J Paleolimnol (2011) 45:273–285 DOI 10.1007/s10933-011-9497-x

ORIGINAL PAPER

Late Holocene Ade´lie penguin population dynamics at Zolotov Island, Vestfold Hills,

Tao Huang • Liguang Sun • Yuhong Wang • Deming Kong

Received: 24 December 2009 / Accepted: 9 October 2010 / Published online: 14 January 2011 Ó Springer Science+Business Media B.V. 2011

Abstract We inferred late Holocene Ade´lie pen- ago, which we interpret as a response to the Little Ice guin occupation history and population dynamics on Age, or a neoglacial cooling event. Zolotov Island, Vestfold Hills, Antarctica, using geochemical data from a dated ornithogenic sediment Keywords Ade´lie penguin Antarctic climates Ice core (ZOL4). Radiocarbon dates on fossil penguin core Ornithogenic sediments Western Antarctic bones in the core indicate that Ade´lie penguins Peninsula Little Ice Age occupied the island as early as 1,800 years before present (yr BP), following the retreat of the SØrsdal glacier. This occupation began *1,200 years later than that observed at and King George Introduction Island, in the . Phosphorus was identified as the most indicative bio-element for Polar seabirds provide important linkages between penguin guano in core ZOL4, and was used to infer marine ecosystems and terrestrial environments. past penguin population dynamics. Around They transport marine-derived nutrients and contam- 1,800 years ago, the Ade´lie penguin populations at inants onto land via their guano and physical remains both Zolotov Island and Ardley Island increased (Sun and Xie 2001a; Blais et al. 2005, 2007; Xie and rapidly and reached their highest levels *1,000 yr Sun 2008; Yin et al. 2008; Brimble et al. 2009; BP. For the past *900 years, the penguin popula- Keatley et al. 2009; Michelutti et al. 2010). Lake tions at Zolotov Island have shown a general rising sediments in areas visited by seabirds can therefore trend, with fluctuations, while those at Ardley Island contain materials of both marine and lacustrine origin have shown a moderate decreasing trend. The Ade´lie (Sun and Xie 2001b). In Antarctica, chemical signa- penguin populations at both Ardley Island and tures from penguin droppings and physical signatures Zolotov Island showed a clear decline *300 years such as bones, feathers and hairs in lake sediments have been used to infer the past population dynamics of penguins and seals, as well as their responses to T. Huang L. Sun (&) Y. Wang D. Kong changing climate and human activities (Hodgson and Institute of Polar Environment, University of Science Johnston 1997; Sun et al. 2000, 2004a, b, 2005; Wang and Technology of China, 230026 Hefei, China et al. 2007; Huang et al. 2009a; Yang et al. 2010). For e-mail: [email protected] example, the penguin populations at Ardley Island Y. Wang and King George Island, South Shetland Islands, National Institutes of Health, Bethesda, MD 20892, USA showed a dramatic decline in the neoglacial period, 123 274 J Paleolimnol (2011) 45:273–285

2,300–1,800 yr BP (Sun et al. 2000), indicating the Many Ade´lie penguin colonies are present on its negative impacts of cooling climate on penguin western coastal islands. During the 1981/1982 sea- populations. sons, it was estimated that there were 196,592 pairs Ade´lie penguins (Pygoscelis adeliae) are the most of breeding Ade´lie penguins on these islands, of abundant seabirds in Antarctica and the bellwether of which 17,496 were on Zolotov Island (Whitehead and Antarctic climate change (Ainley 2002). Their pop- Johnstone 1990). In a previous study, we recon- ulation dynamics are influenced by climatic and structed an 8,500-year record of Ade´lie penguin environmental factors such as sea ice extent and population dynamics at , Vestfold duration, sea surface temperature, air temperature and Hills and examined associations with climate and snow cover (Fraser et al. 1992; Wilson et al. 2001; environmental changes (Huang et al. 2009a). Tem- Jenouvrier et al. 2006; Bricher et al. 2008), and thus poral resolution of the inferred penguin population they provide an integrated response to ecological and shifts at Gardner Island, however, was relatively low, climate changes (Croxall et al. 2002). In the past few and detailed penguin population changes over the decades, observational records of changing Ade´lie past 2,000 years were not resolved. In the present penguin populations have shown strikingly different study, we explore geochemical and chronological trends in the Antarctic Peninsula and East Antarctica data from sediment core ZOL4 taken in a lake on regions. The Ade´lie penguin populations in East Zolotov Island. We extracted the bio-elements and Antarctica have shown a sustained increase, while inferred late Holocene Ade´lie penguin occupation those in the Antarctic Peninsula region have and population dynamics. We also compared the decreased (Woehler et al. 2001). These opposite penguin population changes at Zolotov Island with population trends are likely associated with differ- records from Ardley Island in the South Shetland ences in regional climate and environment, such as Islands over the past 1,800 years, and examined sea ice extent and related changes in prey abundance associations with regional climate changes. (Fraser and Hofmann 2003; Forcada et al. 2006). Records of penguin population changes over larger spatial and temporal scales are required to provide a Materials and methods long-term record of natural variability and to under- stand the population responses of Adelie penguins to Study site and sample collection changes in climate and marine ecosystems. The Vestfold Hills is one of the larger East Zolotov Island is located about 10 km southwest of Antarctic oases (Fig. 1), located east of Prydz Bay. the Australian Antarctic in Vestfold

South Shetla nd Isla nds N 2 km

Vestfold Hills

62º10'S Antarctica

Fildes Peninsula 68º30'S

Ross Sea East Antarctic ice sheet

Ardley Island Davis Station

62º12'S DG4 Y2

Grea t Wa ll Sta tion N ZOL4 Sorsdal glacier 5 km 58º59'W 58º56'W 78ºE

Fig. 1 Map of the Vestfold Hills and South Shetland Islands, including the sampling sites on Zolotov Island (ZOL4) and Ardley Island (Y2) 123 J Paleolimnol (2011) 45:273–285 275

