Abstract: Travail s'appuyant sur les recherches menées entre 1989 et 1995 sur le mouflon à Kerguelen dans le cadre du programme IPEV 276 (responsable J.L. Chapuis)

Abstract: Training period report (advisor: Michel Piat)

Abstract: Tracking data of marine predators are increasingly used in marine spatial management. We developed a spatial data set with estimates of the monthly distribution of 6 pelagic seabird species breeding in the Northeast Atlantic. The data set was based on year-round global location sensor (GLS) tracking data of 2356 adult seabirds from 2006-2019 from a network of seabird colonies, data describing the physical environment and data on seabird population sizes. Tracking and environmental data were combined in monthly species distribution models (SDMs). Cross-validations were used to assess the transferability of models between years and breeding locations. The analyses showed that birds from colonies close to each other (<500 km apart) used the same nonbreeding habitats, while birds from distant colonies (>1000 km) used colony-specific and, in many cases, non-overlapping habitats. Based on these results, the SDM from the nearest model colony was used to predict the distribution of all seabird colonies lying within a species-specific cut-off distance (400-500 km). Uncertainties in the predictions were estimated by cluster bootstrap sampling. The resulting data set consisted of 4692 map layers, each layer predicting the densities of birds from a given species, colony and month across the North Atlantic. This data set represents the annual distribution of 23.5 million adult pelagic seabirds, or 87% of the Northeast Atlantic breeding population of the study species. We show how the data set can be used in population and spatial management applications, including the detection of population-specific nonbreeding habitats and identifying populations influenced by marine protected areas.

Abstract: Trace metals such as Cu, Hg, and Zn have been widely investigated in marine ecotoxicological studies considering their bioaccumulation, transfer along trophic webs, and the risks they pose to ecosystems and human health. Comparatively, Li has received little attention, although this element is increasingly used in the high-tech, ceramics/glass, and medication industries. Here, we report Li concentrations in more than 400 samples, including whole organisms and different organs of bivalves, cephalopods, crustaceans, and fish. We investigated species from three contrasting biogeographic areas, i.e. temperate (Bay of Biscay, northeast Atlantic Ocean), tropical (New Caledonia, Pacific Ocean), and subpolar climates (Kerguelen Islands, southern Indian Ocean), among diverse trophic groups (filter-feeders to meso-predators) and habitats (benthic, demersal, and pelagic). Although Li is homogeneously distributed in the ocean (at 0.18 μg/mL), Li concentrations in soft tissues vary greatly, from 0.01 to 1.20 μg/g dry weight. Multiple correspondence analyses reveal two clusters of high and low Li concentrations. Li distributions in marine organisms appear to be mostly geographically independent, though our results highlight a temperature dependency in fish muscles. Li is consistently bio-reduced through the trophic webs, with filter-feeders showing the highest concentrations and predatory fish the lowest. Strong variations are observed among organs, consistent with the biochemical similarity between Na and Li during transport in the brain and in osmoregulatory organs. Fish gills and kidneys show relatively high Li concentrations (0.26 and 0.15 μg/g, respectively) and fish brains show a large range of Li contents (up to 0.34 μg/g), whereas fish liver and muscles are Li depleted (0.07 ± 0.03 and 0.06 ± 0.08 μg/g, respectively). Altogether, these results provide the first exhaustive baseline for future Li ecotoxicology studies in marine coastal environments.

Abstract: Total column ozone reduction in the Arctic is evaluated each winter since 1993/1994 by the transport method (3-D CTM passive ozone minus measurements). The cumulative loss from 1 December to the end of the season ranges from 5–10% during warm winters like 1998/1999, 2000/2001 and 2001/2002 up to 30%–32% during cold winters like 1994/1995 and 1995/1996. The 23% cumulative loss observed during the winter 2002/2003 is similar in amplitude to the 20–24% measured in 1996/1997 and 1999/2000 but the timing is different. It started unusually early in December after the occurrence of very low temperature at all stratospheric levels between 550 K and 435 K allowing PSC formation and thus chlorine activation. The early ozone loss of 2002/2003 is well captured by current 3-D CTM models.