Emperor Penguins in a Changing World: Unveiling Current Trends and Predicting Future Scenarios

Emperor penguins (Aptenodytes forsteri), have a specialized life history that is intimately tied to Antarctic sea ice, allowing them to survive through six months of darkness and freezing temperatures in the coldest and driest place on Earth (Ainley et al. 2010, see video the life cycle of emperor penguins is tied to sea ice). They rely on sea ice to breed, rest, moult, feed, and seek refuge from predators. Their relationship with sea ice is finely balanced; insufficient sea ice impacts marine foodwebs and reduces food availability, and also means that penguins can’t find a stable breeding habitat (particularly fast ice, which is sea ice that is attached to the shore or to ice shelves, Massom et al. 2009); whilst too much sea ice increases the travel distance between nesting and feeding grounds, limiting their ability to adequately feed their chicks (see Antarctic Sea Ice #1: Physical Role and Function).

The emperor penguin is the only warm-blooded species to breed in Antarctica during the depths of winter. Penguins arrive at their breeding sites in March or April after the start of austral autumn sea ice formation (Prevost 1961). In mid-May the female lays a single egg and then makes a long foraging trip to open water to feed. The male incubates the single egg for 65 to 75 days throughout the Antarctic winter while the female is away. During this period the male must survive without eating. By the time the female penguins return to feed their newly-hatched chicks at the end of winter, the males have lost almost half their body weight after fasting for four months.

To survive the cold, and the frequent violent storms and blizzards during the winter, the male penguins huddle together for warmth. Hundreds of penguins cluster together, continually shifting their positions within the huddle. Thus, each individual continually attempts to avoid the colder external part of the huddle, benefiting from the warmer center parts. This coordinated movement contributes to sustaining an ambient temperature within the huddle above 0 °C, even in the face of an average external temperature of -17 °C; the sheltered inside of the huddle can reach 37.5 °C (Gilbert et al. 2006).

The females return in August, just after the chicks hatch. Both parents then take turns, alternating between foraging and provisioning the chick, and also keeping it warm. By September, both adults must leave their chick simultaneously so that they can forage to meet the chick’s growing demands. In December, adults and chicks leave the colony and return to the ocean. The adults must then feed before they moult to replace all their feathers. During the moult, the adults do not feed and require a stable sea ice platform as their plumage is no longer adequate to insulate them from the cold water.

To feed, emperor penguins can dive for almost 30 minutes on a single breath, leading to dives that are longer and deeper (up to  a maximum of  546m, though most are shallower (Wienecke et al. 2007)) than any other bird species. Emperor penguins primarily feed on Antarctic silverfish, Antarctic krill, Arrow squid, and Antarctic neosquid, with sea ice conditions strongly influencing both krill and silverfish availability (see review by Trathan et al. 2020).

Knowledge about emperor penguin population trends has improved thanks to both long-term studies (Barbraud and Weimerskirch 2001, Wienecke 2010) and advances in high-resolution satellite remote sensing allowing the detection of emperor penguin colonies (Fretwell and Trathan, 2009).

Emperor penguin life cycle is tied to sea ice: implications for fluctuations in colony size

Figure 1. Map illustrating the location of existing colonies and their average population sizes from 2009 to 2018. Colony numbers refer to the site names listed in Table 1. The colonies highlighted in red represent those with available population size estimates according to La Rue et al. 2024. Colonies marked in green are additional locations mentioned in the literature, including references from Fretwell and Trathan 2020 and Fretwell 2024. The graph inset in the Antarctic continent displays the index of global abundance of emperor penguins, calculated annually (La Rue et al 2024). This calculation involves combining the area of “penguin pixels” from Very High-Resolution (VHR) imagery within a Bayesian modeling framework over the study period from 2009 to 2018. The abundance indices of emperor penguins changed from approximately 252,000 birds in attendance during springtime in 2009 to approximately 226,000 birds in 2018. The red ribbon in the graph represents the 95% equal-tailed credible interval for the annual index, while the central red line denotes the median of the posterior distribution. The dotted black line shows the linear trend.

Table 1. Description of the location of emperor penguin colonies and their averaged population index of abundance according to La Rue et al. 2024. The numbers refer to Figure 1.