Hills, East Antarctica (68°390S, 77°520E; Fig. 1). The USA). Standard sediment reference materials were island is about 2 km long and 1.5 km wide, and has a included with every batch of samples. The analytical maximum altitude of 28 m above sea level. It values for major elements and trace elements are possesses a large number of breeding Ade´lie pen- within ±0.5% and ±5% of the certified standards, guins. Sediment core ZOL4 was retrieved from a lake respectively. TC, TN and S were measured by vario near a large Ade´lie penguin colony. The catchment is EL III (Elementar, Germany) with a relative error of located in a low-lying basin at the center of this 0.1%. island, and is about 120 m long and 55 m wide. It lies We ran R-mode clustering analysis, Principal at an altitute of *7 m. During field investigations, Component Analysis (PCA) and Pearson correlation the lake was very shallow. A 12-cm-diameter PVC analysis on these data using SPSS16.0, to examine pipe was pushed vertically into the deepest part of the the relationship between element concentrations in lake to collect the sediment core. After the PVC pipe the sediments and their controlling factors. The was retrieved, its bottom and top were sealed. In the concentration data of fresh penguin guano and laboratory, the 40-cm core was opened and sectioned bedrock at nearby Gardner Island were also included at 1-cm intervals. Penguin remains such as bones, in this study to help discern the bedrock signature feathers and eggshell fragments were found in the from that of penguin guano. upper 17 cm, and hand picked. The 40 subsamples were stored frozen prior to analysis. Before chemical analyses, each subsample was air-dried in a clean Results laboratory and homogenized. Sedimentology and chronology Radiocarbon dating and geochemical analyses Sediment core ZOL4 is 40 cm long. Two sedimen- We dated four fossil penguin bones (collagen) and tary units were identified from macroscopic descrip- two bulk sediment samples (organic carbon) by tion, color, smell and sedimentology (Fig. 2). Unit 1 Accelerator Mass Spectrometry (AMS) 14C to estab- extends from the base to 17 cm, and is characterized lish the depth/age profile for core ZOL4. Dates on by dominance of greyish deposits consisting of fossil bone were corrected for the marine carbon mainly sand, a moderate amount of silt, and some reservoir effect using the dataset of Marine04 (Hug- small gravels. Unit 2 spans from 17 cm to the hen et al. 2004) to give a DR 880 ± 15 years, the age surface and consists of olive to dark olive grey of local modern penguin bone (Huang et al. 2009b), sediment, which compared with Unit 1, has more and calibrated using the CALIB 5.1.0 program silt and clay. Unit 2 also contains many physical (Stuiver et al. 2005). In this study, the calibrated penguin remains such as bones, feathers and 14C dates were reported in calendar years before eggshells, and has a strong smell of penguin guano, present (cal yr BP). and is identified here as a penguin ornithogenic All air-dried subsamples were measured to sediment layer, i.e. sediments that contain penguin determine the concentrations of 16 major and trace guano and body remains. Unit 1 has a distinct elements (P (P2O5), S, Cu, Zn, Ni, Cd, Pb, K (K2O), sedimentology compared with Unit 2, and it is Na (Na2O), Ca (CaO), Mg (MgO), Fe (Fe2O3), Al unlikely amended by penguin guano. TC and TN are (Al2O3), Mn, Cr and Ti), total carbon (TC) and total very low in Unit 1, but very high in Unit 2 (Fig. 2; nitrogen (TN). For chemical element analyses, Table 1), indicating a substantial increase in organic subsamples were sieved though a 70-lm mesh, inputs to Unit 2. and then ground to powder after removal of large Radiocarbon results are shown in Table 2. The rock fragments. About 0.25 g of each powder mean calibrated ages of penguin bones at 1, 8, 14, and sample was taken, weighed, and digested (HNO3- 17 cm are about 80, 930, 1,280, and 1,765 cal yr BP, HF-HClO4) in a Teflon crucible with electric respectively, and they show a fairly linear trend with heating. The digested samples were analyzed for depth (r = 0.99, n = 4). Ages of the two bulk P, Cu, Zn, Ni, Cd, Pb, K, Na, Ca, Mg, Fe, Mn, Cr, sediment samples at 30 and 39 cm depth are Al and Ti using an ICP-OES DV2100 (PerkinElmer, 14,110 ± 50 yr BP and 17,080 ± 60 yr BP 123 276 J Paleolimnol (2011) 45:273–285

P2O5 (%) MgO (%) TC (%) TN (%) Cu (ug/g) Zn (ug/g) Stratigraphy 0246468101 2 3 4 0 1 2 3 0 200 400 50 100 150 0 0 cm 5 unit 2 10 15 17 20 25 unit 1

Depth (cm) 30 35 40 40 45

S (%) Ni (ug/g) Cd (ug/g) Pb (ug/g) Mn (mg/g) K2O (%) Stratigraphy 0.2 0.3 0.4 0.5 70 140 210 0123468100.6 1.2 1.8 1.2 1.8 2.4 0 0 cm 5 10 unit 2 15 17 20 25 unit 1

Depth (cm) 30 35 40 40 45

Al2O3 (%) CaO (%) Fe2O3 (%) Ti (mg/g) Cr (ug/g) Na2O (%) Stratigraphy 45678345674681012135770 140 210 2.5 3.0 3.5 4.0 0 0 cm 5 unit 2 10 15 17 20 25 unit 1

Depth (cm) 30 35 40 40 45 Legend Feather Bone Eggshell Gravel

Fig. 2 The sedimentology and elemental concentration profiles in sediment core ZOL4

(14C dates), respectively, very different from the ages Element concentrations determined on bones. The ‘‘old carbon’’ reservoir effect on bulk sediments is difficult to estimate Mean concentrations of elements in the sediments of precisely because there are multiple sources of ZOL4 are listed in Table 1. Among the major carbon to the sediment. Nevertheless, these dates elements, Fe has the highest mean concentration of suggest that Unit 1 includes material of late glacial 8.33%, followed by Mg (6.50%), Al (6.23%), Ca age, and predates Unit 2. The chronology of the upper (5.61%), Na (3.18%), K (1.67%), Ti (4.38%) and Mn unit (17–0 cm) was established using linear interpo- (1.15%). The mean concentration of P is 1.96% and lation between the four calibrated AMS 14C dates on as high as 4.17% in Unit 2. The elemental concen- macrofossils (Fig. 3). tration profiles of ZOL4 are plotted in Fig. 2. They

