Polar Biol (2010) 33:493–503 DOI 10.1007/s00300-009-0726-y ORIGINAL PAPER Population trends and reproductive success at a frequently visited penguin colony on the western Antarctic Peninsula Heather J. Lynch Æ William F. Fagan Æ Ron Naveen Received: 16 June 2009 / Revised: 8 September 2009 / Accepted: 9 September 2009 / Published online: 25 September 2009 Ó Springer-Verlag 2009 Abstract Petermann Island (65°100S, 64°100W), one of Introduction the Antarctic Peninsula’s most frequently visited locations, is at the epicenter of a rapid shift in which an Ade´lie Despite its small size (*1 mi long and mi wide), penguin dominated fauna is becoming gentoo penguin Petermann Island (65°100S, 64°100W; Fig. 1) represents an dominated. Over the course of five seasons, the breeding important breeding area for both penguins and flying birds productivity of Ade´lie and gentoo penguins breeding at on the Antarctic Peninsula. Resident species include gentoo Petermann Island were monitored to identify drivers of this penguins (Pygoscelis papua), Ade´lie penguins (P. adeliae), rapid community change. The impact of tourist visitation blue-eyed shags (Phalacrocorax atriceps), south polar on breeding success was also investigated. Consistent with skuas (Stercorarius maccormicki), brown skuas (S. ant- larger trends in this region, the Ade´lie penguin population arctica), south polar 9 brown skua hybrids, kelp gulls decreased by 29% and the gentoo penguin population (Larus dominicanus), Wilson’s storm petrels (Oceanites increased by 27% between the 2003/2004 and 2007/2008 oceanicus), and snowy sheathbills (Chionis alba). Prior to seasons. Reproductive success among Ade´lie penguins the recent southward expansion of gentoo penguin breeding ranged from 1.09 to 1.32 cre`ched chicks/nest, which was (Lynch et al. 2008), Petermann Island was also the south- higher than or comparable to other sites and is an unlikely ernmost documented breeding site for gentoo penguins explanation for the precipitous decline of Ade´lie penguins (Croxall and Kirkwood 1979). Although no longer the at Petermann Island. Whereas gentoo penguin reproductive southernmost gentoo colony, Petermann Island remains success was lowest in colonies frequently visited by tour- the fourth largest gentoo penguin colony on the western ists, Ade´lie penguin colonies frequently visited by tourists Antarctic Peninsula [after Cuverville Island (64°410S, had higher reproductive success than those visited only 62°380W), Duroch Island (63°180S, 57°540W), and Gerlache occasionally. These results are placed in the context of Island (64°350S, 64°160W); H. J. Lynch, unpublished data] other studies on reproductive success and the impact of and probably represents an important source of migrants to tourist visitation on breeding colonies of Ade´lie and gentoo new colonies further south. In addition to its biological penguins. importance, Petermann Island is also the sixth most fre- quently visited penguin colony on the Antarctic Peninsula Keywords Ade´lie penguin Á Gentoo penguin Á (International Association of Antarctic Tour Operators Long-term monitoring Á Breeding success Á Tourism 2008), and thus presents an opportunity to assess the potential impacts of tourist visitation on the breeding suc- cess and population trajectories of Antarctic penguins. For H. J. Lynch (&) Á W. F. Fagan Department of Biology, University of Maryland, these reasons, Oceanites Inc. (a non-profit research orga- College Park, MD 20742, USA nization monitoring breeding birds on the Antarctic e-mail: [email protected] Peninsula) initiated a 5-year census and reproductive study in 2003 with the goal of understanding the relative fates of H. J. Lynch Á R. Naveen Oceanites Inc., P.O. Box 15259, Chevy Chase, the increasing gentoo population and the declining Ade´lie MD 20825, USA population (Lynch et al. 2008). The principal goals of the 123 494 Polar Biol (2010) 33:493–503 project were: (1) to obtain nest and chick census counts that penguin colonies and site-wide censuses of the breeding conformed to the standards defined by the CCAMLR Eco- penguins. In the austral spring of the succeeding three system Monitoring Program (Scientific Committee for the seasons of the study (2005/2006–2007/2008), a total of 60 Conservation of Antarctic Marine Living Resources 2004), Ade´lie nests in 12 colonies and 100 gentoo nests in 20 and (2) to determine if breeding performance was related to colonies were chosen after nest establishment but before year, colony, or the level of visitation each colony received. clutch initiation for the reproductive study. (Note that we Recently, Trathan et al. (2008) published the results of use the word ‘‘colony’’ to describe a contiguously nesting their research on the gentoo penguin colony at Port Lockroy, group of penguins. Penguin colonies, which may