Following the start of routine satellite passive-microwave monitoring in 1979, Antarctic-wide sea ice extent displayed a slight net increase until 2016, though trends varied markedly by region and season (Parkinson 2019). However, in recent years there have been dramatic decreases in sea ice extent to record low levels (Turner et al. 2022, see Antarctic Sea Ice #3: Trends and Future Projections). Changing atmospheric and oceanic factors, including faster winds and ocean freshening and warming (Hobbs et al. 2016, Purich and Doddridge 2023), have influenced these sea ice trends.

Emperor penguin mortality and reproductive output are both significantly impacted by sea ice conditions, and there is an optimal sea-ice range within which populations thrive (Jenouvrier et al. 2012). For example, in the late 1970s the population at the Pointe Géologie colony (#38 Figure 1) declined by half when the mortality of males increased during consecutive years of low sea ice extent (Barbraud and Weimerskirch 2001, Jenouvrier et al. 2005). Low sea ice extent reduces the abundance of key prey species and males are more vulnerable to this than females because of their long winter fast. Moreover, the Pointe Géologie colony’s population has also experienced massive breeding failures in 1982, 1990, 1992, 1994, 1995, 2013, and 2014 (Jenouvrier et al. 2009, Barbraud et al. 2015) when the extensive sea ice extent increased the distance between the nearest open water and the colony during the chick-rearing period (Massom et al. 2009; Labrousse et al. 2021). Such population impacts are important as the emperor penguin has a generation time of 16 years (Jenouvrier et al. 2014), so disturbances to breeding and reproductive output take a long time to stabilize.

The year-to-year changes in population size at a given colony varies markedly, with some colonies having dramatic declines whilst nearby colonies have notable increases. Not all such variation can be easily understood (e.g. Barber-Mayer et al. 2008), but some major changes in population can be linked to major physical perturbations (e.g. Kooyman et al. 2007). For example, at the world’s second largest emperor penguin colony (Halley Bay, #4 Figure 1), the breeding population decreased to ~zero in 2015, with consequences for the nearby Dawson-Lambton colony (#5 Figure 1), only 55 km to the south, which subsequently experienced a more than tenfold increase in penguin numbers (Fretwell and Trathan 2019). Halley Bay suffered three years of almost total breeding failure linked to a shift in atmospheric conditions and hence sea ice. Those poor conditions almost certainly led to penguins relocating, with most moving to Dawson-Lambton over the ensuing years. Elsewhere, penguins are also thought to shift between breeding sites in some years (Kooyman and Ponganis, 2017), whilst some colonies disappear in some years and reappear in the following years (Fretwell and Trathan 2021), highlighting the complexity of the factors driving population trends and habitat suitability for the emperor penguin.

Pan-Antarctic Distribution of Emperor Penguins and Global Population Dynamics

Seventy colonies of emperor penguins are currently identified around Antarctica (Fretwell and Trathan 2019, Fretwell and Trathan 2021, Fretwell 2024, Figure 1 and Table 1). Utilizing circumpolar high-resolution satellite imagery, defined fast ice metrics, and various geographical and biological factors, Labrousse et al. (2023) reported that there are no discernible differences between the fast ice habitats occupied by penguin colonies and those lacking them. Nevertheless, persistent polynyas (Massom et al. 1998), sea ice cracks, flaw leads, and ephemeral short-term polynyas (Labrousse et al. 2019), may be key habitat features providing access to potential foraging habitats for emperor penguins.

Recently, the discovery of new colonies has increased the estimate of the global population, whilst also expanding the geographic scale of monitoring, including now in areas offshore from the coastal margins (Fretwell and Trathan 2021, Fretwell 2024). In addition, LaRue et al. (2024) have combined remote sensing, validation surveys, and Bayesian modeling to assess the population size and trajectory of adult emperor penguins across the entire population range. The new analysis estimates a global index of abundance of 252,000 breeding adults, with an estimated median decrease in the population of 9.6% (95% credible interval -26.4% to +9.4%) between 2009 and 2018 (LaRue et al. 2024). There are significant year-to-year fluctuations around this trend, exhibiting first the decline, but followed by an apparent population increase toward the latter part of the time series, however, the modelling suggest an 81% likelihood of an overall decline. In addition, the prospect of emigration to and from colonies outside the 50 colonies examined by the LaRue et al. (2024) study could contribute to the observed decrease over the period 2009-2018. This and other explanations cannot be entirely discounted. The ecological mechanisms behind the observed changes are not fully understood, emphasizing the need for urgent efforts for more extensive monitoring through international collaboration.