123 J Paleolimnol (2011) 45:273–285 277

Table 1 Element mean concentrations in the sediments of ZOL4, together with reference data from penguin guano and bedrock Elements Range CV (%) ZOL4 (0–40 cm) ZOL4 (0–17 cm) ZOL4 (17–40 cm) *Guano (n = 3) *Bedrock (n = 3)

TC (%) 1.16–3.74 37.2 2.16 3.05 1.50 10.25 0.23 TN (%) 0.14–2.05 97.0 0.61 1.17 0.19 2.84 0.03 P (%) 0.19–5.72 125.4 1.96 4.17 0.33 6.31 0.33 S (%) 0.26–0.41 76.0 0.31 0.35 0.29 2.11 0.46 Mg (%) 4.82–9.63 25.6 6.50 8.08 5.34 5.57 4.62 Cu (lg/g) 66–315 52.5 164 252 99 310 144 Zn (lg/g) 65–126 21.0 90 106 78 130 209 Ni (lg/g) 88–198 24.3 127 155 106 22 52 Cd (lg/g) 0.69–2.29 32.5 1.30 1.66 1.03 0.95 0.40 Fe (%) 4.91–11.66 23.4 8.33 5.83 10.19 3.52 17.7 Ti (mg/g) 2.02–6.44 32.8 4.38 2.54 5.75 2.42 11.59 Al (%) 4.24–7.47 15.5 6.23 5.01 7.12 7.04 7.51 Ca (%) 3.79–6.83 14.4 5.61 4.79 6.21 4.21 9.64 Pb (lg/g) 4.91–9.48 15.3 7.43 7.82 7.14 5.83 9.52 K (%) 1.34–2.23 11.1 1.67 1.62 1.70 0.78 0.21 Na (%) 2.59–3.72 8.3 3.18 3.02 3.31 3.11 1.42 Mn (%) 0.85–1.55 13.1 1.15 1.13 1.16 0.03 0.19 Cr (lg/g) 109–217 13.6 163 146 175 – – CV is coefficient of variation; * Collected from Gardner Island, which is adjacent to Zolotov Island in Vestfold Hills; – not determined

Table 2 AMS 14C dates and calibrated ages using CALIB 5.1.0 and the Marine04 datasets (DR 880 ± 15) UCIAMS number Sample number Sample material Depth (cm) Conventional 14C age (yr BP) Calibrated age (cal yr BP) Mean Range (2r)

*39438 DG-34 Bone Modern 880 ± 15 Modern Modern 55716 ZOL4–1 Bone 1 1,340 ± 15 83 0–139 55717 ZOL4–8 Bone 8 2,250 ± 15 928 904–963 55736 ZOL4–14 Bone 14 2,595 ± 15 1,281 1,252–1,309 55718 ZOL4–17 Bone 17 3,030 ± 15 1,765 1,702–1,817 55794 ZOL4–30 Bulk sediments 30 14,110 ± 50 – – 55795 ZOL4–39 Bulk sediments 39 17,080 ± 60 – – The dates were measured at the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory, University of California Irvine (KCCAMS UCI); * 39438 was presented in Huang et al. (2009b) show substantial fluctuations, consistent with the show an opposite trend. Their concentrations are high sedimentology. Similar to TC and TN, the concen- in sediments from 40–17 cm and low in sediments trations of most elements show an abrupt shift at from 17–0 cm. The large changes in elemental 17 cm depth. The elements P, Mg, Cu, Zn, Ni, S, Cd concentrations at 17 cm indicate a change in the and Pb have similar concentration profiles in that source of material delivered to the sediment. The their concentrations are high in the upper, 17–0 cm elements K, Na, Cr and Mn do not show clear trends, (Unit 2), but very low and stable in the lower and thus appear to have been less impacted by the 40–17 cm (Unit 1). The elements Al, Ca, Fe and Ti changes in the sedimentary source materials.

123 278 J Paleolimnol (2011) 45:273–285

P, Mg, TN, TC, Cu, Zn, S, Ni and Cd are significantly correlated with each other and they are negatively correlated with lithologic elements Al, Ca, Fe, Ti, Cr and Na. PCA is an effective statistical method for separat- ing elemental assemblages and elucidating the main controlling factors on chemical components in lake sediments (Hodgson et al. 2001; Liu et al. 2006). We therefore performed PCA on the elemental concen- Fig. 3 The age profile of ZOL4 based on linear interpolation trations in sediments of core ZOL4. Two factors 14 between the calibrated C dates (Note that the dates below accounting for approximately 80.97% of the variance 20 cm are not plotted on the same scale) in the data were obtained, and their loadings are plotted in Fig. 5. Component 1 accounts for 64.13% of the total variance, and is the controlling factor. The variables P, Mg, TN, TC, Cu, Zn, S, Ni and Cd have very high positive loadings on Component 1. These nine elements are significantly inter-correlated and show similar concentration profiles (Table 3; Fig. 2). Their mean concentrations in the penguin ornitho- genic layer (Unit 2) are higher than those in Unit 1 (Table 1), and therefore Component 1 is likely linked with the input of penguin guano or guano-derived materials. Al, Ca, Fe and Ti have negative loadings on Component 1, likely caused by the input of a large amount of penguin guano and the consequent dilution and reduction in their relative concentrations. Com- ponent 2, which accounts for 16.84% of the total variance, is characterized by high positive loadings of K, Cr, Pb, Na and Mn and the negative loadings of P, Fig. 4 R-mode clustering results for the elements measured in sediment core ZOL4 Mg, TN, TC and Cd. S, Ni, Cu, Zn, Al, Ca, Fe and Ti have positive loadings on Component 2, therefore Component 2 probably represents the geochemical Statistical analysis contributions from other sources such as the local bedrock. Assemblages of bio-elements in lake sediments have To confirm the results of PCA on separating the been used successfully to infer their material sources element concentrations in ZOL4, we included the (Sun et al. 2000, 2004a; Liu et al. 2006; Huang et al. concentrations of elements in fresh penguin guano 2009a). We performed R-clustering, Pearson corre- and bedrock from nearby Gardner Island (Table 1). lation analyses and PCA to obtain an assemblage of The concentrations of P, TN, TC, S, Cu, Cd and Mg bio-elements for the penguin ornithogenic sediments in penguin guano are much higher than those in of ZOL4. bedrock, suggesting that these elements are derived The R-clustering results for the elemental concen- mainly from penguin guano on Gardner Island and trations in ZOL4 are shown in Fig. 4 and show that P, Zolotov Island. In contrast, Zn, Ni, Pb, Mn, Fe, Ti, Ca TN and S belong to the first group, TC, Cu, Ni, Zn, and Al show geochemical characteristics different Mg and Cd the second, and Al, Ca, Na, K, Cr, Mn, from penguin guano, indicating that guano is not their Pb, Fe and Ti the third. Pearson correlation analyses primary source. were performed to confirm the clustering results. The As indicated above, P, Mg, Cu, S and Cd in the coefficients are listed in Table 3 and are consistent ornithogenic sediment layer of ZOL4 are derived with the R-clustering results. The concentrations mainly from penguin guano and are thus used here as 123 J Paleolimnol (2011) 45:273–285 279

