HAL - La Rochelle Université, Archives Ouvertes
Not a member yet
19516 research outputs found
Sort by
"Kangxi et Louis XIV" 《康熙与路易十四》
No full textInternational audienceL’année 2024 marque le 60e anniversaire de l'établissement des relations diplomatiques entre la Chine et la France. Les relations culturelles entre les deux pays remontent aux premières années de la dynastie Qing et à une figure centrale, celle de Joachim Bouvet (1656-1730), appelé Bai Jin par les Chinois mais également Ming Yuan, qui fut l'un des cinq premiers missionnaires jésuites envoyés en Chine par le roi de France Louis XIV. Arrivé en Chine en 1687, il mourut à Pékin et fut enterré dans le cimetière du temple de Zhengfu également connu sous le nom de cimetière français. Grâce à ses réalisations pionnières dans de nombreux domaines, Joachim Bouvet a toujours été la figure la plus observée parmi les missionnaires français en Chine
Composition et préservation du carbone organique dans les sédiments de zones humides côtières macrotidales : apports des biomarqueurs et des signatures isotopiques
No full textInternational audienceCoastal wetlands store high amounts of organic carbon (OC) in their sediments, but long-term preservation of this carbon depends on habitat type, sediment depth, and the molecular characteristics of organic matter (OM). This study explores the dynamics of OC deposition and preservation across vertical profiles (0–30 cm) in two adjacent coastal habitats—mudflat, and salt-marsh—within the macrotidal system of the Aiguillon Bay (France). A multi-tracer approach was applied, combining stable isotopes δ13C, C/N ratios, lignin phenols, and fatty acids. Sediment OC content ranged from 13.4 to 23.2 mgC g−1, with the highest concentrations found in the salt-marsh. δ13C and C/N signatures revealed dominant marine source in the mudflat, with a secondary contribution from microphytobenthos, and mixed marine–C₃ plant inputs in the salt-marsh. Fatty acids and lignin compositions supported this partitioning, with surface mudflat layers enriched in labile microbial and algal-derived compounds, whereas deeper salt-marsh sediments retained more resistant, C3 plant-derived signatures resembling those of terrestrial OM source. OM degradation rates were closely linked to source composition and depth. Degradation was concentrated within the top 5 cm of salt-marsh and the top 10 cm of mudflat. Below these depths, biomarker profiles changed minimally, delineating a transition to longer-term preservation. First-order degradation constants were three times higher in mudflat (0.53 yr−1) than in salt-marsh (0.17 yr−1), despite similarly high sedimentation rates (1.8 and 2.2 cm yr−1, respectively). This reflects differences in OM lability, with even minor contributions from microphytobenthos enhancing reactivity in mudflats. Salt-marshes, with their intermediate OM reactivity and high sedimentation rates, emerged as hotspots of carbon accumulation (366 gC m−2 yr −1), while mudflats also contributed substantially to coastal carbon sequestration (239 gC m−2 yr −1). These results highlight the value of depth-resolved, biomarker-based approaches to identify habitat-specific degradation dynamics; ultimately better understanding carbon accumulation in coastal ecosystems
Revealing cryptic lineages: genetic structure and conservation units in the Eurasian Stone-curlew (Burhinus oedicnemus)
No full textInternational audienc
The Southern Ocean’s bioluminescent environment revealed by deep-diving Southern elephant seals
No full textInternational audienceAs a consequence of climate change, marine communities and ecosystems are shifting at all scales, but characterizing these shifts and their extent remains challenging. Notably lacking is information on mid-trophic level components such as zooplankton and micronekton. As ca. 70% of water column organisms are bioluminescent, we hypothesized that measuring changes in deep-water bioluminescence signals can provide valuable additional information for further characterizing the mid-trophic level composition of mesopelagic habitats. To test this, southern elephant seals Mirounga leonina from the Kerguelen Islands and from Peninsula Valdés in Argentina were equipped with light and motion sensors. Since these predators are large consumers of mesopelagic fauna, we expected them to target high-biomass regions and thus efficiently sample the bioluminescence fields in the Indian sector of the Southern Ocean and the southwestern sector of the Atlantic. Combining light, depth, GPS and motion data, we