This newly developed population modelling accounts for distinct colony-level trends, occasional colony disappearance and reappearance, and daily variation in abundance during the spring survey period. The modeling also addresses imperfections in counts from aerial surveys and imprecise satellite observations of occupied areas, considering the potential changes in expected counts over the survey period. Importantly, the new index includes only a proportion of the population present in late spring (the breeding adults), and so provides a conservative estimate of the total number of adult emperor penguins because it does not include failed breeders and young birds that are yet to start breeding.

This index reveals an average annual global population decline of -1.3% per year between 2009 and 2018 (95% CI = -3.3% to +1.0%) and suggests that declines are likely to have occurred in four out of eight different fast ice regions, regardless of habitat conditions. The majority of emperor penguin breeding sites occur on fast ice and if fast ice breaks up too early in the season before chicks have fully fledged (grown their waterproof feathers), they can drown in the water, leading to massive breeding failures (Fretwell and Trathan 2019; Fretwell et al. 2023). Accurate measurements of fast ice extent remain challenging. Information on large-scale change and variability in circum-Antarctic fast-ice coverage is derived from analysis of satellite image time series (Fraser et al. 2021). These and other studies show the extent, persistence and seasonality of the fast-ice habitat to be influenced by the presence/absence of a protective pack-ice “buffer” (Massom et al. 2009; Fraser et al. 2021). In the long-term, sea ice at the large-scale likely determines the extent of fast ice as a breeding platform (Fraser et al. 2021).

Under future climate change scenarios with warmer global temperatures (IPCC, 2023- Lee at al. 2023), Antarctic sea ice will form later and break up earlier than it does today. Despite uncertainties regarding the magnitude of future changes in sea ice, coupled Earth system models (ESMs) consistently project the loss of Antarctic sea ice throughout all seasons (Fox-Kemper et al. 2021).

Penguin population projections are determined using state-of-the-art demographic models that include the complex relationships between reproduction, mortality, and sea ice throughout the entire life cycle of emperor penguins and are intricately linked to sea ice projections derived from ESMs (Jenouvrier et al. 2012, 2013, 2014, 2017, 2020,  2021). These demographic models do not currently incorporate fast ice’s impact on reproduction, because fast ice projections are not currently available from any ESM, however, the stability of fast ice is strongly influenced by adjacent sea ice (Fraser et al. 2021), so as sea ice declines, it is likely fast ice will also. The anticipated decline in sea ice is virtually certain to result in a reduction of the global emperor penguin population (Jenouvrier et al. 2014, 2021). Indeed, every emperor penguin colony will likely decline by the end of the century, regardless of the greenhouse gas emission scenario, with or without emission mitigation.

Figure 2. (a) Landsat8 image from October 2016 showing the location of the Halley Bay colony at Windy Creek, together with the location of the Dawson-Lambton Colony on the Dawson-Lambton Glacier. Area around Windy Creek on (b) 3 November 2016; (c) 17 November 2016; and (d) 1 December 2016 in images obtained from the European Union Copernicus Sentinel 1 satellite, using the EO Browser. In 2016, breeding failed at Halley Bay due to the sea ice loss shown in panels b-d; many birds are thought to have moved to Dawson-Lambton in subsequent years. Note the Brunt Ice Shelf experienced a major calving event in January 2023, and it is uncertain yet if emperor penguins have begun to return (Press release British Antarctic Survey, January 23rd 2023).

As well as looking at long-term average outcomes, scientists have also considered the effect of extreme environmental events in their population models (Jenouvrier et al. 2021). For example, satellite imagery showed that at Halley Bay in 2016 there was late formation and early break-up of sea ice on which the colony was located (Figure 2), and more than 10,000 chicks probably died (Fretwell and Trathan 2019). Additional extreme events, such as the calving of glaciers, ice tongues, and ice shelves, have the potential to undermine the stability of fast ice, necessitating the relocation of penguin colonies (Fretwell et al. 2014). The calving of ice shelves, in particular, can lead to the creation of large tabular icebergs that may impede penguin access to foraging grounds. This disruption has the potential to impact breeding success by increasing the distances to feeding areas and causing damage to fast ice breeding platforms (Kooyman et al. 2017).