Table 3 Correlation coefficients between the elements in the ZOL4 sediment core P Mg TC TN Cu Zn S Ni Cd Al Ca Fe Ti Cr Na Pb Mn K

P1 Mg .98* 1 TC .89* .84* 1 TN .95* .93* .89* 1 Cu .89* .86* .91* .80* 1 Zn .80* .76* .77* .65* .90* 1 S .72* .70* .70* .68* .79* .71* 1 Ni .72* .74* .79* .60* .93* .88* .71* 1 Cd .69* .62* .67* .57* .75* .88* .58* .69* 1 Al -.95* -.91* -.91* -.91* -.90* -.79* -.60* -.76* -.68* 1 Ca -.88* -.85* -.76* -.90* -.72* -.56* -.54* -.53* -.44* .89* 1 Fe -.95* -.89* -.91* -.92* -.87* -.81* -.73* -.69* -.76* .93* .85* 1 Ti -.94* -.89* -.90* -.90* -.92* -.82* -.81* -.78* -.72* .92* .87* .96* 1 Cr -.65* -.57* -.63* -.68* -.47* -.44* -.19 -.27 -.53* .69* .67* .72* .61* 1 Na -.56* -.49* -.44* -.51* -.48* -.54* -.07 -.33* -.57* .70* .63* .61* .54* .71* 1 Pb .29 .28 .27 .22 .43* .37* .72* .46* .20 -.21 -.23 -.29 -.45* .26 .10 1 Mn -.09 -.02 -.08 -.23 -.01 .09 .12 .26 .01 .24 .34* .19 .18 .33* .46* .09 1 K -.22 -.13 -.23 -.29 .01 -.07 .32* .18 -.21 .30 .28 .31* .10 .70* .52* .64* .35* 1 * Significant at the level of 0.01 (2-tailed)

unit, 17–0 cm, are affected by penguin guano input, but those in the lower unit, 40–17 cm, are not. Thus elemental concentrations between 40 and 17 cm mainly reflect the background values of local bedrock. Here we use elemental enrichment ratios to evaluate the impacts of penguin guano on the elemental concentrations in the ZOL4 sediments. We define the

enrichment ratios as ERi = Cmean 17–0/Cmean 40–17, where Cmean 17–0 is the mean elemental concentration in the upper 17 cm and Cmean 40–17 is the mean concentration in the lower unit, 40–17 cm. The elemental enrichment ratios are plotted in Fig. 6. Phosphorus has the maximum enrichment ratio, 12.64,

Fig. 5 Distributions of the factor loadings from the PCA analysis of the elements and oxides in sediment core ZOL4 geochemical markers of penguin populations. These elements, however, also exist in the weathered bed- rock, especially Mg, Cd and Cu (Table 1). Therefore it was necessary to identify those elements that have very low concentrations in local weathered bedrock, but high concentrations in penguin guano. We define these as optimal bio-elements. The sediments in the upper Fig. 6 Enrichment ratios of elements in sediment core ZOL4 123 280 J Paleolimnol (2011) 45:273–285 and was identified as the optimal bio-element to component of fresh penguin guano and ornithogenic represent Ade´lie penguin guano in this study. soils (Tatur and Keck 1990). Elements P, Mg, Cu, S and Cd in ZOL4 are therefore mainly derived from penguin guano, and Discussion these elements are considered immobile in Antarctic lake sediments (Sun et al. 2000; Liu et al. 2005). Bio-elements as geochemical markers of penguin Consequently, concentrations of these elements in guano ZOL4 are considered to be good geochemical mark- ers for changes in penguin guano inputs and thus Relative to typical Antarctic soils and bedrock, many penguin numbers. Of these elements, P is most elements are enriched in penguin guano, ornithogenic enriched and can therefore be used to infer past soils and ornithogenic sediments. Sun et al. (2000) Ade´lie penguin population dynamics around the lake reported that nine elements including F, P, S, Se, Cu, catchment. Zn, Ca, Sr and Ba were enriched significantly in penguin ornithogenic sediments in core Y2 relative to Late Holocene Ade´lie penguin occupation the surrounding Antarctic soils on Ardley Island and and population dynamics at Zolotov Island King George Island, South Shetland Islands. Hofstee et al. (2006) reported that P, S, Mg, Ca, As, Cu, Zn, Ade´lie penguins are sensitive to Antarctic climate and and Cd were concentrated in penguin guano from environmental changes. The establishment of their in northern Victoria Land. Similar breeding colonies is well correlated with the degla- enrichment of certain elements in fresh penguin ciation and formation of ice-free areas in Antarctica guano has also been reported at , (Baroni and Orombelli 1994; Emslie and McDaniel Terra Nova Bay and Admiralty Bay, King George 2002; Emslie et al. 2003; Huang et al. 2009b). In the Island, South Shetland Islands (Ancora et al. 2002; Vestfold Hills, there was a large increase in sea ice Zdanowski et al. 2005). Previously, we reported very extent between 2,500 and 2,000 yr BP (corrected 14C high concentrations of F, P, S, Se, Cu, Sr and As in dates). This event, termed the Chelnock Glaciation by Adelie penguin guano compared to bedrock at Adamson and Pickard (1986), has been identified Gardner Island, Vestfold Hills, East Antarctica from terrestrial evidence (Adamson and Pickard (Huang et al. 2009a). In the present study, Se, Sr 1986), lake sediments (Roberts and McMinn 1999), and As in ZOL4 were not determined. Elements P, and marine sediments (McMinn et al. 2001; Taylor Mg, Cu, S and Cd were present in high concentrations and McMinn 2002). After 2,000 yr BP, the sea-ice in the ornithogenic layer of ZOL4, due to their high extent was reduced, but it was still substantially levels in guano. However, Zn and Ni were present in greater than that prior to 2,500 yr BP (McMinn et al. higher concentrations in local bedrock (Table 1). 2001). The sediments of Abel and Platcha Bays in the Elements Zn and Ni are easily enriched by organic southern Vestfold Hills recorded a possible ice cap matter in the earth’s surface (Goldschmidt 1954), and retreat at approximately 1,750 yr BP (corrected 14C their high levels in the ornithogenic layer of ZOL4 may dates) (McMinn 2000), coinciding with a warm period be due to a combination of guano and local bedrock (2,000–1,750 yr BP) recorded in the sediments from inputs. Similar high levels of Zn and Ni in ornithogenic nearby Lake Nicholson (Bronge 1992). In the present sediment core DG4 from nearby Gardner Island were study, the AMS 14C dates on the fossil penguin bones also observed (Huang et al. 2009a). High levels of P, S from sediment core ZOL4 indicate that Ade´lie and Cu in guano mainly originate from Antarctic krill penguins occupied Zolotov Island only in the late (Euphausia superba), a main dietary component of the Holocene, since *1,765 cal yr BP, corresponding to Ade´lie penguins (Tatur and Keck 1990), and high this period of local warm climate and the retreat of levels of Cd in guano are likely from upwelling nearby SØrsdal glacier, as recorded in lake sediments Antarctic deep water (Ancora et al. 2002). Mg and P in (Bronge 1992; McMinn 2000). the ornithogenic layer of ZOL4 show significant Bulk sedimentation rates in the upper 17 cm of correlations (r = 0.98, p \ 0.001), indicating likely ZOL4 show only minor fluctuations, therefore the P deposition of struvite (Mg(NH4)PO46H2O), a typical concentration in the upper 17 cm is considered to be a 123 J Paleolimnol (2011) 45:273–285 281