found that (1) bioluminescence events took place at depths ranging from 100 to 700 m, occurring deeper in Peninsula Valdés than in the Kerguelen Islands and in shallower waters during the night than during the day, suggesting organism diurnal migrations; and that (2) bioluminescence events were more intense, longer, deeper and more likely to be composed of multiple peaks when associated with a prey-capture attempt. We show that bioluminescence is a good proxy to describe marine mid-trophic levels. Combining it with other biologging sensors such as micro-sonars would yield extremely valuable data on these organisms and could be a key component in improving our understanding of mesopelagic habitats
CCOR: A DRL-Based Framework for Efficient Data Replication in Computing Continuum Architectures
No full textSmart cities generate massive amounts of heterogeneous data to support critical urban services such as transportation, healthcare, and energy. Meeting the strict Service Level Objectives (SLOs) of these services requires efficient data management across the Computing Continuum. Replication, beyond its role in fault tolerance, plays a crucial role in reducing data access latency through locality-aware placement. However, deciding where to place replicas across the Computing Continuum layers requires balancing conflicting objectives, including latency, cost, energy consumption, and availability. In this paper, we present CCOR (Coordinated Continuum Optimized Replication), a distributed framework for adaptive and multi-objective data replication management using Deep Reinforcement Learning (DRL). Unlike prior approaches that fail to adapt to heterogeneous SLOs as well as dynamic infrastructures, CCOR learns placement policies that jointly optimize data availability, storage cost, and energy consumption while ensuring applications latency SLO compliance. Simulations based on the Bordeaux metropolitan area demonstrate that CCOR achieves near-optimal availability while using up to 70% fewer replicas, reducing energy and cost by 10% and 5%, respectively, compared to specialized baselines
In-situ marine corrosion experiments during extended immersion: Metallic coatings deposited onto mild steel
No full textInternational audienc
Off-target: Herbicides in cereal fields favour competitive weeds over non-target species
No full textInternational audienceWeed diversity plays an important role in maintaining resilient agroecosystems, yet agricultural practices, such as pesticide applications, significantly shape weed communities. Previous studies have primarily focused on comparing organic and conventional farming, with a particular emphasis on land-use intensification. However, less attention has been given to the specific effects of herbicides, insecticides, and fungicides on weed communities, including the species affected and the mechanisms driving these changes, based on farmer-recorded application data. Our study examined the impact of herbicide, fungicide, and insecticide useboth in terms of Treatment Frequency Index (TFI), i.e., intensity of application, and quantity (QA) -on weed communities in 96 non-organic cereal fields over 4 years. We did not detect any effect of insecticide metrics on weed communities. Interestingly, our results indicate that TFI of fungicide decreased weed abundance, whereas QA did not. In contrast, for herbicides, QA had a stronger negative impact on weed communities than TFI. This suggests that TFI of fungicide, which may reflect low-dose but frequent applications, could exert indirect and long-term effects on weed suppression. In comparison, QA of herbicide more directly reflects the toxic load delivered to weeds and is therefore a better predictor of their suppression. Our results reveal that herbicides are the main factor shaping weed communities by decreasing the abundance of non-target/non-competitive species while replacing them with competitive species. These findings point out the potential ineffectiveness of herbicide application on problematic weeds and emphasise the need to reconsider the use of herbicides to maintain weed diversity in agricultural landscapes
FAQ - Données de la recherche, par le Méta-Atelier de la Donnée en Nouvelle-Aquitaine
No full textCe document propose une foire aux questions (FAQ) de base pour la communication sur les données de la recherche. Réalisé dans le cadre du Méta-Atelier de la Donnée en Nouvelle-Aquitaine, il aborde des aspects génériques liés aux données de la recherche, indépendamment des disciplines ou des structures
