HAL - La Rochelle Université, Archives Ouvertes
Not a member yet
19516 research outputs found
Sort by
Mercury contamination of sympatric seabirds and associated health risks in an Antarctic ecosystem
No full textInternational audienceAntarctic marine ecosystems are located far from industrial pollution sources, yet mercury (Hg) contamination remains an important threat to regional biodiversity. Seabirds occupy mid- to high trophic positions in Antarctic food webs, and can show high levels of Hg contamination due to biomagnification. Here, total Hg (THg) concentrations and stable isotopes of carbon (δ13C) and nitrogen (δ15N) were measured in red blood cells of brown skuas Stercorarius antarcticus lonnbergi (n = 44) and south polar skuas S. maccormicki (n = 50) at King George Island/Isla 25 de Mayo (South Shetland Islands) in the 2022/23 and 2023/24 breeding seasons. The objectives were to: (i) determine current levels of Hg contamination at King George Island and identify the underlying drivers (e.g., species, sex, year, δ13C and δ15N); (ii) compare contaminant levels with other breeding sites; and (iii) examine potential Hg-associated health risks. At King George Island, south polar skuas had higher THg concentrations (mean ± SD, 3.85 ± 2.99 μg g-1 dw) than brown skuas (1.67 ± 1.25 μg g-1 dw), potentially due to their greater reliance on mesopelagic fish and carry-over effects from their non-breeding distributions. THg concentrations of males were higher than females, with deposition into eggs by females being the likeliest explanatory factor, and were positively related to δ15N, reflecting the biomagnification process. THg concentrations of brown skuas in this study were higher than at Hope Bay (Antarctic Peninsula), but lower than at South Orkney Islands (Antarctica), South Georgia and Kerguelen Islands (subantarctic). THg concentrations of the south polar skuas analysed here were higher than at Hope Bay and Adélie Land (Antarctic continent). Comparisons with toxicity benchmarks suggest that skuas are currently at low risk of Hg-associated health impacts at King George Island
La première voie de quai dans le port de Bordeaux sous le Second Empire. Enjeux, résistances et controverses
No full textInternational audienc
biolutoxR: An R-Shiny package for easy performing data analysis of a toxicity test based on bacterial bioluminescence inhibition
No full textInternational audienceIn the 21st century, in the context of an environmental crisis, chemical pollution has become a major global concern. In addition to chemical analysis, many bioassays have been developed and have proved to be interesting tools for considerably improve our understanding of the effects of human activities on species and ecosystems. Compared with animal or plant bioassays, bacterial bioassays based on bioluminescence inhibition have the advantage of being relatively economical and reproducible. This toxicity test uses the bioluminescent metabolic response of bacteria exposed for a short time to a solution of interest in the aim of assessing and quantifying its toxicity. Although routine tools are available to pre-process the data obtained, to our knowledge no flexible research tool is available for the data analysis following these tests, which could limit the access to the results for novices or experienced people unfamiliar with data processing software. To overcome this lack of tools for this bioassay, an R-Shiny package is proposed to generalise data analysis following a toxicity test based on bacterial bioluminescence inhibition. The traditional paper-based working environment is reproduced digitally in this package, which ultimately facilitates data entry and cleaning, makes the creation of relevant dynamic graphs, and simplifies access to toxicity data (e.g. dose-response curve and median effective concentration, i.e. EC 50 ). The aim of this tool is to provide the target community with a high-performance tool that can be used to obtain toxicity test results based on the inhibition of bacterial bioluminescence
Major and trace nutrient concentrations in sub-Antarctic forage fish around the Kerguelen Islands: All are not equal for nutrient recycling by top predators
No full textInternational audienceWhile the nutrient content of forage species is crucial to study how marine top predators release nutrients in their waste (faeces and urine) and possibly facilitate nutrient recycling at lower trophic levels of the food webs, it is poorly documented in many ecosystems. Here, we analyse the concentrations of five major nutrients, eight essential trace nutrients, and four non-essential nutrients in 34 mesopelagic and demersal forage fish species from the Kerguelen Plateau area (Southern Indian Ocean). This biogeochemically singular region benefits from natural iron (Fe) fertilisation and is an oasis for marine life in the high nutrient low chlorophyll Southern Ocean. Its nutrient dynamics have been subject to intense research for decades, but the contribution of the biota remain unexplored. Local forage fish were found to be significantly heterogeneous in their absolute nutrient concentrations, and in their relative nutrient content. Using hierarchical clustering, we showed that some species were enriched in nutrients found in limited concentrations in Southern Ocean surface waters, such as Fe and manganese (Mn), whereas others were enriched in nutrients likely