Fig. 7 Ade´lie penguin a Zolotov Island population dynamics and 6 wider regional climate records. a Penguin population changes at 5 Zolotov Island as indicated

by the P concentrations in (%) ZOL4; b Climate changes 5 4 O 2

over the past 2,000 years P from d18O in the Taylor Dome ice core (5-points 3 smoothed); the original data were reported by Steig et al. (2000), and in the Law 2 increase populations Penguin 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Dome ice core (figure from Morgan 1985); c Penguin population changes at b Law Dome and Taylor Dome -21.5 -39 Ardley Island as indicated by the P concentrations in sediment core Y2 (Sun et al. 2000) -22.0 -40

-22.5 O (‰) O (‰) 18 18

-41

-23.0

Law Dome Ta ylor Dom e

-23.5 -42 0 200 400 600 800 1000 1200 1400 1600 1800 2000

12 c Ardley Island 10

8 (%)

5 6 O 2

P 4

2

0 Penguin populationsincrease 0 200 400 600 800 1000 1200 1400 1600 1800 2000 yr BP reliable indicator of the relative abundance of penguin BP and the last *200 years, and low levels during guano and therefore penguin numbers. Overall *930–810 cal yr BP and *320 cal yr BP, although P-inferred Ade´lie penguin population dynamics at some of the latter periods are indicated by single data Zolotov Island show a rising trend for the past points and are therefore less reliable. 1,800 years (Fig. 7). The P-inferred Ade´lie penguin The inferred Ade´lie penguin population dynamics populations at Zolotov Island show high levels during at Zolotov Island can be compared with the climate the periods *1,160–990 cal yr BP, *690–450 cal yr changes over the past 1,800 years inferred by Morgan

123 282 J Paleolimnol (2011) 45:273–285

(1985) from the oxygen isotope data in an ice core penguin to semi-permanent sea ice blocking shore- from Law Dome, East Antarctica. During the isotope- lines. Around 1,000 years ago, Ade´lie penguin pop- inferred warm period from 1,700 to 1,000 yr BP, the ulations at Zolotov Island and Ardley Island reached penguin population displayed a rising trend and high levels (Fig. 7a, c), corresponding to a warm peaked between *1,160 and 990 cal yr BP (Fig. 7a, b). period recorded in East Antarctic ice cores (Morgan The dramatic reduction of the penguin population 1985; Masson et al. 2000; Steig et al. 2000) and in between *930 and 810 cal yr BP coincided with marine sediments from the Antarctic Peninsula rapid cooling between 1,000 and 700 yr BP. During (Domack et al. 2003). Between 1,200 and 600 cal yr the relatively mild climate from 650 to 400 yr BP, the BP, Ade´lie penguins reoccupied Prior Island along penguin population maintained high levels between the Victoria Land coast (Baroni and Orombelli 1994), 690 and 450 cal yr BP. The oxygen isotope data in indicating favorable climate conditions, similar to the the ice core from Law Dome recorded a cooling from northern hemisphere Medieval Warm Period. The *400 yr BP to a cold period between 250 and 200 yr penguin populations at Zolotov Island decreased BP, followed by rapid warming, and this is consistent dramatically to a very low level *930 cal yr BP, with the decline of the penguin population at whilst those at Ardley Island only showed a moderate *320 cal yr BP and the increase in the last reduction (Fig. 7a, c). However, since *900 cal yr *200 years. Similar associations between penguin BP, the penguin populations at Zolotov Island show populations and climate were also observed on an overall rising trend with fluctuations, but those at adjacent Gardner Island. There, the inferred Ade´lie Ardley Island show a continued, moderately decreas- penguin populations reached high levels ing trend, except for a single dramatic decline *4,700–2,400 cal yr BP during the warmer mid- *300 cal yr BP that coincides with a clear decline Holocene, and then the populations showed a signif- in penguin populations between 450 and 200 cal yr icant decline, corresponding to the onset of local BP inferred from another ornithogenic sediment core neoglaciation (Huang et al. 2009a). (Y4) on the same island (Liu et al. 2005). These Such associations can be explained by the impacts opposing trends are similar to observational data for of climate-related changes on penguin nesting and the past few decades, during which Ade´lie penguin foraging. During warm climates, the island is exposed populations have shown an increasing trend in East adequately and thus provides ample nesting and Antarctica, while decreasing in King George Island breeding sites for penguins. Warm climate may also and the Palmer station area, Western Antarctic increase marine productivity, thereby providing suf- Peninsula (Woehler et al. 2001). These opposite ficient food supply to support the growth of penguin population trends over the past decades are likely populations. caused by differences in regional sea ice extent and prey availability, driven by local climate (Forcada Comparison of late Holocene penguin populations et al. 2006). Opposite penguin population trends in Antarctica inferred over the past *900 years at these two islands may be a consequence of internal biological Sun et al. (2000) reported a 3,000-year record of processes such as predation pressure. Physical pro- penguin population changes at Ardley Island, King cesses such as icebergs blocking access (Arrigo et al. George Island, South Shetland Islands, and the results 2002) may also play a role, in addition to external indicated a rapid increase in local penguin popula- climatic factors. tions from *1,800 cal yr BP (Fig. 7c). Ade´lie pen- The penguin populations at both Zolotov Island guins occupied Zolotov Island, Vestfold Hills from and Ardley Island experienced an abrupt decline *1,765 cal yr BP (Fig. 7a), and their populations around *300 cal yr BP (Fig. 7a, c), corresponding to experienced a similar and rapid increase. In the Ross a cold climate anomaly recorded in ice cores from Sea area, after the known ‘penguin optimum’ East Antarctica (Morgan 1985; Masson et al. 2000;Li between about 4,000 and 2,000 cal yr BP (Emslie et al. 2009) and West Antarctica (Mayewski et al. et al. 2003, 2007), there was a disappearance of 2004), and in marine sediments from the South Ade´lie penguins between 2,000 and 1,100 cal yr BP. Shetland Islands (Yoo et al. 2009; Milliken et al. Emslie et al. (2007) attribute the disappearance of the 2009) and Antarctic Peninsula area (Domack et al. 123 J Paleolimnol (2011) 45:273–285 283

