The emperor penguin - vulnerable to projected rates of warming and sea ice loss Philip N. Trathan1, *, Barbara Wienecke2, Christophe Barbraud3, Stéphanie Jenouvrier3, 4, Gerald Kooyman5, Céline Le Bohec6, 7, David G. Ainley8, André Ancel9, 10, Daniel P. Zitterbart11, 12, Steven L. Chown13, Michelle LaRue14, Robin Cristofari15, Jane Younger16, Gemma Clucas17, Charles-André Bost3, Jennifer A. Brown18, Harriet J. Gillett1, Peter T. Fretwell1 1 British Antarctic Survey, Natural Environment Research Council, High Cross, Madingley Road, Cambridge CB30ET UK 2 Australian Antarctic Division, 203 Channel Highway, Tasmania 7050, Australia 3 Centre d'Etudes Biologiques de Chizé, UMR 7372, Centre National de la Recherche Scientifique, 79360 Villiers en Bois, France 4 Biology Department, MS-50, Woods Hole Oceanographic Institution, Woods Hole, MA, USA 5 Scholander Hall, Scripps Institution of Oceanography, 9500 Gilman Drive 0204, La Jolla, CA 92093-0204, USA 6 Département d’Écologie, Physiologie, et Éthologie, Institut Pluridisciplinaire Hubert Curien (IPHC), Centre National de la Recherche Scientifique, Unite Mixte de Recherche 7178, 23 rue Becquerel, 67087 Strasbourg Cedex 02, France 7 Scientific Centre of Monaco, Polar Biology Department, 8 Quai Antoine 1er, 98000 Monaco 8 H.T. Harvey and Associates Ecological Consultants, Los Gatos, California CA 95032 USA 9 Université de Strasbourg, IPHC, 23 rue Becquerel 67087 Strasbourg, France 10 CNRS, UMR7178, 67087 Strasbourg, France 11 Applied Ocean Physics and Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA, USA 12 Department of Physics, University of Erlangen-Nuremberg, Henkestrasse 91, 91052 Erlangen, Germany 13 School of Biological Sciences, Monash University, VIC 3800 Australia 14 University of Minnesota, Minneapolis and St. Paul, Minnesota MN 55455 USA 15 University of Turku, Turku, FI-20014 Turun Yliopisto, Finland 16 Milner Centre for Evolution, University of Bath, Claverton Down, Bath, BA2 7AY UK 17 Atkinson Center for a Sustainable Future and Cornell Lab of Ornithology, Cornell University, Ithaca, NY 14850, USA 18 Information Services, University of Cambridge, 7 JJ Thomson Avenue, Cambridge CB3 0RB UK * Author for correspondence (E-mail: [email protected]; Tel.: +44 1223 221602). 1 SUPPLEMENTARY MATERIAL, PART 1 Given the regional variability of climate change around the Antarctic (e.g. Turner et al., 2017; Rignot et al., 2019), different emperor penguin colonies are likely to be affected in different ways and over different temporal scales. Understanding the vulnerability of different habitat types will therefore help increase confidence when making projections about future scenarios. Individual colony projections are best viewed in the context of the local environment and geography (e.g. Kooyman et al., 2007, Barber-Meyer et al., 2008, Kooyman & Ponganis, 2016). To explore the current variation amongst colonies, we classified them according to current observed sea ice, near-surface temperature and wind (both zonal and meridional). We used sea ice concentration data obtained from satellite microwave sensors processed using the Bootstrap v3.1 algorithm (see https://nsidc.org/data/nsidc-0079); for near-surface temperature and winds, we used data from the ERA-Interim reanalysis (see https://www.ecmwf.int/en/forecasts/datasets/archive- datasets/reanalysis-datasets/era-interim). However, it should be noted that this explores variation in large-scale pack ice, and not specifically fast ice. The assumption is therefore made that “so as pack ice goes, so goes fast ice,” with a level of appreciable uncertainty. To some degree extensive pack ice cover does shield fast ice from ocean swells and wind waves, thus inhibiting early break up (Ainley et al., 2015; Kim et al., 2018). On the other hand, however, persistent winds can also retard fast ice formation, or lead to successive early break outs, at the same time contributing to more extensive pack ice (Ainley et al., 2010). Clearly, the situation is complex and data are sparse, stemming from the fact that microwave imagery cannot differentiate between ice-covered land and adjacent fast ice. For both data sets, we extracted seasonal means for January through March, April through June, July through September and October through December for the years 1979-2017. Data were gridded at 99 km spatial resolution and the values for the cells nearest to each colony extracted; we then used K- means clustering based on a scree plot of the within groups sum of squares to estimate the appropriate number of clusters. Clustering was carried out in R (R version 3.3.1 Copyright © 2016 The 2 R Foundation for Statistical Computing) using RStudio (Version 1.0.136 – © 2009-2016 RStudio, Inc.) and the library cluster. Four clusters were evident based on the first two principal components (PC1 and PC2); these jointly explained 62.8% of the variance. PC1 showed positive weightings for temperature and negative weightings for both meridional and zonal winds. PC2 showed positive weightings for zonal winds and negative weightings for meridional winds. PC1 also separated sea ice