Pre-proof Exposure of a small Arctic seabird , the little auk ( Alle alle ) breeding in Svalbard , to selected elements throughout the course of a year

The Arctic marine ecosystem can be altered by processes of natural and anthropogenic origin. Spatio-temporal variation in species exposure to contamination is still poorly understood. Here, we studied elemental concentrations in the non-lethally collected samples from the most numerous seabird in European Arctic, the little auk (Alle alle) nesting in one breeding colony in Svalbard. This seabird spent the breeding season in the high-Arctic zone and the non-breeding period in sub-Arctic areas what may implicate spatio-temporal variation in elements bioaccumulation. We determined concentrations of 19 elements in adults feathers to determine levels of exposure during part of the pre-breeding (n = 74) and post-breeding (n = 74) seasons, feathers from nestlings (n = 18) to determine local Journal Pre-proof

J o u r n a l P r e -p r o o f 3 human activities, both past and present, including runoffs from active and shut-down coal mines, ship transportation and increased tourism (Rose et al. 2004;Granberg et al. 2017).
As top predators, seabirds are considered as a good model group for studying the bioaccumulation and biomagnification of elements (Burger et al. 2008). Moreover, they play an important role in Arctic ecosystems by being two-environmental organisms; foraging at sea and breeding on land, they may easily transport various types of matter, including contaminants, from marine areas to terrestrial ones (Stempniewicz et al. 2007;Burger et al. 2007).
The aim of this study was to examine the concentrations of certain elements in samples collected from the High Arctic seabird, the little auk (Alle alle). We took advantage of the fact that birds excrete the elements into external tissues such as feathers (changed regularly during the moulting, e.g. Jaspers et al. 2004;Burger et al. 2008;Fort et al. 2016;Pacyna et al. 2017), and eggshells (females; Lam et al. 2005). Thus, sampling those allowed us to examine birds exposure to contaminants relatively non-invasively. Since little auks moult twice a year in different time and area, various feather types can be used to trace changes in exposure to elements over an annual cycle and to examine changes in food-chain contamination (Fort et al. 2016). We sampled throat and body feathers: throat feathers are changed during the partial pre-breeding moult (Fort et al. 2014; SEATRACK), thus represent contamination exposure during the wintering; body feathers are moulted after the breeding (Fort et al. 2014;2016;SEATRACK). Elemental concentrations in chick feathers, being affected by fewer factors (including the fact that food comes from a more restricted foraging area of the parents during the breeding season) can be used to examine local contamination effect (Evers et al. 2005). Chick down and post-hatching eggshells can be used to examined females contamination prior to egg-laying, thus also a maternal effect in the contamination Journal Pre-proof J o u r n a l P r e -p r o o f 4 level of the chicks (Ackerman et al. 2016). Sampling the body and throat feathers of adults, post-hatching eggshells, and the body feathers and down of chicks we were able to build a full-year picture of the spatio-temporal exposure of the local population to various elements, including contaminants. Due to differences in the adults moulting location (outside the breeding colony, i.e. various quarters of the Northern Atlantic, including sub-arctic area) and chicks (breeding colony, High Arctic) we expected higher concentration of various elements in adults. Also, since the isotopic niches (reflecting foraging niches) occupied by little auks from the colonies in East Greenland and Spitsbergen change throughout their annual cycle (Fort et al. 2010), we also expected differences between adults feathers being moulted in different time (preand post-breeding). Due to sex specific deposition of the elements into egg, we also expected sex differences in elemental concentration in post-breeding feathers of adults.
Additionally, owing to the fact of sampling the body feathers of the same adult individuals in two consecutive breeding seasons we could examine temporal patterns of individual contamination levels.
J o u r n a l P r e -p r o o f 5 (Steen et al. 2007;Wojczulanis et al. 2006, Jakubas et al. 2011. In winter, when the availability of Calanus sp. decreases dramatically due to its migration to greater depths (Fort et al. 2010), little auks switch to higher trophic level prey, such as larger amphipods, krill or small fish (Fort et al. 2010;Stempniewicz 2001;Rosing-Asvid et al. 2013).
The little auk has been proven to be a valuable bio-indicator of the temporal trends of environmental contamination within the Arctic pelagic zone (Borgå et al. 2006;Jaeger et al. 2009;Fife et al. 2015;Fort et al. 2016). It has been recognized as the most numerous seabird in the European part of the Arctic (Stempniewicz, 2001) and a keystone species for the functioning of the Arctic terrestrial ecosystem through its transportation of nutrients from the sea to the land (Stempniewicz et al. 2007). Thus, the knowledge about little auks contamination is important for understanding the circulation of elements in both marine and terrestrial Arctic environments.