1995; Leventer et al. 1996; Smith et al. 1999). These References recorded cold climates in Antarctica coincide approx- imately with the northern hemisphere Little Ice Age Adamson DA, Pickard J (1986) Cainozoic history of the (Lamb 1977; Grove 1988; Mann et al. 2009), but the Vestfold Hills. In: Pickard J (ed) Antarctic Oasis. Aca- demic Press, Sydney, pp 63–97 timing of the Little Ice Age in Antarctica is different Ainley DG (2002) The Ade´lie penguin: bellwether of climate (Masson et al. 2000; Bentley et al. 2009; Verleyen change. Columbia University Press, New York et al. in press). Ancora S, Volpi V, Olmastroni S, Focardi S, Leonzio C (2002) Studies have shown that both extreme warm and Assumption and elimination of trace elements in Adelie penguins from Antarctica: a preliminary study. Mar cold climates could reduce Ade´lie penguin populations Environ Res 54:341–344 by impairing their access to nesting and foraging Arrigo KR, van Dijken GL, Ainley DG, Fahnestock MA, habitats (Fraser et al. 1992; Sun et al. 2000). There Markus T (2002) Ecological impact of a large Antarctic seems to be an optimal combination of climatic and iceberg. Geophys Res Lett 29(7):1104. doi:10.1029/2001 GL014160 marine ecosystem conditions for Ade´lie penguins, Baroni C, Orombelli G (1994) Abandoned penguin rookeries as and the marked ‘penguin optimum’ around 4,000– holocene paleoclimatic indicators in Antarctica. Geology 2,000 cal yr BP in Antarctica (Baroni and Orombelli 22:23–26 1994; Emslie et al. 2003, 2007; Huang et al. 2009a)isa Bentley MJ, Hodgson DA, Smith JA, Cofaigh CO, Domack EW, Larter RD, Roberts SJ, Brachfeld S, Leventer A, period when such conditions likely occurred. Hjort C, Hillenbrand CD, Evans J (2009) Mechanisms of holocene palaeoenvironmental change in the Antarctic Peninsula region. Holocene 19:51–69 Conclusions Blais JM, Kimpe LE, McMahon D, Keatley BE, Mattory ML, Douglas MSV, Smol JP (2005) Arctic seabirds transport marine-derived contaminants. Science 309:445–445 AMS 14C dates on fossil penguin bones from the Blais JM, Macdonald RW, Mackay D, Webster E, Harvey C, ZOL4 core show that Ade´lie penguins occupied Smol JP (2007) Biologically mediated transport of con- Zolotov Island, Vestfold Hills from at least taminants to aquatic ecosystems. Environ Sci Technol 41:1075–1084 1,765 cal yr BP, corresponding to the local retreat Bricher PK, Lucieer A, Woehler EJ (2008) Population trends of of the SØrsdal Glacier. Phosphorus was identified as Adelie penguin (Pygoscelis adeliae) breeding colonies: a the optimal bio-element indicative of penguin guano spatial analysis of the effects of snow accumulation and input to the sediments in ZOL4, and was used to infer human activities. Polar Biol 31:1397–1407 Brimble SK, Blais JM, Kimpe LE, Mallory ML, Keatley BE, penguin population changes for the past 1,800 years. Douglas MSV, Smol JP (2009) Bioenrichment of trace Ade´lie penguin population dynamics at Zolotov elements in a series of ponds near a northern fulmar Island correlated well with regional climate changes. (Fulmarus glacialis) colony at Cape Vera, Devon Island. Ade´lie penguin populations at Ardley Island and Can J Fish Aquat Sci 66:949–958 Bronge C (1992) Holocene climatic record from lacustrine Zolotov Island both reached high levels around sediments in a freshwater lake in the Vestfold Hills, 1,000 cal yr BP, and experienced a decline around Antarctica. Geogr Ann 74:47–58 *300 cal yr BP, suggesting that cold climates are Croxall JP, Trathan PN, Murphy EJ (2002) Environmental detrimental to Ade´lie penguin populations. change and Antarctic seabird populations. Science 297:1510–1514 Domack EW, Ishman SE, Stein AB, McClennen CE, Jull AT Acknowledgments This study was funded by the National (1995) Late holocene advance of the Mu¨ller Ice Shelf, Natural Science Foundation of China (No. 40730107), the Antarctic Peninsula: sedimentological, geochemical and National Science and Technology Supporting Program palaeontological evidence. Antarct Sci 7:159–170 (2006BAB18B07), the Major State Basic Research and Domack EW, Leventer A, Root S, Ring J, Williams E, Carlson Development Program of China (973 program, No. D, Hirshorn E, Wright W, Gilbert R, Burr G (2003) 2010CB428902) and the Australian Antarctic Division Marine sedimentary record of natural environmental var- Science Project (AAD2873). We thank the Chinese Arctic iability and recent warming in the Antarctic Peninsula. In: and Antarctic Administration, Polar Research Institute of Domack EW, Leventer A, Burnett A, Bindschadler R, China, Australian Antarctic Division for logistical support in Convey P, Kirby M (eds) Antarctic Peninsula climate field, Dr. Renbin Zhu for collection of samples and Dr. William variability: historical and paleoenvironmental perspec- Cooper for assistance with radiocarbon dating. We especially tives. Antarctic Research Series 79, American Geophysi- thank the reviewers for their critical comments and Dr. cal Union, pp 205–222 Dominic Hodgson and Dr. Mark Brenner for their help in Emslie SD, McDaniel JD (2002) Ade’lie penguin diet and improving this manuscript. climate change during the middle to late holocene in

