trip was excluded from some analyses: trip duration, maximum distance from the colony, and total distance travelled. In total, 146 and 280 unfiltered satellite positions were obtained from Granite Island and West Island penguins, respectively. The filter removed 8 (0.5%) and 64 (23%) locations from Granite and West Islands respectively, thus the average number of filtered locations per day were: 14 and 27 for Granite Island and West Island penguins respectively. The mass of adult birds at Granite (mean: 1260 ± 21 g) and West Islands (mean: 1130 ± 13 g) did not differ significantly (F = 1.86, P > 0.05) (Table 4), nor did the mass of chicks (Granite Island mean: 281 ± 126; West Island mean: 321 ± 118; F = 1.019, P > 0.05). However, when all behavioural and oceanographic parameters were compared across islands, ANOSIM detected a significant

24 difference in foraging behaviour between Granite and West Island little penguins (one-way ANOSIM, P1000 = 0.01, Global R = 0.204).

Summary of foraging behaviour Most penguins from Granite Island travelled away from the colony on a median bearing of 135 (SE) degrees and travelled total distances of between 20 and 60 km (mean: 40.9 ± 3.9 km) (Table 5). Penguins from West Island travelled at a median bearing of 207 (SW) degrees and travelled total distances between 35 and 242 km (mean: 76.3 ± 28.1) (Fig. 3). The direction of travel did not differ significantly for little penguins at both colonies (Watson- Williams F-test, F = 0.18, df = 9, P > 0.100) (Fig.4).

4 3 2 1

270 4 3 2 1 1 2 3 4 90

1 2 3 4

180

Figure 3. Circular histogram of the mean direction of travel of breeding little penguins from Granite (blue) and West Islands (red) in 2006.

Foraging parameters The duration of foraging trips did not differ significantly between Granite Island (0.54 ± 0.05 days) and West Island (0.97 ± 1.12 days) (Z = -1.39, P > 0.05). Penguins from West Island reached a greater maximum distance (mean: 18.7 ± 4.8 km) from the colony compared to Granite Island birds (mean: 10.0 ± 1.0 km) (Z = -2.57, P < 0.010) (Table 4). The average speed travelled by West Island penguins (mean: 3.6 ± 0.4 km/h) and Granite Island penguins (mean: 3.4 ± 0.3 km/h) did not differ significantly (Z = -0.08, P > 0.05). Total distance travelled was negatively correlated to bird mass (r = -0.63 P < 0.01). The

25 maximum distance from the colony was positively correlated to total distance travelled (r = 0.56, P < 0.05), average speed (r = 0.54, P < 0.05) and chick mass (r = 0.48, P < 0.05) (Table 4).

Figure 4. Summary of the inferred foraging routes undertaken by little penguins at Granite Island (red) and West Island (blue) based on satellite tracking in 2006.

Figure 5. The proportional time spent foraging in oceanic areas of 1 x 1 km grids by 10 little penguins from Granite Island. Proportional time spent in each grid is indicated by colour: where red represents regions where penguins spent more time followed by orange, yellow, green and finally blue areas where penguins spent relatively little time.

Figure 6. The proportional time spent foraging in oceanic areas of 1 x 1 km grids by 8 little penguins at West Island. Proportional time spent in each grid is indicated by colour: where red represents regions where penguins spent more time followed by orange, yellow, green and finally blue areas where penguins spent relatively little time.

Oceanography There were no significant differences in water depth used by little penguins from Granite Island (median depth: 38.4 m and mean depth: 34.1 ± 2.4 m) and West Island (median 38.1 and mean depth: (39.5 ± 1.2 m) (F = 10.46, P = 0.05), nor in bathymetric gradient between Granite Island (0.34 ± 0.04 per km) and West Island (0.34 ± 0.04 per km) (F = 0.18, P > 0.05) (Table 5).

Table 4. Summary of the foraging parameters of little penguins at Granite and West Islands in 2006.

Penguin Colony Month Penguin Mass at Foraging trip Foraging trip Total distance Median bearing Maximum distance Mean horizontal no. sex retrieval (g) commenced duration (days) travelled (km) (degrees) from colony (km) speed (m/s) 665 Granite July F 1060 26/07/2006 0.5 37.0 164 6.3 3.1 707 Granite July F 1100 11/07/2006 0.5 49.4 171 11.8 4.0 787 Granite July F 1500 24/07/2006 0.5 20.4 48 5.0 1.7 844 Granite July F 1000 12/07/2006 0.5 50.8 142 10.2 4.6 671 Granite July M 1000 25/07/2006 0.7 49.9 181 13.2 3.0 688 Granite July M 1180 14/07/2006 0.5 60.8 157 11.9 5.0 705 Granite July M 1235 11/07/2006 0.5 37.3 142 8.0 4.0 726 Granite July M 1200 15/07/2006 0.5 25.1 76 6.3 2.1 821 Granite July M 1450 09/07/2006 0.5 36.5 133 13.1 3.5 972 Granite July M 1300 23/07/2006 0.6 42.6 138 14.4 3.0 7a West Sep F - 12/09/2006 3.0 190.8 115 63.0 3.5 624 West Sep F 1260 17/09/2006 0.6 50.8 223 18.1 3.6 681 West Sep F 1000 19/09/2006 0.6 43.7 217 10.4 3.2 23 West Sep M - 12/09/2006 0.6 46.2 202 13.8 3.3 124 West Sep M 1180 13/09/2006 0.5 78.7 174 15.5 6.2 596 West Sep M 950 15/09/2006 3.5 242.3 206 47.5 2.9 647 West Sep M 1240 12/09/2006 0.5 35.2 223 14.7 3.1 674 West Sep M 1150 18/09/2006 0.5 37.3 207 11.3 2.9

Females Granite (mean ± sd, median) 0.5 ± 0.01, 0.5 39.4 ± 14.1, 43.2 153 8.3 ± 3.2, 8.3 3.4 ± 1.3, 3.6 Males Granite (mean ± sd, median) 0.5 ± 0.1, 0.5 42.0 ± 12.3, 39.1 140 11.1 ± 3.2, 12.5 3.4 ± 0.1, 3.3 Females West (mean ± sd, median) 1.4 ± 1.4, 0.6 95.1 ± 82.9, 50.8 216 30.5 ± 28.4, 18.1 3.5 ± 0.2, 3.5 Males West (mean ± sd, median) 1.1 ± 1.3, 0.5 87.9 ± 88.0, 46.2 205 20.6 ± 15.1, 14.7 3.7 ± 1.4, 3.1 Juvenile males (mean ± sd, median) a Incomplete foraging trip because the transmitter failed at sea.

28 Table 5. Summary data of the oceanographic characteristics of the areas that the little penguins foraged in from Granite and West Islands in 2006.

Mean Median Mean gradient Penguin Penguin depth depth (change in depth Colony No. Month sex (m) (m) (m) per horiz. km) Granite 665 July F 20 27 0.4 Granite 844 July F 21 23 0.4 Granite 707 July F 30 36 0.2 Granite 787 July F 6 5 0.6 Granite 726 July M 13 15 0.4 Granite 705 July M 21 22 0.3 Granite 821 July M 27 30 0.2 Granite 688 July M 29 35 0.6 Granite 671 July M 29 35 0.3 Granite 972 July M 30 36 0.1 West 624 Sept F 25 27 0.4 West 681 Sept F 28 35 0.7 West 7 Sept F 36 39 0.1 West 647 Sept M 24 27 0.9 West 674 Sept M 31 35 0.3 West 23 Sept M 32 35 0.2 West 596 Sept M 32 33 0.2 West 124 Sept M 33 36 0.2

Granite females (mean) 19 27 0.3 Granite males (mean) 25 32 0.8 West females (mean) 28 31 0.3 West males (mean) 31 35 0.4

Overlap of foraging areas between Granite and West Island penguins Niche overlap indices calculated that the foraging areas used by the birds from Granite versus West Islands overlapped by 9.0%. When the incomplete foraging trip of bird 7 from West Island was included the proportion of spatial overlap in foraging areas was slightly less, 7.0%. This suggests that there was little spatial overlap in the foraging space of little penguins from each island in 2006.