In addition to revealing the effects of extreme environmental events, satellite imagery has also helped unravel other effects driving the size of colonies. Satellite images show that emperor penguin movements between colonies can cause colonies to grow or shrink. With this knowledge, scientists developed models to incorporate movement behaviors to assess how populations might change as penguins move between colonies. These models assume that penguins will leave poor quality habitat to search over both short and long distances for higher-quality habitat that maximizes their chances of breeding successfully. However, even including dispersal activities and the formation of new colonies will not reverse the anticipated population decline.

By combining the effects of dispersal and extreme events in emperor penguin demographic models, Jenouvrier et al. (2021) projected that 98% of colonies will be extinct by 2100 under a high climate warming scenario (‘baseline’ RCP8.5 scenario, in which greenhouse gas emissions remain unchanged throughout the 21^st century), and that the global population will decline by 99% compared to its historical size (Figure 3). Under a scenario where the Paris Agreement is met, limiting the average global temperature increase to no more than 2°C, with additional efforts to restrict the increase to only 1.5°C, the projected decline in colony numbers persists but to a much lesser extent. Under this scenario, 61% of colonies are anticipated to be at risk of extinction by 2100, thus decelerating, but not preventing, the overall global population decline (Jenouvrier et al. 2019, Table S2 in Jenouvrier et al. 2021 , Figure 3).

Figure 3. Maps of emperor penguin colonies and sea ice changes that contrast the future for a world without or with climate mitigation and impact of climate change on emperor penguin global population (updated from Jenouvrier et al. 2021).
For maps, the sea ice loss is represented from severe loss by 2100 relative to today in dark blue to much less changes in sea ice in white. The color and size of dots shows population decline by target dates 2050, 2080 and 2100.
Right panel (without climate mitigation): Most of the colonies will have disappeared under a scenario where greenhouse gas emissions continue their current course (RCP 8.5, Riahi et al 2011). 98% of the colonies will go quasi-extinct by the end of the century (red dots). Importantly, the few colonies that will not have disappeared, shown in black here, will be declining.
Left panel (with climate mitigation): If actions are taken to reduce greenhouse gas emissions and the Paris Agreement objectives are met (Sanderson et al. 2018), viable emperor penguin refuges will continue to exist in Antarctica by 2100. The model still projects declines, but the trends will be much less severe so that fewer colonies will go to extinction.
For the global population trends: total number of breeding pairs of emperor penguins from 2009 to 2100 projected for various climate scenarios (red: greenhouse gas emissions continue their current trends; blue: countries take actions to reduce their greenhouse gas emissions in order to reach the goals of the Paris Agreement).

The Climate Action Tracker (CAT, https://climateactiontracker.org/) plays a crucial role in quantifying and evaluating climate change mitigation efforts, analyzing targets, policies, and actions at both national and global levels. According to the CAT’s assessments, current policies are insufficient to adequately limit global warming. The combined estimate now suggests average warming of 2.7°C (plausible range between 2.5°C and 2.9°C) by 2100, with temperatures continuing to rise beyond that point. Such global temperature increases pose significant threats to emperor penguins. Under such scenarios the global population will have more than halved by 2050 (Jenouvrier et al. 2021).

Human activities may potentially transfer native Antarctic species to areas within Antarctica where they are not found naturally (intra-regional transfer) (Hughes et al., 2019; Bergstrom, 2022). Human-mediated dispersal of species could disrupt established terrestrial and freshwater ecosystems and alter the distinct biogeographic regions found within Antarctica (Hughes and Convey, 2010; Terauds and Lee, 2016; Hughes et al., 2019; Cukier et al., 2023). Furthermore, human movement within Antarctica may transfer existing non-native species to other Antarctic areas (Hughes et al., 2019). For example, laboratory research has shown that a flightless midge (Eretmoptera murphyi), accidentally introduced to Signy Island, South Orkney Islands, could survive and complete its life cycle c. 750 km further south on the Antarctic Peninsula and some grasses could survive beyond their current Antarctic distributions (Hughes et al., 2013; Pertierra et al., 2017a).