Studied area and types of sampled tissues
The study was performed in the little auk breeding colony on the Ariekammen slope in  (Fort et al. 2010;2013;2014;2016;SEATRACK). From nestlings we collected down and newly grown feathers: down (hereafter CHDOW) was collected from chicks at 7-14 days old and body feathers from the same chicks at 14-23 days old (hereafter CHBFR).
Residual post-hatching eggshells (hereafter EGGSCH) were also collected from focal nests just after hatching (little auk parents remove eggshells from the nest burrow a few days after hatching, thus our collection could not affect the birds' welfare). For 20 adults sampled in 2017, body feathers collected in the previous season (2016) were available. They were used to investigate an inter-annual variation in elemental concentration.
After collection, feathers were kept at an ambient temperature in sealed string bags.
Eggshells were kept in plastic containers in a refrigerator (-20°C).  (Griffiths et al. 1998). This size difference was clearly visible under UV-light when the fragments were dyed with Midori green and separated on a 2% agarose gel.
In total, feathers from 74 adult birds were analysed, but a reliable sex determination was only possible for 54 individuals (26 males and 28 females).

Sample preparation
All feather samples were cleaned once with acetone and twice with deionised water (Mili-Q Water, France) to remove external contaminations, and air-dried for about 24h.
Next, the dry feathers were homogenised by being cut into small parts and weighed to the nearest 0.01 mg (mean sample mass of adult body feathers: 24.33 ± 5.81 mg, throat: 7.65 ± 2.07 mg, chick feathers: 16.09 ± 4.32 mg, chick down feathers: 27.21 ± 12.16 mg). In the case of eggshells, the internal membrane was discarded, as it was too contaminated, and samples were washed with acetone and deionised water. Samples were left to dry for 24h, then crushed into small pieces and weighed (mean 203.30 ± 4.10 mg). All samples were placed individually into a clean Teflon vessel with 65% HNO 3 (Merck, Suprapure). Digestion was carried out using a high-pressure microwave emitter (Microwave Digestion System, Anton J o u r n a l P r e -p r o o f 9 For statistical analysis, the arithmetic mean was calculated if at least 65% of the samples had concentrations of the compound >LOD.

Statistical analyses
Although feather samples were collected from family members, potentially imposing dependency in the data set, they were treated as independent data points in all the analyses. This is because feathers of both types (pre-and post-breeding) collected from adult males and females have grown independently (both in a geographical and temporal sense). In addition, all the individuals were represented in each sample group in similar numbers, thus the issue of possible pseudoreplication could be considered negligible. To examine the level of contamination of chemical elements in little auks in respect to sample type and birds sex, a data mining approach was applied, using various methods: The molar ratio of Se and Hg in the examined samples was also analysed. This is because both elements can be bioaccumulated in significant amounts (Borghesi et al. 2016).
Se is an essential element which level is physiologically regulated within the body and plays an important role in the organism's proper functioning, including protection against the adverse effects of Hg, by the creation of an Hg-Se complex (Hg binds to Se with an extraordinarily high affinity; Berry and Ralston 2008; Khan and Wang 2009;Øverjordet et al. 2015). Hg toxicity is observed when Hg has a substantial molar excess of Se (Berry and Ralston 2008). We used the Kruskal-Wallis tests, with post-hoc Mann-Whitney U tests, to compare the molar ratio, Se:Hg between the sample types. Eggshells and chick feathers were excluded from this analysis because of the high representation of samples with Hg or Se concentrations <LOD (see Table 1).
To analyse sex differences in elemental concentrations in post-breeding feathers of adults, t-tests were used. To investigate inter-annual consistency in elemental concentrations in feathers of the same individuals captured in 2016 and 2017 Pearson correlation was used.
PCA, and PERMANOVA analyses were performed on log(x + 1) transformed data in PAST software (Hammer et al. 2001) and all other analyses in R software (R Core Team 2018).