Penguin Diet In all, 57 birds were captured from July to October to obtain diet samples from Granite Island (n = 43) and West Island (n = 14). Of these birds, only 25 (58 %) from Granite and 11 (78 %) from West Island had prey remains in their stomachs. The average wet mass of samples was 45.0 ± 33.1 g (range: 8.5 – 113.7 g) at Granite Island and 46.0 ± 35.0 g (range: 10.0 g – 125.2 g) at West

29 Island. There was no significant difference in the mean sample mass between the two islands (F = 0.12, df = 1, P > 0.05). Mean mass of stomach contents was not correlated with body mass (r = -0.24, P> 0.05).

A total of 421 otoliths were found and measured, and classified to be in either pristine or poor condition. This resulted in the exclusion of 332 otoliths from the length and weight analyses because they were judged to be too eroded. The prey species in this study comprised nine species of fish and two species of cephalopods. In all, there were 191 individual prey items identified. The most frequently detected prey for penguins at both colonies were anchovies, with a FO value of 31.3% and 32.0% at Granite and West Island colonies, respectively. Anchovies occurred in the diet in all months. The next most frequent prey species were southern sea garfish 20.3% and 12.0%; pilchards 14.1% and 8.0%; Gould’s squid (Nototodarus gouldi) 9.4% and 12.0%; and unidentified fish larvae 7.8% and 8.0% at Granite and West colonies respectively. Other species of fish occurred but they were infrequently recovered and typically found in low quantities (Table 6).

Table 6. The prey consumed and their relative numerical abundance (NA) and frequency of occurrence (FO) in stomach contents of little penguins from the Granite and West Island colonies in 2006. The number of samples examined, the number with prey remains and the minimum number of individuals consumed were 37, 37 and 191 respectively.

Prey type NA FO Granite Is. West Is. Granite Is. West Is. Anchovy (Engraulis australis ) 50.0 42.5 31.3 32.0 South sea garfish (Hyporhamphus melanochir ) 11.6 4.6 20.3 12.0 Pilchard (Sardinops sagax ) 4.6 6.5 14.1 8.0 Unidentified larvae 17.9 17.2 7.8 8.0 Gould's squid (Nototodarus gouldi ) lower beaks 6.5 8.8 9.4 12.0 Tommy rough (Arripis georgianus ) 2.07.41.68.0 Unidentified fish 4.2 2.8 6.3 4.0 Sandy sprat ( Hyperlophus vittatus ) 0.05.60.04.0 Spotted pipe fish (Stigmatopora argus ) 2.20.06.30.0 Redbait (Emmelichthys nitidis ) 0.01.90.04.0 Snapper (Pagrus auratus) 0.01.90.04.0 Blue mackerel (Scomber australasicus ) 0.00.90.04.0 Red cod (Pseudopycis bachus ) 0.70.01.60.0 Calamari squid (Sepioteuthis australis ) lower beaks 0.20.01.60.0 Total 100 100 100 100

Size and weight of prey Table 9 shows the mean, minimum and maximum lengths and weights of anchovies, southern sea garfish and pilchards consumed by little penguins in this study. The estimated length of each prey species did not differ significantly between colonies: anchovy (t = 0.66, df = 18, P > 0.05); garfish (t = 0.38, df = 7, P > 0.05), and pilchard (t = -0.958, df = 9, P > 0.05). The heaviest fish was a garfish weighing 24.3 g with most fish weighing under 10 g. Only garfish exceeded 12 cm in length (89 % of fish) with the longest fish being 21 cm. There were no significant differences in the mass of the different prey between colonies, anchovy (t = 1.15, df = 18, P > 0.05), garfish (t = 0.93, df = 7, P > 0.05) and pilchard (t = -1.033, df = 9, P > 0.05).

Table 7. The mean, minimum and maximum lengths (mm) and weights (g) of anchovies, southern sea garfish and pilchards consumed at Granite and West Islands. Length Mass Max Min Max Min Species Colony (mm) SD (g) SD length length mass mass Anchovy Granite 84.1 12 6 2.4 124 55.7 20.4 1.3 Anchovy West 79.8 6.9 4.7 1.4 89.9 97.8 6.9 2.8 Southern sea garfish Granite 131.7 34.4 5.2 4.4 219.8 66.9 24.6 0.3 Southern sea garfish West 116.5 3.2 2.2 0.2 118.8 112.5 2.4 1.9 Pilchard Granite 64.4 15.7 3.2 2.6 94.4 41.9 8.9 0.7 Pilchard West 75.9 22.7 5.5 4.8 100.5 52.8 10.9 1.5

Comparison of diet between colonies The species and number of prey consumed by little penguins were not

significantly different within and between colonies (one-way ANOSIM P1000 = 0.447, Global R = -0.03). Based on the niche overlap index, there was high overlap in both the numerical abundance and frequency of occurrence of species consumed between Granite and West Islands (80% and 70% respectively). The differences in diets between colonies were accounted for by differential consumption of anchovies (numerical abundance: 14 %, frequency of occurrence: 1.4%) southern sea garfish (13%, 16%), tommy rough (10%, 12%), sprat (10%, 7.8%), pilchards (3.6%, 11%) and pipe-fish (12%, 4.3%) (Table 8).

31 New Zealand fur seal diet In all, 34 samples were collected, of which 29 (83%) contained prey remains. Fish were found in 20 (68%) of the samples containing remains, penguin feathers were found in 12 (41%) samples and octopod beaks were found in 3 samples (10 %) and 1 sample had a squid beak (>1 %). In total, 5 species of fish, 1 squid family, 1 Octopus genus, and penguins were identified. The minimum number of individual prey identified was 98. In all, 80 individual fish, 12 penguins and 6 squid were consumed. Yellow-eyed mullet was the most prevalent prey item (FO 35.8%), followed by; little penguin (FO 30%), garfish (FO 7.5%), Octopus sp. (10%), cardinal fish (FO 5.0%), barracouta (FO 2.5%), unidentified fish (FO 2.5) and Ommastrephidae (FO 2.5%).

Table 8 The percent numerical abundance (NA) and frequency of occurrence (FO) of prey types found in New Zealand fur seal scats at Granite and West Islands. The number of scats examined, the number with prey remains and the minimum number of individuals consumed are 29, 29 and 98, respectively. Prey type NA FO Yellow Eyed Mullet (Aldrichetta forsteri ) 35.8 35.0 Little penguin (Eudyptula minor ) 33.7 30.0 Garfish (Hyporhamphus melanochir )6.97.5 Octopus sp. 6.2 10.0 Red Bait (Emmelichthys nitidus )5.85.0 Cardinal (Vincentia conspersa )4.95.0 Baracouta (Thyrsites atun )3.42.5 Unidentified Fish 2 3 Ommastrephidae 1.1 2.5 Total 100 100

32 Number of New Zealand fur seals hauled-out New Zealand fur seals were present throughout the region for the duration of this study. Seal Rocks consistently had the greatest number of fur seals, with a mean maximum of 32 ± 2.2 seals in June and a minimum of 13 ± 1.1 seals in October (mean 22.4 ± 7.9 per month). At West Island the mean maximum number of seals 12 ± 3.5 peaked in September and was lowest in May 5 ± 1.4 (mean 8 ± 2.5 per month). Overall, there were fewer fur seals on the Granite Is than in the other locations, a mean maximum of 7 ± 1.4 in July and a minimum of 1 ± 1.2 in Aug and Sept (mean 2.6 ± 1.8 per month) (Fig. 8).

