Adélie penguins foraging consistency and site fidelity are conditioned by breeding status and environmental conditions

Circumpolar assessment of mercury contamination: the Adélie penguin as a bioindicator of Antarctic marine ecosystems

Due to its persistence and potential ecological and health impacts, mercury (Hg) is a global pollutant of major concern that may reach high concentrations even in remote polar oceans. In contrast to the Arctic Ocean, studies documenting Hg contamination in the Southern Ocean are spatially restricted and large-scale monitoring is needed. Here, we present the first circumpolar assessment of Hg contamination in Antarctic marine ecosystems. Specifically, the Adélie penguin (Pygoscelis adeliae) was used as a bioindicator species, to examine regional variation across 24 colonies distributed across the entire Antarctic continent. Mercury was measured on body feathers collected from both adults (n = 485) and chicks (n = 48) between 2005 and 2021. Because penguins' diet represents the dominant source of Hg, feather δ13C and δ15N values were measured as proxies of feeding habitat and trophic position. As expected, chicks had lower Hg concentrations (mean ± SD: 0.22 ± 0.08 μg·g‒1) than adults (0.49 ± 0.23 μg·g‒1), likely because of their shorter bioaccumulation period. In adults, spatial variation in feather Hg concentrations was driven by both trophic ecology and colony location. The highest Hg concentrations were observed in the Ross Sea, possibly because of a higher consumption of fish in the diet compared to other sites (krill-dominated diet). Such large-scale assessments are critical to assess the effectiveness of the Minamata Convention on Mercury. Owing to their circumpolar distribution and their ecological role in Antarctic marine ecosystems, Adélie penguins could be valuable bioindicators for tracking spatial and temporal trends of Hg across Antarctic waters in the future.

The role of individual variability on the predictive performance of machine learning applied to large bio-logging datasets

Animal-borne tagging (bio-logging) generates large and complex datasets. In particular, accelerometer tags, which provide information on behaviour and energy expenditure of wild animals, produce high-resolution multi-dimensional data, and can be challenging to analyse. We tested the performance of commonly used artificial intelligence tools on datasets of increasing volume and dimensionality. By collecting bio-logging data across several sampling seasons, datasets are inherently characterized by inter-individual variability. Such information should be considered when predicting behaviour. We integrated both unsupervised and supervised machine learning approaches to predict behaviours in two penguin species. The classified behaviours obtained from the unsupervised approach Expectation Maximisation were used to train the supervised approach Random Forest. We assessed agreement between the approaches, the performance of Random Forest on unknown data and the implications for the calculation of energy expenditure. Consideration of behavioural variability resulted in high agreement (> 80%) in behavioural classifications and minimal differences in energy expenditure estimates. However, some outliers with < 70% of agreement, highlighted how behaviours characterized by signal similarity are confused. We advise the broad bio-logging community, approaching these large datasets, to be cautious when upscaling predictions, as this might lead to less accurate estimates of behaviour and energy expenditure.