There are 18 species of penguin in the world, six of which breed regularly on the Antarctic Peninsula and across the Scotia Sea (Figure 1): Adélie [Pygoscelis adeliae], chinstrap [P. antarctica], gentoo [P. papua], king [Aptenodytes patagonicus], emperor [A. forsteri] and macaroni [Eudyptes chrysolophus]. Populations of penguins in this region have been changing over the past century (e.g., 1-2), with marked declines recorded for many colonies of macaroni, Adélie, and chinstrap penguins. For example, macaroni penguins have declined in SubArea 48.3 (3), and while Adélie and chinstrap penguins are stable or increasing in SubArea 48.1 in Marguerite Bay (2, 4), they have declined throughout most of the western Antarctic Peninsula (WAP) to the north of Marguerite Bay (2). Adélie penguins are also stable or increasing in the Weddell Sea region, as well as in SubArea 48.4 (4). Gentoo penguins are increasing in abundance on the WAP in 48.1 and at the South Orkney Islands in 48.2 and are expanding their breeding range to the south (2); elsewhere in the region, gentoo populations do not appear to be trending significantly (4), though large interannual fluctuations in gentoo abundance can complicate trend assessment. King penguins have increased in SubArea 48.3 and appear to be extending their range south to the South Shetland Islands in 48.1 (5), and whilst emperors have now disappeared from the Dion Islands on the WAP (6) they continue to breed near Snow Hill Island. These findings highlight considerable spatial heterogeneity in penguin population trends and the importance of comparative work across long-term study sites.
Figure 1: Antarctic Peninsula and Scotia Arc. The pink area represents that portion of the western Antarctic Peninsula where Adélie and chinstrap penguins have declined in abundance and gentoo penguins have increased in abundance; yellow areas represent where Adélies and chinstraps are either stable or increasing in abundance (NB: no chinstraps are found in the Weddell Sea area or in Marguerite Bay). The green shading indicates where macaroni penguins have experienced substantial declines. The FAO statistical SubArea boundaries used by CCAMLR are shown in black.
Existing datasets on Antarctic Peninsula and Scotia Arc penguins
Long-term datasets (20 years or longer) on abundance, diet, demographics, and phenology include studies from the South Shetland Islands (e.g. 1,7,8), Port Lockroy (9), Anvers Island and Biscoe Point (10), Cierva Point (11), Signy Island (12), Avian Island (13) and Bird Island (3). Operating across the Antarctic Peninsula and South Orkney Islands, the Antarctic Site Inventory program (4) has surveyed penguin nests and chicks (and other seabirds) since 1994 (4) using tourist vessels to provide a broader geographic coverage complementary to that provided by other long-term research programmes.
The most up-to-date and most comprehensive population data for Antarctic penguins comes from the MAPPPD project (4). This database includes all published data on Antarctic penguin populations, including historical data held by the British Antarctic Survey, the Antarctic Site Inventory program, the Australian Antarctic Division, and other national programs. MAPPPD data also include population estimates derived from very high resolution satellite imagery for Adélie penguins (14), drone counts, ground photographs, and offshore vessel counts.
In Subarea 48.1, the most up-to-date counts in MAPPPD (extract made on 14 September 2018, but updated with recent data 15) for 88 extant Adélie penguin breeding sites, show that 98.5% of the regional population has been surveyed since 2004. For chinstrap penguins, of 233 existing breeding sites, only 10.2% of the population has been surveyed since 2004, and that most of the population has not been counted in over 30 years. For gentoos 83.4% of the population across 101 breeding sites have been surveyed since 2004, but most of the remainder have counts older than 30 years.
In Subarea 48.2, 81.4% of the Adélie population across 20 sites has been surveyed since 2004. For chinstraps, 46.7% of the population across 79 sites has been surveyed since 2004, but with most of the population not having been surveyed in over 30 years. Across six gentoo breeding sites, only one colony, representing 23.4% of the population, has been surveyed since 2004.
Up-to-date data are vital for management purposes, including for managing the krill fishery, for understanding the regional impacts of climate change, and for understanding how ecosystems change, given the ongoing recovery of marine mammals. New tools (e.g. 16, 14) are now becoming urgent if managers are to adequately understand penguin population trajectories.
Drivers of population change
Several hypotheses have been suggested to explain why penguin populations are changing, including: (i) ecosystem changes associated with regional climate change; (ii) population changes as marine mammals recover from historical harvesting; (iii) the commercial fishery for Antarctic krill [Euphausia superba]; and (iv) impacts from tourism.
Environmental variability and change affect penguin population processes either directly through physiological impact, or indirectly via the distribution or availability of their mid-trophic-level prey. On short time scales (days to years), breeding behaviour, as well as breeding success and survivorship rates, can be affected by random events (such as unusual storm events), seasonal variability, or interannual variability in environmental conditions. Over longer periods of time (years to decades), persistent changes can drive shifts in distribution and abundance, which are easily detected by long-term survey programmes (2, 17, 18).
The continuing effects of historical sealing, whaling and fishing, make it difficult to determine the influence of climate change per se (19-20). Indeed, it is highly probable that no unifying solution will be found to account for all observed changes in penguin populations. Different drivers of change may be impacting different species and operating at different temporal and spatial scales that range from that of a single colony, to the regional scale, or even to the circumpolar scale. Nevertheless, there is some compelling evidence that indicates climate change can have a direct effect on penguin abundance and distribution. Increased snowfall resulting from increased warm, wet conditions may have contributed towards Adélie population declines close to Palmer Station (21), a feature that may also affect populations in the South Shetlands. Colonies that have experienced more snow accumulation have decreased more rapidly than penguin colonies where wind scour abates snow accumulation. Similarly, gentoo penguins breed in areas with <50% sea ice concentration in the austral spring and following the decline in spring sea ice along the WAP, they have expanded their breeding range southward into new areas (2).
Predicting and monitoring future change
Disentangling the effects of climate change is complex as it might potentially impact penguins at any or all of their breeding sites, foraging sites, moulting sites or wintering sites. A broader understanding of penguin life-history and their ecological interactions is necessary. Long-term studies of fixed populations have played an essential role in our current understanding of changes whilst opportunistic vessel-based surveys (e.g. 4) have provided complementary data on penguin abundance and distribution. New technologies, such as remote sensing imagery (14,16), unmanned-aerial vehicles (UAVs), animal tracking devices (satellite- or light-based geolocators), and automated individual identification systems (radio frequency identification), as well as remote automated camera systems, are now providing important details about abundance, life cycle, demography and distribution on land and at sea (foraging and migration), at local and regional scales. Integration of these disparate data streams, and their analysis alongside complementary datasets on krill abundance (and availability to foraging penguins) and environmental conditions (e.g. sea ice), will be required to predict better the future of penguins along the Antarctic Peninsula. Focused study regions, including reference areas where fishing is restricted, would help reduce the complexity of interacting drivers (12).