35 Seal Rock 30 West Island

25 Granite Island breakwater 20

0 r h y ly st e rc u u a Ma June J M Aug ptember Octob e S

Figure 7. The average number of New Zealand fur seals hauled-out at Seal Rocks, West Island and Granite Island between March-October 2006.

33 6 Discussion

This study investigated factors that may be responsible for the population decline in little penguins at Granite Island. The breeding success of little penguins at Granite and West Islands were compared, as were aspects of the diet and foraging behaviour. Results comparing the reproductive success and foraging ecology of each population, the current size of each population and the importance of little penguins in the diets of New Zealand fur seals are discussed below.

Reproductive parameters Although the number of active burrows fluctuated at Granite Island, overall the number of active burrows has decreased since 2001. The number of active burrows found in 2006 at West Island was much less than expected given the large number of birds banded there in the early 1990s. However, the number of active burrows at West Island was underestimated in 2006 because some areas were inaccessible (approximately 50 m by 50 m), although there did not appear to be many birds entering this area. Another census is required to confirm that the West Island population has declined, although findings from this study indicate that the population has declined significantly since the early 1990s.

Seabirds typically breed when their prey peak in abundance in near-shore waters, which minimises the time adults spend commuting between foraging grounds and the colony (Grémillet et al. 2004). Therefore, when prey abundance is poor near the colony the onset of breeding can be delayed (Dann et al. 2000). The breeding season in 2006 was delayed by 2 and 3 months at Granite and West Islands respectively. Interestingly, commencement of breeding at nearby Troubridge Island (Fig. 1) was also delayed in 2006 (A. Wiebkin unpubl. data). The Bonney Upwelling, which nourishes this region, was reported to be poor in 2006 because of a lack of winds from the SE (P. Gill and A. Levings pers. comm.). Although the

34 distribution of little penguin prey was not measured in this study, it is possible that the peak abundance of their prey was delayed or relatively small, because the poor upwelling resulted in depressed primary production (Ward and Staunton-Smith 2002).

Such factors may have influenced the delayed onset of breeding of little penguins breeding at these colonies this year. The late finish of the moult this year may have also influenced the late onset of breeding. At Granite Island, the moulting period was longer than usual in 2006 (N. Gilbert pers. comm.), which has been associated with a late onset of breeding at other little penguin colonies (Reilly and Cullen 1983). When egg laying is delayed, the ability of a little penguin to raise a second clutch is reduced because they abandon their dependent chicks in preparation for the annual moult, which potentially halves the annual breeding success (Perriman et al. 2000; Knight and Rogers 2004; Robinson et al. 2005).

Both colonies showed marked inter-annual variation in breeding success, however little penguins at West Island produced relatively more fledglings per pair. Inter-annual variation in breeding success is common among penguins and in other seabirds and has been linked to food availability and predation (Bertram 1995; Rindorf et al. 2000; Wienecke et al. 2000). Overall, the breeding success at Granite and West Islands was relatively poor, compared to Philip Island (Victoria) and Lion Island (New South Wales), where the respective average numbers of chicks fledged was 1.0 ± 0.4 and 0.88 ± 0.18 per breeding pair per year (Rogers et al. 1995; Dann et al. 2000). Based on rankings system to assign to breeding success at Phillip Island (Robinson et al. 2005), little penguins at Granite and West Islands had poor breeding success in 66% and 25% of the years presented in this study, respectively.

In 2006, the success of earlier breeders was lower than those that bred later in the season, in contrast to the findings of other studies (Chiaradia and Kerry 1999). This finding is presumably because earlier breeders typically have shorter incubation and foraging trips, that reduces the time between meals for

35 chicks and leads to greater fledging success (Reilly and Cullen 1981). Although, this observation may not be as pronounced when food availability is poor early in the breeding season. The findings presented here based on the timing of breeding, indicates that prey may have been in relatively low abundance early in the breeding season.

Impacts of tourism and predation by rats on little penguins Interestingly, the breeding success of penguins at Granite Island is comparable to that at Penguin Island, where penguins are also a tourist attraction (Klomp et al. 1991). The colonial nesting of seabirds in large aggregations in often scenic coastal locations attracts many tourists (Yorio et al. 2001). To accommodate the needs of the expanding tourism industry on Granite Island, a restaurant, gift shop and penguin information centre were built in 1995 and vehicles regularly use the area where little penguins come ashore.

At Granite Island little penguins nest in burrows and they may therefore be less sensitive to disturbance than those that nest on the surface. This implies that disturbance when little penguins are returning to their burrows and when they return to the sea is most critical. Penguins are highly faithful to their landing sites, where disturbance delays breeding birds from coming ashore and may deter young birds from breeding in these areas (Weerheim et al. 2003). Klomp et al. (1991) demonstrated that little penguins breeding in areas with many tourists had lower breeding success (26%) compared to those breeding in areas of low tourism impact (40%). At Granite Island, considerable effort has been made to reduce disturbance to the little penguins nesting in the areas frequented by tourists. Nest boxes and board-walks have been constructed and access to the island is restricted at dusk to people taking part in an organised penguin tour. Despite these efforts, it is possible that tourists have negative impacts on the Granite Island population.

Numerous species of seabirds are negatively affected by predation by introduced rat species (Martin et al. 2000; BirdLife International 2004).

36 Predation by rats was shown to reduce breeding success by 24% at a little penguin colony in New Zealand (Perriman et al. 2000). At Granite Island, rats preyed on relatively small chicks, post incubation, rather than on eggs, which has been the case in other seabird studies (Igual et al. 2006). Chicks being incubated by their parents are less vulnerable to predation by rats, because adult penguins could keep rats at bay. In many cases, it was not possible in the current study to distinguish whether chicks had died as a result of wounds inflicted by rats or if rats had scavenged dead chicks, as has been noted in other studies (Cunningham and Moors 1994; Stapp 2002). In addition, when chicks disappeared from their burrows it was not possible to determine the cause of death. Daily checks of burrows would be required to determine how many chicks were killed by rats, which was not feasible in the current study. In addition, extensive rat baiting was carried out during the penguin breeding season in 2006 on Granite Island, which may have reduced the number of chicks killed by rats. In previous years, the number of chicks killed has been observed to be greater (N. Gilbert pers. comm.). Despite the absence of tourists and terrestrial predators on West Island, there appears to have been a substantial decline in the number of birds nesting on this island as well. This indicates that, in addition to onshore predation and disturbance caused by tourists at Granite Island that other perturbations may be placing pressure on both populations.