Results
Elemental concentrations in all samples are reported in Table 1. In the case of Bi in chick body feathers, the results suggested possible external contamination, as extraordinary variations were found (65% of samples were <LOQ, range was from <LOD-92.40 μg · g ˗1 dw).

Differences in all elemental concentrations between the studied types of samples
Principal component analysis (PCA) revealed that 75.7% of the total variance in the elemental concentrations in the studied types of samples was explained by the two axes.
The first axis (explaining 55.4% of total variance) was the most correlated with K (r = 0.625), Ca (r = 0.495) and Sr (r = 0.406), and the second axis (explaining 20.3% of total variability) with K (r = -0.712) and Ca (r = 0.463) (Fig. 2). PREBR and POSTBR clustered in similar positions and partly overlapped with CHDOW. EGGSCH and CHBFR clustered in position different from any other groups not overlapping with any of them (Fig. 2).

Differences in particular elements between the studied types of samples
Kruskal-Wallis inter-group tests comparing the concentrations of particular elements in the studied types of samples, i.e. feathers PREBR, POSTBR, CHDOW, CHBFR and EGGSCH revealed significant differences for all 19 elements (p < 0.05) (Supplementary Materials 1,   Fig. ES1-ES8). Post-hoc U Mann-Whitney tests revealed that most of differences were significant (p < 0.05) (Electronic Supplementary Materials 1, Fig. ES1-ES7).

Differences in elemental concentration (by sex and year of sampling)
Ca and Zn concentrations in the POSTBR body feathers of females were significantly higher than in males. Concentrations of Hg also tended to be higher in females, but results were not statistically significant (p = 0.057). No significant sex differences were found in concentrations of other studied elements in POSTBR body feathers (Table 2).
An analysis of inter-annual consistency in elemental concentrations in the same individuals sampled both in 2016 (side body feathers) and 2017 (back feathers) revealed no statistically significant relationships, except for a positive correlation for Pb (Table 3). Table 1. Elemental concentration in little auk samples mean ± SD (median; min-max) μg · g ˗1 dw; all sampled individuals; Se:Hg molar ratio calculated only when more than 65% of both Se and Hg were above LOD Journal Pre-proof Table 2. Summary of element concentration [μg · g ˗1 dw] in adult body feathers sampled during the post-breeding period. Difference between males (n = 26) and females (n = 27) analysed with t-test Table 3. Inter-annual consistency in element concentrations (Pearson correlation) in the back feathers of adult little auks captured both in 2016 and 2017

Discussion
The elemental exposure of the most numerous High Arctic seabird throughout the year was characterized based on non-lethally collected samples. Several significant differences between the concentrations of the studied elements were found (only elements of known relevance and toxicity are discussed).

