Introduction
The Adélie penguin (Pygoscelis adeliae) is a widely distributed and abundant penguin species in the Antarctic (Woehler Reference Woehler1993), and as a consumer of krill and fish is an important component of the Antarctic marine ecosystem. There is considerable interest in the status and trends of Adélie penguin breeding populations because they are considered likely to reflect the influence of human-induced perturbations such as fishing (Convention for the Conservation of Antarctic Marine Living Resources (CCAMLR) 2003) and climate change (Croxall and others Reference Croxall, Trathan and Murphy2002, Weimerskirch and others Reference Weimerskirch, Inchausti, Guinet and Barbraud2003, Ainley and others in press) on the ecosystem.
Given the broad distribution of the species, and the remoteness and inaccessibility of much of its breeding habitat, it has not been possible to undertake a comprehensive, synoptic survey of the circumpolar distribution of breeding sites, nor, consequently, of abundance. Instead, knowledge of circumpolar distribution and abundance has been arrived at by collating the results of many separate, independent surveys undertaken over many years at smaller scales (Woehler Reference Woehler1993; Woehler and Croxall Reference Woehler and Croxall1997; Ainley Reference Ainley2002). Woehler (Reference Woehler1993), for example, derived a circumpolar abundance estimate of 2.5 million breeding Adélie penguin pairs from count data published in numerous reports in the primary and secondary literature prior to 1993, but noted that the estimate should be considered a minimum because there were likely to be many areas and sites for which no data existed.
The Mac.Robertson Land coast between longitudes 62–64°E has many hundreds of islands offering potential breeding habitat for Adélie penguins. Prior to this study the number of breeding sites known to exist along this section of coast was 21 (Horne Reference Horne1983; Woehler and others Reference Woehler, Johnstone and Burton1989). However, only one of these 21 breeding sites occurs in the many scores of islands in the eastern half of this section of coast (that is between longitudes 63–64°E), collectively known as the Robinson Group of islands, hereinafter cited as the Robinson Group (Fig. 1). Horne (Reference Horne1983) gives a count of 434 chicks at Macey Island on 11 February 1972 by the Australian National Antarctic Research Expedition (ANARE). From inspection of published information and collations thereof it is difficult to determine whether the Robinson Group is largely unoccupied by breeding Adélie penguins, or whether it is largely unsurveyed: as the extent of search effort is sometimes not reported and findings of absences are rarely reported in the literature. However, the recent finding of a hand-drawn map of a portion of the Robinson Group by a member of the 1972/73 ANARE (K. Kerry unpublished data) indicated that the published record of breeding sites was definitely incomplete. This motivated the first comprehensive survey of breeding sites in the region, which is reported in this paper.
Map of the Robinson Group of islands. Islands at which breeding Adélie penguins were present or absent are filled in black or grey respectively. The large area of grey at the bottom of the figure is continental ice.
Methods
During 21–26 November 2005, a 380 km2 area off the Mac.Robertson Land coast, east Antarctica was surveyed for islands occupied by breeding Adélie penguins. This area, known as the Robinson Group, is a collection of approximately 150 islands lying between 25 and 55 km to the east of Australia's Mawson Station (67.602°S, 62.879°E). The survey encompassed the known extent of these islands from Child Rocks (67.427°S, 63.246°E) to an un-named island 5.2 km to the south-east of Austskjera (67.548°S, 64.106°E), and from the edge of the continent out to the Douglas Islands (67.380°S, 63.389°E) 15 km to the north. During the survey the party was based at a small field hut on Macey Island (67.438°S, 63.818°E), which is situated towards the eastern edge of the Robinson Group. Observations at Bechervaise Island near Mawson station indicate that, at the time of year during which the survey was conducted, all breeding penguins have established nests which are occupied by at least one breeding adult (Kerry and others Reference Kerry, Clarke, Else, Dann, Norman and Reilly1995; C. Southwell unpublished data).
At the time of the survey, the sea ice conditions were very good, enabling travel between the islands using 4WD quad bikes (all-terrain vehicles) equipped for sea ice conditions. The islands in this region are patchily distributed into small clusters of islands (in this paper referred to as sub-Groups) and a series of individual islands. From maps and satellite images, 27 waypoints were created using a geographic information system (GIS; Oziexplorer version 3.90.3) and these waypoints were used to navigate to all areas in which islands were known to exist. In addition, tracks were plotted between these survey points that allowed visual scanning for islands in areas that would not otherwise be visited. Because of the excellent field conditions during the survey, surrounding islands and island sub-Groups could be seen from each survey location. This permitted the identification and survey of a number of islands not recorded on existing maps. In situations in which uncertainty existed concerning whether an island had been visited before, the island was approached and GPS was checked concerning the location of the nearest logged waypoint. From this the status of each island was determined (surveyed/unsurveyed). Because of this systematic approach, there was confidence that all islands with ice-free areas on them were visited.