Predation by New Zealand fur seals This study confirmed that little penguins formed a significant part of the diet of New Zealand fur seals around Granite and West Islands. Little penguins were the second most prevalent prey item suggesting predation pressure could be important. Based on these findings, if we conservatively set the number of resident fur seals in the region at 20, and if they consume a single penguin per week, more than 1000 penguins would be consumed in a year. Given that the number of little penguins at Granite and West Island probably only number in the hundreds, this estimate of fur seal predation would not be sustainable. A more conservative estimate of fur seal predation of 20 seals eating one penguin per month would equate to 240 adult penguins being taken a year.

37 The average number of resident seals is likely to be greater than 20 (if surveys in this study are an indication), and based on the dietary data predation rates on penguins could be even higher. However, the importance of little penguins in the diet of New Zealand fur seals from nearby Seal Rocks, the largest haul-out in the region, is unknown. Predation of adult birds during the breeding season has dual impacts because not only is an adult lost, but its clutch is also likely to be abandoned.

Fur seals prove a threat to birds by direct consumption and by competition for breeding spaces. For example, the consumption of little penguins by New Zealand fur seals around Kangaroo Island is thought to have caused the local extinction of a little penguin breeding colony at Cape Gantheaume (Page et al. 2005). Whereas, competition for space with Cape fur seals in recent years has resulted in several seabird colonies being displaced on islands in Namibia (Crawford et al. 1989). Populations of New Zealand fur seal on Kangaroo Island are currently increasing by 12.5% per annum, and have increased almost 10-fold in size over an 18-year period (based on data presented in Shaughnessy 2006). As populations of New Zealand fur seals continue to recover, predation rates are likely to increase, escalating the risk of more localised extinctions, including populations at Granite and West Islands.

Foraging parameters The results obtained in this study provide insights into the foraging behaviour of little penguins during the chick raising period. Unexpectedly, there was very little spatial overlap in the foraging areas used by penguins from the two colonies. Fish stocks are known for their heterogenous distribution and therefore, the utilisation of different foraging areas by penguins from each island may reflect movement of fish during the breeding season (Weimerskirch et al. 2005) and the differences in timing of satellite tracker deployments on birds from the two islands.

Although, penguins may reduce competition for prey with con-specifics from nearby islands by using foraging areas that do not overlap. Aggregations of breeding seabirds may deplete local food resources (Ashmole 1963; Birt et al.

38 1987), which increases the separation of breeding and feeding habitats and amplifies the cost of commuting to provision dependent young, which remain at the central place (Orians and Pearson 1979). Ideally, seabird colonies should be sufficiently spaced to reduce intra-specific competition for prey (Cairns 1989), however, the availability of islands suitable for breeding often occur in groups (Forbes et al. 2000). There are many examples of how intra- colony competition can result in different foraging strategies. For example, magellanic penguins (Spheniscus magellanicus) foraged close to their colonies and consumed fish that were of relatively poor quality rather than feeding in distant waters, which were used by penguins from other colonies (Forero et al. 2002). In contrast, cape gannets (Morus capensis) at nearby colonies in South Africa exhibited complete separation in foraging areas, despite the islands being within a cape gannets foraging range (Grémillet et al. 2004). West and Granite Islands are located well within the foraging ranges of little penguins breeding on either island. Considering the small size of the populations it is unlikely that there is density dependent localised prey depletion and resultant foraging separation around the islands.

With the exception of two birds from West Island (one of which did not return and is not considered further), all little penguins undertook <16 hr foraging trips, which were within 20 km of the colony. Little penguins forage by sight and are therefore only able to forage during daylight, and their arrival and departure times from their colonies reflect this restriction (Ropert-Coudert et al. 2006). The average maximum distance travelled in a day was greater at Granite (41 km) and West Islands (48 km) than little penguins at Phillip Island (33 km) (Weavers 1992), indicating that little penguins in this study travelled relatively further in search of prey. Little penguins have been shown to make longer foraging trips and travel further when prey availability is relatively low (Hobday 1992; Weavers 1992). The greater travel speeds and travel distances of little penguins in this study suggest that the prey availability around Granite and West Islands was reduced during in 2006.

39 The average travel speed in this study at Granite Island (3.4 ± 0.3 km per hour) and West Island (3.6 ± 0.4 km per hour) was also greater than at Phillip Island (1. 5 ± 1.1 km per hour (Weavers 1992). Marine animals exploiting patchy environments would be expected to move slowly in areas where resources are plentiful and quickly move through areas where resources are scarce, because searching for plentiful patches is typically more beneficial than remaining in sparse ones (Pinaud and Weimerskirch 2005). These findings support the conclusions of relatively low prey availability around Granite and West Islands during 2006. Although, considering that little penguins from West Island on average travelled further in search of prey, yet they had greater reproductive success suggests that in the current season increased effort in searching for food did not lower reproductive success at West Island compared to Granite Island.

Similarity in prey diversity and meal sizes The relatively late breeding season, low breeding success, long duration and distance of foraging trips, which were documented in this study, indicate relatively poor prey availability. Although prey availability was not measured, the diet of penguins on both islands was quantified, and provides information on the size of individual prey available and the size of meals fed to chicks. However, seabird diet studies must be interpreted with caution, because they are subject to biases as a result of differential rates of digestion, which depend on: 1) stomach fullness, 2) the time the food has spent in the stomach and 3) the size of the prey (Gales 1988; Gales and Pemberton 1990). Consequently, otoliths can be eroded to varying degrees, which can affect estimates of the size of individual prey consumed, and some otoliths may be completely digested. Although otoliths that were completely digested cannot be accounted for, the impact of these biases was likely to be minimal as only otoliths that were in pristine condition were used to estimate the size of prey consumed.

As expected, there was significant overlap in the diet of penguins at both islands, despite the relatively disjunct foraging habitats used by these

40 populations. The diet of little penguins at Granite and West Islands reflects the generalist foraging behaviour of little penguins at other locations (Montague and Cullen 1988). Fish were the most common prey in the diet, with only two squid species being recovered. Fish are also usually the most frequently consumed prey of little penguins in Tasmania and Victoria (Gales and Pemberton 1990). Although many species of fish and squid were eaten, anchovies (80-100 mm in length) occurred most frequently, which is similar to the findings of other dietary studies on little penguins in South Australia (A. Wiebkin unpubl. data).

Relatively little is known about small pelagic fishes and their distributions and abundance in South Australia (Ward et al. 2001; Ward and Staunton-Smith 2002; Ward et al. 2006). Anchovies and pilchards dominate the clupeoid assemblages in temperate Australia and both species form large schools, occupy similar habitats and undergo comparable fluctuations in abundance (Ward et al. 2001). Little penguins typically prey on juvenile anchovies and pilchards, which are less than 1 year of age (Hobday 1992). The availability of suitably sized prey exploited by little penguins is likely to be linked to the strength of coastal upwellings, which are thought to strongly influence the breeding success of small pelagic fish (Ward et al. 2006). The peak in anchovy and pilchard spawning generally occurs from January to April in South Australia, which is associated with the timing of coastal upwelling that brings nutrient rich waters, giving rise to relatively high primary, secondary and tertiary production (Ward et al. 2006). In years of strong coastal upwelling, penguins at Phillip Island attain better body conditions and breed earlier, possibly because of increased prey availability (Mickelson et al. 1992). Therefore, it is possible that the timing of breeding of little penguins at Granite and West Islands is also influenced by the strength of regional coastal upwellings.