limiting primary productivity in terrestrial ecosystems, such as phosphorus (P) or calcium (Ca). These fish species could be major contributors to these nutrients' biological cycling in different ecosystems when consumed by predators. This functional typology further demonstrates that prey are not interchangeable. The consumption of one or another prey species modulates not only the ability of the predator to meet its requirements and exposure to contaminants, but also predator-mediated nutrient recycling. This work provides valuable baseline knowledge that can be used in studies of nutrition, nutrient dynamics, and contaminant transfer. It also paves the way for the inclusion of predator-prey relationships in local biogeochemical studies to help unravel the complex processes at work in this unique ecosystem
Des femmes en Amérique: Une histoire des États-Unis de Pocahontas à #MeToo
No full textNational audienceRemember the ladies...En 1776, alors qu'est préparée la Déclaration d'indépendance des États-Unis, l'épouse d'un de ses rédacteurs, Abigail Adams, enjoint son mari à " se souvenir des dames ". Pourtant, comment peut-on prétendre comprendre le passé d'un si grand pays en se plaçant uniquement d'un point de vue masculin ? Comment appréhender les rapports de force qui caractérisent l'histoire américaine à différentes époques – depuis la formation des premières sociétés coloniales au XVIIe siècle jusqu'à l'apogée actuel de l'hyperpuissance qu'est le pays de l'Oncle Sam – sans s'intéresser de très près à la moitié de sa population
La pucele qui voloit voler - Édition, traduction et notes d'après le manuscrit Bern, Burgerbibliothek, Cod. 354
No full textÉdition, traduction et notes de La pucele qui voloit voler d'après le ms. Bern, Burgerbibliothek, Cod. 35
Le pescheor de Pont seur Saine - Édition, traduction et notes d'après le manuscrit Berlin, Staatsbibliothek und Preussischer Kulturbesitz, Hamilton 257
No full textÉdition, traduction et notes du Pescheor de Pont seur Saine d'après le manuscrit Berlin, Staatsbibliothek und Preussischer Kulturbesitz, Hamilton 25
La dame escoillee - Édition et notes d'après le manuscrit BnF fr. 19152
No full textÉdition et notes de La dame escoillee d'après le ms. BnF fr. 1915
Local Change Detection from Video Scene: An End-to-End Encoder–Decoder-Type Deep Neural Network Architecture with Feature Pooling
No full textInternational audienceIn the field of computer vision tasks, substantial importance lies in moving object detection. Background subtraction (BS) is a frequently used scheme for identifying objects in motion in video scenes. However, the BS technique’s efficacy is dependent on background modeling. As a result, background modeling is effective in dealing with a variety of issues, such as variation in illumination, dynamic background, and poor resolution. Therefore, in this work, we have developed end-to-end deep learning frameworks for BS. Initially, we proposed a VGG-19 deep neural architecture encoder with a transfer learning (TL) strategy-induced feature pooling module (FPM) capable of extracting multi-scale objects at different scales. The proposed decoder framework contains several transposed convolution layers that efficiently project feature space to image space. Further, we have developed a BS scheme where a ResNet-50 deep neural network (DNN) with an FPM block learns to project from image to feature space efficiently. The proposed decoder framework contains stacked convolutional layers that can separate the foreground and background pixels effectively for the target frame. The efficacy of the designed techniques is validated by comparing the outcomes obtained by the VGG-19-based network against 17 existing techniques and the results achieved by the ResNet-50-based network against 28 state-of-the-art (SOTA) schemes
Development of the ability to assess (fake) news about ecology from adolescence to older adults
No full textInternational audienceMisinformation about ecology spreads rapidly in social networks where teenagers, as well as older adults, spend much of their time. Given that teenagers and older adults have cognitive specificities that could make them more or less vulnerable to fake news, in the present study, we investigated the development of the ability to discern fake news from real news about ecology from adolescence to older adults. Adolescents (15–18 years old), young adults (19–23 years old), middle-aged adults (30–45 years old), and older adults (65–80 years old) assessed fake versus real news on ecology and answered questions regarding their attitudes toward eco-friendly behaviors, the resources available to help them engage in these behaviors and their concerns regarding climate change. Finally, the participants completed the cognitive reflection test (CRT) to examine whether their analytical thinking skills were related to fake news detection. The results revealed that adolescents are better at detecting fake news about ecology than older participants are. Results confirmed that analytical reasoning is involved in media truth discernment at all ages and also revealed the valuable influence of the participants’ ecological profiles on the news valuations. This latter effect is, however, absent in adolescents, suggesting that the advantage in media truth discernment for teenagers might result from the critical analysis of information developed in secondary school lessons