123 284 J Paleolimnol (2011) 45:273–285

northern Marguerite Bay, Antarctic Peninsula. Polar Biol neoglacial episode. J Geophys Res 114:D08117. doi: 25:222–229 10.1029/2008JD011091 Emslie SD, Berkman PA, Ainley DG, Coats L, Polito M (2003) Liu XD, Sun LG, Xie ZQ, Yin XB, Wang YH (2005) A Late-holocene initiation of ice-free ecosystems in the 1300-year record of penguin populations at Ardley Island southern , Antarctica. Mar Ecol-Prog Ser in the Antarctic, as deduced from the geochemical data in 262:19–25 the ornithogenic lake sediments. Arct Antarct Alp Res Emslie SD, Coats L, Licht K (2007) A 45,000 yr record of 37:490–498 Ade´lie penguins and climate change in the Ross Sea, Liu XD, Zhao SP, Sun LG, Luo HH, Yin XB, Xie ZQ, Wang Antarctica. Geology 35:61–64 YH, Liu KX, Wu XH, Ding XF, Fu DP (2006) Geo- Forcada J, Trathan PN, Reid K, Murphy EJ, Croxall JP (2006) chemical evidence for the variation of historical seabird Contrasting population changes in sympatric penguin population on Dongdao Island of the South China Sea. species in association with climate warming. Glob Change J Paleolimnol 36:259–279 Biol 12:411–423 Mann ME, Zhang Z, Rutherford S, Bradley RS, Hughes MK, Fraser WR, Hofmann EE (2003) A predator’s perspective on Shindell D, Ammann C, Faluvegi G, Ni F (2009) Global causal links between climate change, physical forcing and signatures and dynamical origins of the little ice age and ecosystem response. Mar Ecol-Prog Ser 265:1–15 medieval climate anomaly. Science 326:1256–1260 Fraser WR, Trivelpiece WZ, Ainley DG, Trivelpiece SG Masson V, Vimeux F, Jouzel J, Morgan V, Delmotte M, Ciais (1992) Increases in Antarctic penguin populations: P, Hammer C, Johnsen S, Lipenkov VY, Mosley- reduced competition with whales or a loss of sea ice due Thompson E, Petit JR, Steig EJ, Stievenard M, Vaikmae R to global warming? Polar Biol 11:525–531 (2000) Holocene climate variability in Antarctica based Goldschmidt VM (1954) Geochemistry. The Clarendon Press, on 11 ice-core isotopic records. Quat Res 54:348–358 Oxford Mayewski PA, Rohling EE, Stager JC, Karlen W, Maasch KA, Grove JM (1988) The little ice age. Methuen, New York Meeker LD, Meyerson EA, Gasse F, van Kreveld S, Hodgson DA, Johnston NM (1997) Inferring seal populations Holmgren K, Lee-Thorp J, Rosqvist G, Rack F, Staub- from lake sediments. Nature 387:30–31 wasser M, Schneider RR, Steig EJ (2004) Holocene cli- Hodgson DA, Vyverman W, Sabbe K (2001) Limnology and mate variability. Quat Res 62:243–255 biology of saline lakes in the Rauer Islands, eastern McMinn A (2000) Late holocene increase in sea ice extent in Antarctica. Antarct Sci 13:255–270 fjords of the Vestfold Hills, eastern Antarctica. Antarct Hofstee EH, Balks MR, Petchey F, Campbell DI (2006) Soils Sci 12:80–88 of , Cape Hallett, northern Victoria Land, McMinn A, Heijnis H, Harle K, McOrist G (2001) Late- Antarctica. Antarct Sci 18:473–486 Holocene climatic change recorded in sediment cores Huang T, Sun LG, Wang YH, Liu XD, Zhu RB (2009a) Pen- from Ellis Fjord, eastern Antarctica. Holocene 11:291– guin population dynamics for the past 8500 years at 300 Gardner Island, Vestfold Hills. Antarct Sci 21:571–578 Michelutti N, Blais JM, Mallory M, Brash J, Thienpont J, Huang T, Sun LG, Wang YH, Zhu RB (2009b) Penguin Kimpe LE, Douglas MSV, Smol JP (2010) Trophic occupation in the Vestfold Hills. Antarct Sci 21:131–134 position influences the efficacy of seabirds as metal bio- Hughen KA, Baillie MGL, Bard E, Bayliss A, Beck JW, vectors. Proc Natl Acad Sci 107:10543–10548 Bertrand CJH, Blackwell PG, Buck CE, Burr GS, Cutler Milliken KT, Anderson JB, Wellner JS, Bohaty SM, Manley KB, Damon PE, Edwards RL, Fairbanks RG, Friedrich M, PL (2009) High-resolution holocene climate record from Guilderson TP, Kromer B, McCormac FG, Manning SW, Maxwell Bay, South Shetland Islands, Antarctica. Geol Bronk RC, Reimer PJ, Reimer RW, Remmele S, Southon Soc Am Bull 121:1711–1725 JR, Stuiver M, Talamo S, Taylor FW, van der Plicht J, Morgan VI (1985) An oxygen isotope—climate record from Weyhenmeyer CE (2004) Marine04 marine radiocarbon the Law Dome, Antarctica. Clim Change 7:415–426 age calibration, 26–0 ka BP. Radiocarbon 46:1059–1086 Roberts D, McMinn A (1999) A diatom-based palaeosalinity Jenouvrier S, Barbraud C, Weimerskirch H (2006) Sea ice history of Ace Lake, Vestfold Hills, Antarctica. Holocene affects the population dynamics of Adelie penguins in 9:401–408 Terre Adelie. Polar Biol 29:413–423 Smith RC, Ainley D, Baker K, Domack E, Emslie S, Fraser B, Keatley B, Douglas M, Blais J, Mallory M, Smol J (2009) Kennett J, Leventer A, Mosley-Thompson E, Stammer- Impacts of seabird-derived nutrients on water quality and john S, Vernet M (1999) Marine ecosystem sensitivity to diatom assemblages from Cape Vera, Devon Island, climate change. Bioscience 49:393–404 Canadian High Arctic. Hydrobiologia 621:191–205 Steig EJ, Morse DL, Waddington ED, Stuiver M, Grootes PM, Lamb FF (1977) Climate: present, past and future. Vol. 2. Mayewski PA, Whitlow SL, Twickler MS (2000) Wis- Climatic history and the future. Methuen, London consinan and holocene climate history from an ice core at Leventer A, Domack EW, Ishman SE, Brachfield S, McClen- Taylor Dome, western Ross Embayment, Antarctica. nen CE, Manley P (1996) Productivity cycles of Geogr Ann 82A:213–235 200–300 years in the Antarctic Peninsula region: under- Stuiver M, Reimer PJ, Reimer RW (2005) CALIB5.1.0. standing linkages among the sun, atmosphere, oceans, sea http://calib.qub.ac.uk/calib/download ice, and biota. Geol Soc Am Bull 108:1626–1644 Sun LG, Xie ZQ (2001a) Changes in lead concentration in Li YS, Cole-Dai JH, Zhou LY (2009) Glaciochemical evidence Antarctic penguin droppings during the past 3,000 years. in an East Antarctica ice core of a recent (AD 1450–1850) Environ Geol 40:1205–1208