Chick growth rates Breeding success and chick growth are typically correlated to the meal sizes fed to chicks (Barrett et al. 1987; Olsson 1997; Dahdul and Horn 2003; Ramos et al. 2006). Older chicks are preferentially fed when food availability

41 is poor to improve the chances of fledging at least one healthy chick, this validates that chick survival can be reduced by limited food resources (Radl and Culik 1999). Meal sizes fed to chicks at Granite and West Islands (45 g and 46 g) were smaller than those fed to chicks at nearby Troubridge Island (Fig 1), (315 g, A. Wiebkin unpubl. data), but comparable to those in Tasmania (31-47 g) (Gales and Pemberton 1990) and Victoria (80 g) (Montague and Cullen 1988). The amount of food that a chick is fed determines its fledging mass and ultimately influences its survival. At Troubridge Island the average fledging mass was 1300 g (A. Wiebkin unpubl. data), whereas at Granite and West Islands the asymptotic mass was 955 g and 800 g respectively.

Breeding success only indicates the number of chicks that fledge, whereas mass at fledging can be a better predictor of survival to adulthood because chicks, which fledge with higher body mass, have greater survival rates (Giese et al. 2000). Therefore, the lower body mass at fledging of chicks in the current study may result in poor survival post fledging. The lower fledging success at Granite Island was not related to either meal sizes nor chick growth rates, indicating that differences in breeding success between the two islands were not due to a difference in the capacity of adult little penguins to deliver food to their chicks. In fact this study highlights the similarities in the prey species eaten by penguins at both sites.

The future of the little penguin populations at Granite and West Islands Based on estimates of the size of the breeding population of little penguins at Granite and West Islands conducted during this study, both sites appear to be in a state of decline. This result, plus the fact that both sites displayed a similar late commencement of breeding during the 2006 season, and because little penguins targeted similar prey species, suggests that both populations are responding to factors in their marine environment in similar ways. The decline of both little penguin colonies suggests that either predation pressures or changes in food availability may be causing reductions in adult little penguin survival at both colonies. However, this study has identified strong

42 evidence in support for a ‘predation hypothesis’, where recent declines in little penguin populations are attributed to increasing predation pressure from recovering New Zealand fur seal populations.

In addition, the potential impacts of land-based pressures from either rat predation or disturbance by tourists at Granite Island cannot be discounted. Although no differences were detected in meal masses, chick growth rates, foraging trip duration and distance to foraging sites between Granite and West Island, the breeding success was significantly lower at Granite Island, possibly because of lower chick survival. Evidence of young chicks being removed from burrows and killed by rats, suggested that land-based predation of chicks may be a factor responsible for lower reproductive success at Granite Island. However, the potential role of disturbance from tourism in reducing breeding success could not be discounted in this study.

To understand how the combined factors of predation, prey availability and tourism may affect population declines at Granite and West Islands, it would also be necessary to investigate how variations in survival and fecundity affect their population demography (Gardali et al. 2000; Jenouvrier et al. 2005). Seabird populations are considered to be most vulnerable to a reduction in adult survival because they have low fecundity and because an increase in adult mortality has an immediate impact on population size (Lewison et al. 2004). Because rats are only capable of preying on small unattended little penguin chicks, they are not likely to be impacting on the Granite Island population as negatively as predation by fur seals. Therefore, predation by New Zealand fur seals is likely to have a major impact on the population sizes of both colonies.

However, low fecundity and reduced reproductive success and/or poor juvenile survival over the long-term can lead to population decline. For example, a population of emperor penguins (Aptenodytes forsteri) declined by 50% during the 1960s because of climate change (Jenouvrier et al. 2005). The population did not recover, because reproductive success for the next 30 years was poor. Given the continued poor breeding success at Granite Island,

43 it is possible that recruitment will limit population recovery (Thompson and Ollason 2001). Therefore, it is important to bolster the reproductive success and subsequent recruitment of juveniles into the breeding population at Granite Island to ensure the perpetuity of this colony.

44 7 Conclusions

This study identified that the decline in little penguin numbers were not restricted to Granite Island, and they appear to also be decreasing at neighbouring West Island. This suggests that declines could be due primarily to changes in the marine environment that affect penguin survival. The very significant levels of predation by local New Zealand fur seals, as supported by the importance of little penguins in their diet, could account for much of the observed decline, especially as their populations (and presumable predation pressure on little penguins) have increased rapidly over the last 20 years. Reductions in prey availability may also be affecting the status of little penguin populations, but there was not evidence to support this in the current study.

Breeding success of little penguins at Granite Island was significantly lower than at West Island, potentially because of lower chick survival. This may be the result of predation of young chicks by rats. Further research is needed to quantify the impact rat predation has on overall reproductive success at Granite Island. Although the role of disturbance by tourism could not be discounted, results suggest that land-based factors are also contributing to population declines in the Granite Island population. As such, the rates of decline in penguin populations are likely to be greater at Granite Island compared to West Island. Effort should be directed to continuing population surveys at both sites, and at reducing the impact of land based predators and tourism at Granite Island.

The results from this study indicate that no single factors are responsible for the little penguin population decline at Granite Island. The relatively short duration of this project meant that it was not possible to quantify how each factor (adult predation, nestling predation, low nesting effort or nestling mortality due to a lack of food) or the combination of these factors have contributed to the rate of population decline and this remains to be investigated. These findings indicate that because the population is still

45 declining, the number of tourists allowed to go on nightly penguin tours should not increase to further limit the impacts of tourism. Tourist impacts present a difficult challenge for wildlife managers, but the eradication of rats may increase recruitment. New Zealand fur seals and water rats are protected species and therefore the management of their interactions with little penguins also presents a challenging management issue. The inter-play between tourism, predation, fluctuating oceanographic productivity and the relationship between prey availability and foraging success and reproductive success are complex challenges that face little penguins at these sites. Small increases in the impacts of any one of these factors on little penguins may ultimately be sufficient to progress these little penguin subpopulations to local extinction.

46 8 Recommendations for future research

Carry out additional population surveys at West Island to determine whether the little penguin population there is in decline.

Continue monitoring the survival of chicks on both islands to determine post fledging survival.

Continue baiting of black rats on Granite Island and monitor penguin burrows frequently to provide a better estimate of the number of little penguin chicks killed by both black and water rats.

Continue diet studies on both little penguins and New Zealand fur seals.

Carry out studies on the interactions between little penguins and tourists as the little penguins return to the shore in the evening. This will help to differentiate between the impact of rats and tourists on breeding success at Granite Island.