include and comparing results from that study with our own presents more than one species, are separated by areas of unoccu- an opportunity to compare two sites on the western Antarctic pied bare ground, and there are many colonies at the ‘‘site’’ Peninsula with a mix of gentoo colonies receiving high levels Petermann Island.) A group of five contiguous nests were of visitation and those experiencing little or no visitation. We randomly selected from within the colony, and the location note that although Port Lockroy receives more landed pas- of each nest relative to the group was photographed and sengers than Petermann Island (16,398 vs. 11,922 in 2007/ marked by a painted stone placed in the center of the 2008), both sites experience similar temporal patterns of group. These same five nests were then tracked throughout visitation, with visitation beginning and ending within a few the remainder of the breeding season. Nests were selected days of each other at the two sites. Therefore, the timing and equally from both small (B25 nests) and large ([25 nests) intensity of visitation at the two sites with respect to the colonies to assess the importance of colony size in popu- breeding phenology of the penguins involved is similar. lation trend and reproductive success. As Ade´lie penguin Unlike the study at Port Lockroy, Petermann Island also reproductive success has been linked to proximity to col- offers the opportunity to assess the reproductive perfor- ony edge (Spurr 1975; Davis and McCaffrey 1986), we mance of Ade´lie penguins, a species that has been in sharp also recorded whether Ade´lie nests were ‘‘peripheral’’ (i.e. decline on the western Antarctic Peninsula (Ducklow et al. at the colony edge, or one adjacent to a nest at the colony 2007; Lynch et al. 2008; McClintock et al. 2008) and which edge) or ‘‘central’’ (i.e. at least two nests away from the has gone locally extinct at some sites (Rejcek 2008). colony edge). Reproductive study nests were visited daily to assess nest contents and track the fate of reproductive effort. Observation of nest contents was passive (i.e. pen- Methods guins were observed from a distance and were not manipulated in any way) and, whenever possible, observ- The first two seasons of the Petermann Island study (2003/ ers maintained at least 5 m distance from the penguin 2004 to 2004/2005) focused on mapping the island’s colony to minimize unnecessary disturbance. In this way, Northern portion of the island not shown. 0 55 0o 0/ W// N meters 150 60 o0 /0 //W o / // 65 0 0 S Port Lockroy 65 o 0 /0 //W Petermann Island o / // 70 0 0 S Adelie only 70 o 0 / 0 // W gentoo only Both Adelie and gentoo Fig. 1 Map of Petermann Island 123 Polar Biol (2010) 33:493–503 495 observers determined the dates of clutch initiation and groups of passengers. Occasionally Visited Sites were completion for each nest in the study; daily observations those that were fairly easy to access by those passengers ensured that dates were accurate to within 1 day. Egg loss exploring the island separate from a tour group but were and any re-laying events were also recorded. Once each of not areas to which groups of passengers were explicitly the five nests in the group had complete clutches, the group lead by expedition staff. We note that these categories were was visited every 2 days to determine nest contents and chosen prior to and separate from the analysis regarding the record any egg losses or nest failures. Within a week of the relative breeding performance of different areas of the peak of egg laying for each of the two species, a complete island. site-wide census count of nests for that species was com- We considered nine different models for breeding pro- pleted for the island. In the days approaching egg hatching, ductivity [yij; observation j in group (Colony, Year, Type) nests were once again monitored daily so that the exact i], including an intercept-only model (Model 1) which dates of egg hatching could be established for each of the pools all of the data to estimate site-wide mean produc- study nests. For the remainder of the breeding season, nests tivity (b). Models 2–4 are fixed effects models with Colony were observed every other day to record any chick mor- (C), Year (Y), and Colony Type (T; representing the rel- tality and, eventually, the initiation of chick cre`ching ative extent of visitation at each colony) as the fixed within the colony. Within a week of the peak of chick effects, respectively. Models 5–7 are random effects cre`ching for each species, a site-wide census of chicks was models in which Colony (C), Year (Y), and Colony Type completed. It is important to emphasize that the principal (T) are modeled as random effects accounting for vari- goal of these reproductive study nests was to establish the ability around the population mean (b).