123 J Paleolimnol (2011) 45:273–285 285

Sun LG, Xie ZQ (2001b) Relics: penguin population programs. Whitehead MD, Johnstone GW (1990) The distribution and Sci Prog 84:31–44 estimated abundance of Ade´lie penguin breeding in Prydz Sun LG, Xie ZQ, Zhao JL (2000) Palaeoecology—a 3,000-year Bay, Antarctica. Proc NIPR Symp Polar Biol 3:91–98 record of penguin populations. Nature 407:858–858 Wilson PR, Ainley DG, Nur N, Jacobs SS, Barton KJ, Ballard Sun LG, Liu XD, Yin XB, Zhu RB, Xie ZQ, Wang YH (2004a) G, Comiso JC (2001) Adelie penguin population change A 1,500-year record of Antarctic seal populations in in the pacific sector of Antarctica: relation to sea-ice response to climate change. Polar Biol 27:495–501 extent and the Antarctic circumpolar current. Mar Ecol- Sun LG, Zhu RB, Yin XB, Liu XD, Xie ZQ, Wang YH (2004b) Prog Ser 213:301–309 A geochemical method for the reconstruction of the Woehler EJ, Cooper J, Croxall JP, Fraser WR, Kooyman GL, occupation history of a penguin colony in the maritime Miller GD, Nel DC, Patterson DL, Peter HU, Ribic CA, Antarctic. Polar Biol 27:670–678 Salwicka K, Trivelpiece WZ, Weimerskirch H (2001) A Sun LG, Zhu RB, Liu XD, Xie ZQ, Yin XB, Zhao SP, Wang statistical assessment of the status and trends of Antarctic YH (2005) HCI-soluble Sr-87/Sr-86 ratio in sediments and Subantarctic seabirds. Scientific Comm Antarct Res, impacted by penguin or seal excreta as a proxy for his- Hobart torical population size in the maritime Antarctic. Mar Xie ZQ, Sun LG (2008) A 1,800-year record of arsenic con- Ecol-Prog Ser 303:43–50 centration in the penguin dropping sediment, Antarctic. Tatur A, Keck A (1990) Phosphates in ornithogenic soils of the Environ Geol 55:1055–1059 maritime Antarctic. Proc NIPR Symp Polar Biol Yang QC, Sun LG, Kong DM, Huang T, Wang YH (2010) 3:133–150 Variation of Antarctic seal population in response to Taylor F, McMinn A (2002) Late quaternary diatom assem- human activities in 20th century. Chin Sci Bull blages from Prydz Bay, Eastern Antarctica. Quat Res 55:1084–1087 57:151–161 Yin XB, Xia LJ, Sun LG, Luo HH, Wang YH (2008) Animal Verleyen E, Hodgson DA, Cremer H, Emslie SD, Gibson J, excrement: a potential biomonitor of heavy metal con- Hall B, i Imura S, Kudoh S, Marshall GJ, McMinn A, tamination in the marine environment. Sci Total Environ Melles M, Newman L, Roberts D, Roberts SJ, Sabbe K, 399:179–185 Singh SM, Sterken M, Tavernier I, Verkulich S, Van de Yoo KC, Yoon HI, Kim JK, Khim BK (2009) Sedimentolog- Vyver E, Van Nieuwenhuyze W, Wagner B, Vyverman W ical, geochemical and palaeontological evidence for a (2010) Post-glacial regional climate variability along the neoglacial cold event during the late holocene in the East Antarctic coastal margin—evidence from shallow continental shelf of the northern South Shetland Islands, marine and coastal terrestrial records. Earth Sci Rev. doi: West Antarctica. Polar Res 28:177–192 10.1016/j.earscirev.2010.10.006 Zdanowski MK, Zmuda MJ, Zwolska W (2005) Bacterial role Wang JJ, Wang YH, Wang XM, Sun LG (2007) Penguins and in the decomposition of marine-derived material (penguin vegetations on Ardley Island, Antarctica: evolution in the guano) in the terrestrial maritime Antarctic. Soil Biol past 2, 400 years. Polar Biol 30:1475–1481 Biochem 37:581–595

123