47 9 References

Adams, J., Takekama, J. Y. and Carter, H. R. (2004). Stable foraging areas and variable chick diet in Cassin's auklets (Ptychoramphus aleuticus) off southern California. Canadian Journal of Zoology 82: 1578-1595. Ainley, D. G., Ballard, G., Karl, B. J. and Dugger, K. M. (2005). Leopard seal predation rates at penguin colonies of different size. Antarctic Science 17: 335-340. Ainley, D. G., Wilson, P. R., Barton, K. J., Ballard, G., Nur, N. and Karl, B. (1998). Diet and foraging effort with Adélie penguins in relation to pack- ice conditions in the southern Ross Sea. Polar Biology 20: 311-319. Ashmole, N. P. (1963). The regulation of numbers of tropical oceanic birds. Ibis 103: 458-473. Barbraud, C. and Weimerskirch, H. (2001). Emperor penguins and climate change. Nature 411: 183-186. Barrett, R. T., Anker-Nilssen, T., Rikardsen, F., Valde, K., Røv, N. and Vader, W. (1987). The food, growth and fledging success of Norwegian puffin chicks Fratercula arctica in 1980-1983. Ornis Scandinavica 18: 73-83. Bertram, D. F. (1995). The roles of introduced rats and commercial fishing in the decline of ancient murrelets on Langara Island, British Columbia. Conservation Biology 9: 865-872. BirdLife International (2004). State of the world's birds 2004: indicators for our changing world. Cambridge, UK, Birdlife International. Birt, T. P., Birt, V. L., Goulet, D., Cairns, D. K. and Montevecchi, W. A. (1987). Ashmole's halo: direct evidence for prey depletion by a seabird. Marine Ecology Progress Series 40: 205-208. Boersma, P. D. (1998). Population trends of the Galápagos penguin: impacts of El Nino and La Nina. The Condor 100: 245-253. Bolduc, F. and Guillemette, M. (2003). Human disturbance and nesting success of common eiders: interactions between visitors and gulls. Biological Conservation 110: 77-83. Botsford, L. W., Castilla, J. C. and Peterson, C. H. (1997). The management of fisheries and marine ecosystems. Science 277: 509-515. Bryant, D. M., Jones, I. L. and Hipfner, J. M. (1999). Responses to changes in prey availability by common murres and thick-billed murres at the Gannet Islands, Labrador. Canadian Journal of Zoology 77: 1278- 1287. Burger, J. and Gochfield, M. (1994). Predation and effects of humans on island-nesting seabirds. Birdlife Conservation Series 1: 39-67. Cairns, D. K. (1989). The regulation of seabird colony size : A hinterland model. American Naturalist 134: 141-146. Carpenter, R., Gilbert, N. and Peters, P. (2004). Monitoring of the effect of human activities on the breeding success of little penguins on Granite Island, Granite Island Nature Park, Victor Harbor: 1-6. Catard, A., Weimerskirch, H. and Cherel, Y. (2000). Exploitation of distant Antarctic waters and close shelf-break waters by white-chinned petrels rearing chicks. Marine Ecology Progress Series 194: 249-261.

48 Chastel, O., Weimerskirch, H. and Jouventin, P. (1993). High annual variability in reproductive success and survival of an Antarctic seabird, the snow petrel Pagodroma nivea. Oecologia 94: 278-285. Chiaradia, A., Costalunga, A. and Kerry, K. (2003). The diet of little penguins (Eudyptula minor) at Phillip Island, Victoria, in the absence of a major prey - pilchard (Sardinops sagax). Emu 103: 43-48. Chiaradia, A. and Kerry, K. (1999). Daily nest attendance and breeding performance in the little penguin Eudyptula minor at Phillip Island, Australia. Marine Ornithology 27: 13-20. Crawford, R. J. M., David, J. H. M., Williams, A. J. and Dyer, B. M. (1989). Competition for space: recolonising seals displace endangered, endemic seabirds off Namibia. Biological Conservation 48: 59-72. Croxall, J. P. and Prince, P. A. (1987). Seabirds as predators on South Georgia marine resources. Seabirds feeding ecology and role in marine ecosystems. J. P. Croxall. Cambridge, Cambridge University Press: 347-369. Croxall, J. P. and Wood, A. G. (2002). The importance of the Patagonian Shelf for top predator species breeding at South Georgia. Aquatic Conservation: Marine and Freshwater Ecosystems 12: 101-118. Cullen, J. M., Montague, T. L. and Hull, C. L. (1992). Food of little penguins Eudyptula minor in Victoria: comparison of three localities between 1985 and 1988. Emu 91: 318-341. Cunningham, D. M. and Moors, P. J. (1994). The decline of rockhopper penguins Eudyptes chrysocome at Campbell Island, Southern Ocean and the influence of rising sea temperatures. Emu 94: 27-36. Dahdul, W. M. and Horn, M. H. (2003). Energy allocation and postnatal growth in captive elegant tern (Sterna elegans) chicks: responses to high- versus low-energy diets. The Auk 120: 1069-1081. Dann, P. (1992). Distribution, population trends and factors influencing the population size of little penguins Eudyptula minor on Phillip Island, Victoria. Emu 91: 263-272. Dann, P. and Norman, F. I. (2006). Population regulation in little penguins (Eudyptula minor): the role of intraspecific competition for nesting sites and food during breeding. Emu 106: 289-296. Dann, P., Norman, F. I., Cullen, J. M., Neira, F. J. and Chiaradia, A. (2000). Mortality and breeding failure of little penguins Eudyptula minor in Victoria, 1995-6 following widespread mortality of pilchard Sardinops sagax. Marine and Freshwater Research 51: 355-362.

David, J. H. M., Curry, P., Crawford, R. J. M., Randall, R. M., Underhill, L. G. and Meyer, M. A. (2003). Assessing conservation priorities in the Benguela ecosystem, South Africa: analysing predation by predation by seals on threatened seabirds. Biological Conservation 114: 289- 292. Descamps, S., Gauthier-Clerc, Le Bohec, C., Gendner, J. P. and Le Maho, Y. (2005). Impact of predation on king penguin Aptenodytes patagonicus in Crozet Archipelago. Polar Biology 28: 303-310. Diamond, A. W. and Devlin, C. M. (2003). Seabirds as indicators of changes in marine ecosystems: ecological monitoring on Machias seals island. Environmental Monitoring and Assessment 88: 153-175.

49 Doherty, P. F., Schreiber, E. A., Nichols, J. D., Hines, J. E., Link, W. A., Schenk, G. A. and Schreiber, R. W. (2004). Testing life history predictions in a long-lived seabird: a population matrix approach with improved parameter estimation. OIKOS 105: 606-618. Edgar, G., Samson, C. R. and Barrett, R. T. (2005). Species extinction in the marine environment: Tasmania as a regional example of overlooked losses in biodiversity. Conservation Biology 19: 1294-1300. Erikstad, K. E., Fauchald, P., Tverra, T. and Steen, H. (1998). On the cost of reproduction in long-lived birds: the influence of environmental variability. Ecology 79: 1781-1788. Forbes, L. S., Jajam, M. and Keiser, G. W. (2000). Habitat constraints and spatial in seabird colony distributions Ecography 23: 575-582. Forero, M. G., Tella, J. L., Hobson, K. A., Bertellotti, M. and Blanco, G. (2002). Conspecific food competition explains variability in colony size: A test in magellanic penguins. Ecology 83: 3466-3475. Furness, R. W. and Camphuysen, C. J. (1997). Seabirds as monitors of the marine environment. ICES Journal of Marine Science 54: 726-737. Gales, R. (1987). The use of otoliths as indicators of little penguin Eudyptula minor diet. Ibis 130: 418-426. Gales, R. (1988). Sexing adult blue penguins by external measurements. Notorinis 35: 71-75. Gales, R. (1988). The use of otoliths as indicators of Little Penguin Eudyptula minor diet. Ibis 130: 418-426. Gales, R. and Pemberton, D. (1990). Seasonal and local variation in the diet of the little penguin, Endyptula minor, in Tasmania. Australian Wildlife Research 17: 231-259. Gales, R. P. (1987). Validation of the stomach-flushing technique for obtaining stomach contents of penguins. Ibis 129: 335-343. Gardali, T., Ballard, G., Nur, N. and Geupel, G. R. (2000). Demography of a declining population of warbling vireos in coastal California. The Condor 102: 601-609. Giese, M. (1996). Effects of human activity on Adélie penguin Pygoscelis adeliae breeding success. Biological Conservation 75: 157-164. Giese, M., Goldsworthy, S. D., Gales, N., Brothers, N. P. and Hamill, J. (2000). Effects of the Iron Baron oil spill on little penguins (Eudyptula minor) III. Breeding success of rehabilitated oiled birds. Wildlife Research 27: 583-591. Goldsworthy, S. D., Giese, M., Gales, N., Brothers, N. P. and Hamill, J. (2000). Effects of the Iron Baron oil spill on little penguins (Eudyptula minor) II. Post- Release survival of rehabilitated birds. Wildlife Research 27: 573-582. Grémillet, D., Dell'Omo, G., Ryan, P. G., Peters, G., Ropert-Coudert, Y. and Weeks, S. J. (2004). Offshore diplomacy, or how seabirds mitigate intra-specific competition: a case study based on GPS tracking of Cape Gannets from neighbouring colonies. Marine Ecology Progress Series 268: 265-279. Healy, M., Chiaradia, A., Kirkwood, R. and Dann, P. (2004). Balance: a neglected factor when attaching external devices to penguins. Memoirs National Institute of Polar Research 58: 179-182.