with positive weightings for January through March, and negative weightings for April through June and July through September. Based on these principal components, the area of recent rapid warming in West Antarctic formed a separate group (SM Fig. 1, group 1), whilst the Ross Sea and the Weddell Sea, i.e. the deepest embayments of the Antarctic continent, formed another group (SM Fig. 1, group 3); the coast line of East Antarctica comprised two groups (SM Fig. 1, groups 2 and 4). A small number of colonies from each group were located away from the majority of sites for that group. Identifying such different broad habitat types around the continent emphasises the range of habitats currently occupied by the species and the complexity of projecting change. EMPIRICAL OBSERVATIONS AT SELECTED COLONIES To better understand the factors that affect emperor penguin colonies we consider several as exemplars of differing environmental conditions, and differing levels of knowledge about individual sites. Many of the conditions and responses observed at each site almost certainly have wider relevance to colonies elsewhere, including those distant to national research stations. The colonies we highlight are (i) Halley Bay (SM Fig. 1, group 3), a colony next to an ice shelf demonstrating collapse following stochastic events during years with extreme low atmospheric pressure and strong winds; (ii) Pointe Géologie (SM Fig. 1, group 4), a colony within an archipelago and that is now stable or recovering after extreme perturbation; (iii) Rothschild Island (SM Fig. 1, group 1), a colony in an area that is warming rapidly and vulnerable to ice shelf basal thinning and collapse; (iv) Taylor Glacier (SM Fig. 1, group 4), a colony at a fixed location situated on land (v) Ledda Bay (SM Fig. 1, group 1), a colony where intermittent sea ice allows only occasional successful breeding; (vi) Riiser-Larsen 3 Peninsula/Gunnerus Bank and Umebosi Rock (SM Fig. 1, group 2), colonies where numbers were high but then suddenly decreased; and, (vii) Cape Washington and Cape Crozier (SM Fig. 1, group 3), colonies with adjacent colonies in the southwest Ross Sea that show inter-annual variability, but which are relatively stable. (i) Halley Bay (75.55°S, 27.42°W) The emperor penguin breeding site at Halley Bay, Coats Land, Weddell Sea, used to be the second largest of all known colonies (Fretwell et al., 2012). Records are sparse, but various sites in the creeks of the Brunt Ice Shelf have been used regularly by large numbers of breeding pairs (see historical counts in SM Table 1) that congregate at a location generally consistent between years. Since the operational use of remote sensing for population estimation (Barber-Meyer et al., 2007; Fretwell & Trathan, 2009; Fretwell et al., 2012; Fretwell & Trathan, 2019), the site has been surveyed annually to estimate the colony size (SM Fig. 2). This has revealed fluctuations in numbers, and complete breeding failure between 2016 and 2018 due to the early break out of sea ice (Fretwell & Trathan, 2019). Variability in population size is a feature of the colony, with several significant decreases; however, complete breeding failure has not been previously recorded. The first year of complete failure immediately followed the strongest El Niño in more than 60 years (Oceanic Niño Index, ONI; data available from https://climatedataguide.ucar.edu/climate-data/nino-sst-indices-nino-12-3-34-4-oni- and-tni), and one of the highest positive values of the Southern Annular Mode (SAM; data available from http://www.nerc-bas.ac.uk/icd/gjma/sam.html) (Fretwell & Trathan, 2019). In 2017, sea ice had blown out by October, and after sea ice reformed in November, only 2,000 adults were recorded but none were breeding. In October 2018, just a few hundred adults were present, but again, by the end of November the sea ice had blown out, suggesting another year of failed breeding. Fretwell & Trathan (2019) suggested that following 2016, a proportion of the population probably had deferred breeding and had remained at sea in an unobservable state, a proportion was not breeding but had then become observable, while a large proportion, increasing over the three year period, had begun to 4 breed at the nearby Dawson-Lambton colony. This colony is only 55 km from the Halley Bay site, which is unusually close for emperor colony locations (Ancel et al., 2017). It is possible that some birds could also have moved to other colony locations farther away, but due to the natural variability of colony size, smaller changes are more difficult to associate with immigration. Wind speeds at Halley Bay tend to be highest during August, September and October (Turner et al., 2009c), sea ice dispersal at this time would be critical for chick survival, which may explain the colony location in a sheltered creek with firm fast ice. In 2016, average monthly atmospheric pressure was the lowest for September in over 30 years, whilst average wind speed was the highest for September over the same time frame; temperatures were also higher than average (Fretwell & Trathan, 2019).