Feather mercury concentrations in Southern Ocean seabirds: Variation by species, site and time.

We studied mercury contamination in 25 seabird species breeding along a latitudinal gradient across the Southern Ocean, from Gough Island (40°S) through Marion Island (47°S) to Byers Peninsula (63°S). Total mercury concentrations in body feather samples of adults caught at breeding colonies from 2008 to 2011 were determined. Krill (Euphausia spp.) and other zooplankton consumers had low mercury concentrations (gentoo penguin Pygoscelis papua, chinstrap penguin Pseudomonas Antarctica, common diving petrel Pelecanoides urinatrix, broad-billed prion Pachyptila vittata; mean levels 308-753 ng g(-1)), whereas seabirds consuming squid or carrion had high mercury concentrations (ascending order: Kerguelen petrel Aphrodroma brevirostris, southern giant petrel Macronectes giganteus, soft-plumaged petrel Pterodroma mollis, sooty albatross Phoebetria fusca, Atlantic petrel Pterodroma incerta, northern giant petrel Macronectes halli, great-winged petrel Pterodroma macroptera; 10,720-28038 ng g(-1)). The two species with the highest mercury concentrations, northern giant petrels and great-winged petrels, bred at Marion Island. Among species investigated at multiple sites, southern giant petrels had higher mercury levels at Marion than at Gough Island and Byers Peninsula. Mercury levels among Byers Peninsula seabirds were low, in two species even lower than levels measured 10 years before at Bird Island, South Georgia. Replicate measurements after about 25 years at Gough Island showed much higher mercury levels in feathers of sooty albatrosses (by 187%), soft-plumaged petrels (53%) and Atlantic petrels (49%). Concentrations similar to the past were detected in southern giant petrels at Gough and Marion islands, and in northern giant petrels at Marion. There were no clear indications that timing of moult or migratory behavior affected mercury contamination patterns among species. Causes of inter-site or temporal differences in mercury contamination could not be verified due to a lack of long-term data related to species' diet and trophic levels, which should be collected in future together with data on mercury contamination.