50 Hobday, D. K. (1992). Abundance and distribution of pilchard and Australian anchovy as prey species for the little penguin Eudyptula minor at Phillip Island, Victoria. Emu 91: 342-354. Hunt, G. L., Eppley, Z. A. and Schneider, D. C. (1986). Reproductive performance of seabirds: the importance of population and colony size. Auk 103: 306-317.

Igual, J. M., Forero, M. G., Gomez, T., Orueta, J. F. and Oro, D. (2006). Rat control and breeding performance in Cory's shearwater (Calonectris diomedea): effects of poisoning effort and habitat features. Animal Conservation 9: 59-65. Jahncke, J., Checkley, D. M. and Hunt, G. L. (2004). Trends in carbon flux to seabirds in the Peruvian upwelling system: effects of wind and fisheries on population regulation. Fisheries Oceanography 13: 208-233. Jenouvrier, S., Barbraud, C., Cazelles, B. and Weimerskirch, H. (2005). Modelling population dynamics of seabirds: importance of the effects of climate fluctuations on breeding proportions. OIKOS 108: 511-522. Jenouvrier, S., Barbraud, C. and Weimerskirch, H. (2005). Long-term contrasted responses to climate of two Antarctic seabird species. Ecology 86: 2889-2903. Johannesen, E., Houston, D. and Russell, J. (2003). Increased survival and breeding performance of double breeders in little penguins Eudyptula minor, New Zealand: evidence for individual bird quality? Journal of Avian Biology 34: 198-210. Johnson, R. L., Venter, A., Bester, M. N. and Oosthuzien, W. H. (2006). Seabird predation by white shark Carcharodon carcharias, and Cape fur seal, Arctocephalus pusillus pusillus, at Dyer Island. South African Journal of Marine Science 36: 23-32. Klomp, N. I., Meathrel, C. E., Wienecke, B. C. and Wooller, R. D. (1991). Surface nesting by little penguins on Penguin Island, Western Australia. Emu 91: 190-193. Klomp, N. I. and Wooller, R. D. (1988). Diet of little penguins, Eudyptula minor, from Penguin Island, Western Australia. Australian Journal of Freshwater Resources 39: 633-639. Knight, C. and Rogers, T. (2004). Factors influencing fledging production in little penguins (Eudyptula minor). Wildlife Research 31: 339-344. Lewis, R. K. (1981). Seasonal upwelling along the south-eastern coastline of South Australia. Australian Journal of Freshwater Resources 32: 843- 54. Lewison, R. L., Crowder, L. B., Read, R. J. and Freeman, S. A. (2004). Understanding impacts of fisheries bycatch on marine megafauna. Trends in Ecology And Evolution 19: 598-604. Long, D. J. and Gilbert, L. (1997). California sea lion predation on chicks of the common murre. Journal of Field Ornithology 68: 152-154. Lu, C. C. and Ickeringill, R. (2002). Cephalopod beak identification and biomass estimation techniques: tools for dietary studies of southern Australian finfishes. . Museum Victoria Science Reports 5. Melbourne, Museum Victoria. Lyver, P. O. B. (2000). Identifying mammalian predators from bite marks: a tool for focusing wildlife protection. Mammal Review 30: 31-44.

51 Martin, J. L., Thibault, J. C. and Bretagnolle, V. (2000). Black rats, island characteristics, and colonial nesting birds in the Mediterranean: consequences of an ancient introduction. Conservation Biology 14: 1452-1466. McCann, K. S., Botsford, L. W. and Hasting, A. (2003). Differential response of marine populations to climate forcing. Canadian Journal of Aquatic Sciences 60: 971-985. Mecenero, S., Kirkman, S. P. and Roux, J. P. (2005). Seabirds in the diet of Cape fur seals Arctocephalus pusillus pusillus at three mainland breeding colonies in Namibia. African Journal of Marine Science 27: 509-512. Mickelson, M. J., Dann, P. and Cullen, J. M. (1992). Sea temperature in Bass Strait and breeding success of the little penguin Eudyptula minor at Phillip Island, South-eastern Australia. Emu 91: 355-368. Monaghan, P., Uttley, J. D. and Burns, M. D. (1992). Effect of changes in food availability on reproductive effort in arctic terns Sterna paradisaea. ADREA 80: 71-81. Montague, T. L. and Cullen, J. M. (1988). The diet of little penguin Eudyptula minor at Phillip Island, Victoria. Emu 88: 138-148. Norman, F. I. and Ward, S. J. (1992). Foods and aspects of growth in the Antarctic petrel and Southern fulmar breeding at Hop Island, Rauer Group, East Antarctica. EMU 92: 207-222.

Numata, M., Davis, L. S. and Renner, M. (2000). Prolonged foraging trips and eggs desertion in little penguins (Eudyptula minor). New Zealand Journal of Zoology 27: 277-289. O'Connor, R. J. (1984). The growth and development of birds. New York, Wiley. Olsson, O. (1997). Effects of food availability on fledging condition and post- fledging survival in king penguin chicks. Polar Biology 18: 161-165. Orians, G. H. and Pearson, N. E. (1979). On the theory of central place foraging. Columbus USA, Ohio State University Press. Page, B., McKenzie, J. and Goldsworthy, S. D. (2005). Dietary resource partitioning among sympatric New Zealand fur seals and Australian fur seals. Marine Ecology Progress Series 293: 283-302. Perriman, L., Houston, D., Steen, H. and Johannesen, E. (2000). Climate fluctuation effects on breeding of blue penguins (Eudyptula minor). New Zealand Journal of Zoology 27: 261-267. Quinn, G. P. and Keogh, M. J. (2002). Experimental design of data analysis for biologists. Cambridge, Cambridge University Press. Radl, A. and Culik, B. M. (1999). Foraging behaviour and reproductive success in magellanic penguins (Spheniscus magellanicus): a comparative study of two colonies in southern Chile. Marine Biology 133: 381-393. Ramos, J. A., Maul, A. M., Bowler, J., Wood, L., Threadgold, R., Johnson, S., Birch, D. and Walker, S. (2006). Annual variation in laying date and breeding success of brown noddies on Aride Island, Seychelles. EMU 106: 81-86. Reilly, P. (1994). Penguins of the World. Oxford, Oxford University Press Australia.