The survey of the Robinson Group was conducted by navigating to each island sub-Group and systematically moving across the sub-Group in one direction until all islands in that location had been visited. Each island was climbed, and a GPS waypoint recorded (on a Garmin 12XL GPS receiver) at the island's approximate geographical centre. Satellite coverage was very good during the survey, with an estimated position error of ∼ 5 m achieved for all waypoints. In cases where the entire island could not be seen from the summit, it was systematically searched for evidence of Adélie penguins. An island was recorded as being a breeding site for Adélie penguins if there was at least one penguin on a nest at the time of the survey. Thus, our definition of ‘breeding site’ follows Ainley (Reference Ainley2002) who also uses the term ‘breeding locality’. This is a convenient geographical label describing where penguins breed (in our case, an island), with each site potentially containing multiple subcolonies, or itself being a part of a larger penguin colony (see Ainley Reference Ainley2002 for a discussion of these terms). While endeavours were made to plot a separate waypoint for each island, because of the topography of snow and ice it was sometimes difficult to know if two rocks separated by ice were one island or two. Also, in some instances difficulty was experienced in locating islands of low elevation because they were completely covered in ice. Thus, it is possible that some small islands were missed because of complete ice cover, and it is also possible that two rocks counted as separate islands were, in fact, joined. Regardless of these minor concerns, in such cases no penguins would breed on these ice-covered rocks and so the survey still comprehensively assessed all potential breeding sites for Adélie penguins in the region.
On each island where penguins were found, a rough estimate of the size of that population was recorded as a future reference for allocating resources should a more precise abundance survey be undertaken in the future. The categories employed (10's, 100's, 1000's) may range up to an order of magnitude and should not be relied upon for population estimates.
Results
149 islands were surveyed and a GPS location was recorded at the approximate centre of 142 of them. The seven islands for which a central GPS location was not recorded were small ice-covered islands, usually no more than 50 m across. In these instances the state of the sea ice surrounding them or the smooth ice covering them prevented access, and it was possible to obtain a GPS location only for the closest visited point on the surrounding sea ice. Because these seven islands were completely covered by ice in November, there was confidence that none of these islands were occupied by breeding Adélie penguins. Of the 142 islands to which access was possible, 30 islands were occupied by breeding Adélie penguins (Table 1) and 112 had no evidence of breeding Adélie penguins (Table 2). The number of nests on each occupied island ranged from fewer than 10 (for example 67.44019°S, 63.54346°E) to several thousand (for example Macklin Island). With the exception of the area to the south and east of the Macey Islands, in which there were no breeding penguins, the islands containing breeding Adélies were located throughout the Robinson Group.
Position (in decimal degrees) of the 30 surveyed islands where breeding Adélie penguins were found. Names of individual islands or island sub-Groups are given if known. See the text for information on the definitions of population size categories
Position (in decimal degrees) of the 112 surveyed islands that were accessed and found to have no breeding Adélie penguins. Also shown are the approximate positions of seven ice-covered islands that could not be accessed (*)
Discussion
This study adds significantly to our knowledge of the distribution of breeding Adélie penguins in the Mac.Robertson Land coast between 62–64°E by more than doubling the number of known breeding sites from 21 to 51. Until now, only a single published (and, therefore, accessible) record of Adélie penguins breeding in the Robinson Group was known (Horne Reference Horne1983). While the current study was restricted to a reconnaissance survey and did not attempt to obtain accurate counts of breeding penguins, it could provide a platform for future survey work in this area.
Two important aspects of the current study should be kept in mind for future survey work. Firstly an attempt was made to survey comprehensively a pre-defined area in which there was no presupposition that penguins would be found. If reliance had been placed only on historical records of known breeding localities when designing a future survey to estimate breeding population size for this region, the resulting population estimate would have been severely biased in a negative sense. Thus, it is possible that in areas previously surveyed opportunistically, only the breeding localities that are easily accessible or are in the areas commonly traversed are found; this may, or may not, include all populations in that region. Secondly, all islands that were visited during the survey were recorded in order to facilitate any future survey work in this area. If only those locations at which breeding populations of penguins were found, without a clear indication of where penguins were absent, it would difficult for a reader to make a correct interpretation of the search effort, as it would be unclear whether the breeding locations that were found were the only ones in the area, or were the only ones that were visited. For future surveys the unambiguous reporting of all areas visited is encouraged with the recording of which areas did or did not contain the species of interest. From these data, a more accurate record of the distribution and abundance of species can be generated.
Acknowledgements
We thank David Smith and Henk Brolsma from the Australian Antarctic Division Data Centre for assistance with the survey maps, Angela Bender for her help in generating Fig. 1, and David Ainley and an anonymous reviewer for comments on the manuscript. We also thank the 2005 summer crew from Mawson station for logistical support: in particular Greg Barras and Ted Mead for their assistance during the initial part of the survey.