Hg and Se
Hg is an element of primary concern in marine environment. In the present study, 35% of all the analysed adults feathers representing the pre-breeding period exceed the Hg toxicity threshold value for feathers (5 μg · g ˗1 dw; Burger and Gochfeld 1997), with 11% of individuals having concentrations > 10 μg · g ˗1 dw. Lack of samples with Hg levels exceeding 5 μg · g ˗1 dw during the post-breeding period suggests lower Hg contamination in breeding colonies area, located in the High-Arctic zone. One should remember that the established toxicity threshold associated with Hg adverse effect (Burger and Gochfeld 1997) in birds is a J o u r n a l P r e -p r o o f 14 guideline and real effect on avian organism may differ between species. However, one third of individuals with elevated level of Hg during the pre-breeding period is alarming.
Previous study on adult little auks breeding in Greenland has also shown higher Hg concentrations in feathers grown during the pre-breeding period (2.27 -3.73 μg · g ˗1 dw) compared to the post-breeding period (1.00 -2.11 μg · g ˗1 dw) (Fort et al. 2016). In the present study conducted in Svalbard, even higher values were found, i.e. 1.45-17.2 (mean ± SD: 5.02 ± 3.32 μg · g ˗1 dw) during the pre-breeding and 0.58-2.36 (1.39 ± 0.40 μg · g ˗1 dw) the post-breeding period. These results may be explained by the higher exposure of little auks during the pre-breeding time (reflecting contamination outside breeding grounds), when they are more exposed to human activities (oil drilling, oil spills, etc.) compared to the Here, As levels <0.2 μg · g ˗1 dw in body feathers, and <LOD in throat feathers were found. When excluding outliers, levels of As, Pb, and Cd did not suggest any potential ecotoxicological risk.
The highest absolute elemental concentration in both back and throat feathers were found for Mg, Ca and Zn. These elements are necessary for a proper feather formation and are regulated mostly by homeostatic processes (Bocher et al. 2003).

Temporal variation in levels of elements in respect to sex
Exposure to contaminants may differ between females and males due to sexspecific diets, and spatial and/or temporal differences in stop-over or wintering strategies

Pathways of elemental input in chicks
Elemental allocation from the female organism to the chick down serves as a major elimination pathway for Hg accumulated by the mother (Wenzel et al. 1996). Se transferred from females plays a significant role in prenatal Hg intoxication, reducing its toxicity (Berry andRalston 2008, Hargreaves et al. 2011). Both Hg and Se levels as well as the molar ratio of the two elements were higher in chick down than chick body feathers (Table 1). However, as elevated Se level can also be toxic for organism it is difficult to predict its full toxicological effect (Khan andWang, 2009, Hargreaves et al. 2011).
Ca levels tend to increase in a little auk chick's body after the 10 th day of its life, suggesting intense ossification of the skeleton at that time (Taylor and Konarzewski, 1992).
The nutritional deficiencies in chicks' diets, especially in Ca, could result in delayed growth.
Thus Ca concentration being very high in the down (238 ± 66 μg · g ˗1 dw) compared to the chick body feathers (30.20 ± 12.10 μg · g ˗1 dw) could be related to a protective maternal effect and Ca absorption from the shell (Orłowski et al. 2017). Similarly, Mg and K levels could be under a strong maternal influence. The level of these two elements have been reported to be relatively stable throughout the little auk's early development (Taylor and  Internal elemental deposition depends on several factors including feather physiology. It is known that some elements are incorporated into the feather as part of the building blocks of the keratin, while others can enter the developing feather cells in proportion to their abundance in the blood stream (Bortolotti, 2010). Deposition of some elements is time-dependent and depends on feather growth rate. Time-dependent scale, which is based on feather measurements not mass, can be very useful for flight feathers that grow in a highly linear fashion, with relatively uniform elongation during growth (Bortolotti, 2010). Here as we used very small body and throat feathers, as well as chick down and body feathers we applied mass dependent scale measuring elemental concentration, to enable a uniform comparison between them. In the case of some elements it could cause mass dilution bias, but this phenomena requires more examination.

Conclusions
This study revealed that toxic metal levels in the feathers and eggshells of a high Arctic breeder, the little auk, were generally low. However, in the case of Hg, we found that almost one third of adult birds had elevated Hg level in feathers grown during the prebreeding period.
Concentrations of Hg, Se and Mn in adult feathers were significantly higher in prebreeding period reflecting higher exposition of birds to contaminants in areas outside the High Arctic where they performed pre-breeding moult and lower risk in the High Arctic zone where they moult during the post-breeding period. Sex related differences were mostly nonsignificant, beside Ca and Zn.