52 Reilly, P. and Cullen, J. M. (1981). The little penguin Eudyptula minor in Victoria, II: Breeding. Emu 81: 1-19. Reilly, P. and Cullen, J. M. (1983). The little penguin Eudyptula minor in Victoria, IV: moult. Emu 83: 94-98. Rey, A. R. and Schiavini, A. (2005). Inter-annual variation in the diet of female southern rockhopper penguin (Eudyptes chrysocome chrysocome) at Tierra del Fuego. Polar Biology 28: 132-141. Rindorf, A., Wanless, S. and Harris, M. P. (2000). Effects of changes in sandeel availability on the reproductive output of seabirds. Marine Ecology Progress Series 202: 241-252. Robinson, S., Chiaradia, A. and Hindell, M. A. (2005). The effect of body condition on the timing and success of breeding little penguins Eudyptula minor. Ibis 147: 483-489. Rogers, P. J., Eldershaw, G. and Walraven, E. (1995). Reproductive success of little penguins, Eudyptula minor, on Lion Island, New South Wales. Wildlife Research 22: 709-715. Ropert-Coudert, Y., Kato, A., Wilson, R. P. and Cannell, B. (2006). Foraging strategies and prey encounter rate of free-ranging Little Penguins. Marine Biology 149: 139-148. Safina, C., Burger, J., Gochfield, M. and Wagner, R. H. (1988). Evidence for prey limitation of common and Roseate tern reproduction. The Condor 90: 852-859. Schoener, T. (1968). The Anolis lizards of Bimini: resource partitioning in a complex fauna. Ecology 65: 1820-1827. Shaughnessy, P. D. and Dennis, T. (2001). Research on New Zealand fur seals and Australian sea lions in South Australia, 2000-2001. Report to National Parks and Wildlife South Australia. . Canberra, CSIRO. Shaughnessy, P. D., Goldsworthy, S. D. and Libke, J. A. (1995). Changes in the abundance of New Zealand fur seals, Arctocephalus forsteri, on Kangaroo Island, South Australia. Wildlife Research 22: 201-15. Shepherd, S. A. and Womersley, H. B. S. (1970). The sublittoral ecology of West Island, South Australia. Transactions of the Royal Society of Australia 94: 105-128. Stahel, C. and Gales, R. (1987). Little penguin: fairy penguins in Australia. Kensington, New South Wales University Press. Stapp, P. (2002). Stable isotopes reveal evidence of predation by ship rats on seabirds on the Shiant Islands, Scotland. The Journal of Applied Ecology 39: 831-840. Sterling, J. T. and Ream, R. R. (2004). At sea behaviour of juvenile male northern fur seals (Callorhinus ursinus). Canadian Journal of Zoology 82: 1621-1637. Thompson, P. M. and Ollason, J. C. (2001). Lagged effects of ocean climate change on fulmar population dynamics. Nature 413: 417-420. Velarde, E., Ezcurra, E., Cisneros-Mata. M. and Lavín, M. F. (2004). Seabird ecology, El Nino anomalies, and prediction of sardine fisheries in the Gulf of California. Ecological Applications 14: 607-615. Votier, S. C., Bearhop, S., Ratcliffe, N., Phillips, R. A. and Furness, R. W. (2004). Predation by great skuas at a large Shetland seabird colony. Journal of Applied Ecology 41: 1117-1128.

53 Wanless, S., Harris, M. and Greenstreet, S. P. R. (1998). Summer sandeel consumption by seabirds breeding in the Firth of Forth, south-east Scotland. ICES Journal of Marine Science 55: 1141-1151. Ward, T. and Staunton-Smith, J. (2002). Comparison of spawning patterns of fisheries biology of the sardine, Sardinops sagax, in temperate South Australia and sub-tropical southern Queensland. Fisheries Research 56: 37-49. Ward, T. M., Hoedt, F., McCleay, L., Dimmlich, W. F., Jackson, G., Rogers, P. J. and Jones, K. (2001). Have recent mass mortalities of the sardine Sardinops sagax facilitated an expansion in the distribution and abundance of the anchovy Engraulis australis in South Australia? Marine Ecology Progress Series 220: 241-251. Ward, T. M., McLeay, L. J., Dimmlich, W. F., Rogers, P. J., McClatchie, S., Matthews, R., Kampf, J. and Van Ruth, P. (2006). Pelagic ecology of a northern boundary current system: effects of upwelling on the production and distribution of sardine (Sardinops sagax), anchovy (Engraulis australis) and southern bluefin tuna (Thunnus maccoyii) in the Great Australian Bight. Fisheries Oceanography 15: 191-207. Weavers, B. W. (1992). Seasonal foraging ranges and travels at sea of little penguins Eudyptula minor, determined by radiotracking. Emu 91: 302- 317. Weerheim, M. S., Klomp, N. I., Brunsting, A. M. H. and Komdeur, J. (2003). Population size, breeding habitat and nest site distribution of little penguins (Eudyptula minor) on Montague Island, New South Wales. Wildlife Research 30: 151-157. Weimerskirch, H. and Cherel, Y. (1998). Feeding ecology of short-tailed shearwaters: breeding in Tasmania and foraging in the Antarctic? Marine Ecology Progress Series 167: 261-274. Weimerskirch, H., Gault, A. and Cherel, Y. (2005). Prey distribution and patchiness: factors in foraging success and efficiency of wandering albatrosses. Ecology 86: 2611-2622. Wickens, P. A., Japp, D. W., Shelton, P. A., Kriel, F., Goosen, P. C., Rose, B., Augustyn, C. J., Bross, C. A. R., J., P. A. and G., K. R. (1992). Seals and fisheries in South-Africa - competition and conflict. South African Journal of Marine Science 12: 773-789. Wienecke, B. C., Bradley, J. S. and Wooller, R. D. (2000). Annual and seasonal variation in growth rates of young little penguins Eudyptula minor in Western Australia. Emu 100: 139-147. Wilson, R. P. (1984). An improved stomach pump for penguins and other seabirds. Journal of Field Ornithology 55: 109-112. Wilson, R. P., Culik, B., Danfield, R. and Adelung, D. (1991). People in Antarctica - how much do Adelie penguins Pygoscelis adeliae care? Polar Biology 11: 363 - 370. Yorio, P. (2000). Breeding seabirds of Argentina: conservation tools for a more integrated and regional approach. EMU 100: 367-375. Yorio, P., Frere, E., Gandini, P. A. and Schiavini, A. (2001). Tourism and recreation at seabird breeding sites in Patagonia, Argentina: current concerns and future prospects. Bird Conservation International 11: 231 - 245.

10 Acknowledgements

The Fisheries Research and Development Corporation, the Nature Foundation of South Australia and the SA Department for the Environment Wildlife Conservation Fund provided funding for this project.

David Paton also provided academic advice and guidance on project design. Thanks to the fellow students in the GAB lab at SARDI for all of their help and encouragement. Especially Annelise Wiebkin, Al Baylis, Luke Einoder and Lachie McLeay. They showed me the ropes, gave advice on field techniques and data analyses, and they answered my calls for satellite downloads well into the night. I would especially like to thank Annelise who provided advice on field techniques and she was a wealth of knowledge when it came to little penguins.

Field work was carried out with the assistance of a number of eager volunteers, including Sam Adams Denise Bool, Leanne Trott, Augi Facelli, Beiha Malen Yanez, Toni Milne, Dave Turner, Heidi Valstar, Alex Pembshaw and Clare Adams.`

The dedicated staff and volunteers at Granite Island, Natalie Gilbert, Nick, Dorothy, Keith and Ruben provided assistance in the field, warm cups of coffee and plenty of encouragement. In particular, Natalie Gilbert provided me with her historical data, also passed on her knowledge and expertise of the little penguins on Granite and West Islands. I am very grateful for her support and assistance this year.