HAL Collection UNC (Univ. de la Nouvelle Calédonie)
Not a member yet
7579 research outputs found
Sort by
Limits of equilibrium states for coupled weakly interacting systems. Application to the measure of maximal entropy
No full textWe study metastability for symbolic dynamic. We prove that for a global system given by two independent sub-systems linked by a hole, and for a Lipschitz continuous potential, the global equilibrium state converges, as the hole shrinks, to a convex combination of the two independent equilibria in each component. Two kinds of convergence occur, depending on the assumptions on how long an orbit has to stay in each well.As a by-product, we show that this can be applied to a geometrical system inspired from [11] and for the measure of maximal entropy.</div
Morphological traits and machine learning for genetic lineage prediction of two reef-building corals
No full textInternational audienceIntegrating multiple lines of evidence that support molecular taxonomy analysis has proven to be a robust method for species delimitation in scleractinian corals. However, morphology often conflicts with genetic approaches due to high phenotypic plasticity and convergence. Understanding morphological variation among species is crucial to studying coral distribution, life history, ecology, and evolution. Here, we present an application of Random Forest models for coral species identification based on morphological annotation of the corallum and corallites. We show that the integration of molecular and morphological trait analysis can be improved using machine learning. Morphological traits were documented for Porites and Pocillopora coral species that were collected and genotyped through genome-wide, genetical hierarchical clustering, and coalescence analyses for the Tara Pacific Expedition. While Porites only included three tentative species, most Pocillopora species were accounted by included specimens from the western Indian Ocean, tropical Southwestern Pacific, and southeast Polynesia. Two Random Forest models per genus were trained on the morphological annotations using the genetic lineage labels. One model was developed for in-situ image identification and used corallum traits measured from in-situ photographs. Another model for integrative species identification combined corallum and corallite data measured on scanning electron micrographs. Random Forest models outperformed traditional dimension reduction methods like PCA and FAMD followed by k-means and hierarchical clustering by classifying the correct genetic lineage despite morphological clusters overlapping. This machine learning approach is reproducible, cost-effective, and accessible, reducing the need for taxonomic expertise. It can complement molecular and phylogenetic studies and support image identification, highlighting its potential to advance a coral integrative taxonomy workflow
The nature of soil blue carbon varies across mangrove geomorphic settings
No full textInternational audienceMangrove forests store significant amounts of blue carbon, mainly as soil organic matter. Insights into mangrove organic matter are limited, despite its importance for understanding blue carbon accumulation and vulnerability to global change. Here, we quantified soil organic matter preserved through chemical recalcitrance and association with the mineral phase, as key factors influencing blue carbon persistence. We found that the nature of the soil organic matter varied with mangrove geomorphic settings. Delta settings were dominated by presumably the most persistent soil organic matter associated with minerals, while open coast karstic settings contained mostly particulate soil organic matter, likely preserved due to low microbial activity. Across mangrove soil depths, there was little difference in soil organic matter pools. The soil organic matter pool across mangroves’ geomorphic settings exhibited greater variation than that observed across all terrestrial biomes. These findings underscore the need to tailor mangrove conservation and restoration to geomorphic settings
The emergence of grammatical gender: an experimental study on the loss of informative classification
No full textInternational audienceSystems of grammatical gender are widespread, yet evidence on how these systems emerge is surprisingly scarce. We report on novel research using storyboard experiments in six Oceanic languages of Vanuatu and New Caledonia, to probe the initial stages in the emergence of grammatical gender from possessive classifiers. This would appear to be an unlikely source, yet we can demonstrate the development and investigate it through current experiments. Our storyboard experiments were designed to reveal the use of classifiers as reference tracking devices across units of discourse and in different interactional contexts. The particular interest is in the degree to which the use of these classifiers becomes fixed and hence redundant for given nouns. The general interest is the investigation of the way in which apparently useful distinctions are lost, as transparent classification becomes opaque
De la perception aux strategies d’adaptation : pratiques de gestion locale du changement climatique en Nouvelle-Calédonie et Vanuatu
No full textInternational audienceAt the 92nd ACFAS conference, which took place from 5–9 May 2025 in Canada and was themed “Cross-Disciplinary Research Perspectives on Responsible Management Practices”, the CLIPSSA project (South Pacific Climate, Local Knowledge and Adaptation Strategies) showcased Pacific Island agriculture. In a presentation entitled 'From perception to adaptation strategies: local climate change management practices in New Caledonia and Vanuatu', Samson Jean Marie, a doctoral student in anthropology and geography working on the CLIPSSA project, presented the initial results of field surveys conducted as part of the project. He explained how farmers in these territories, who are at the forefront of climate change, are adapting their practices to cope with it.À l’occasion du 92e colloque de l’ACFAS, qui s’est tenu du 5 au 9 mai 2025 au Canada sur le thème « perspectives de recherche croisées sur les pratiques de gestion responsable », le projet CLIPSSA (Climat du Pacifique Sud, Savoirs Locaux et Stratégies d’Adaptation) a porté la voix des agricultures insulaires du Pacifique. Dans une communication intitulée « De la perception aux stratégies d’adaptation : pratiques de gestion locale du changement climatique en Nouvelle-Calédonie et au Vanuatu », Samson Jean Marie, doctorant en anthropologie et géographie au sein de ce projet, a présenté les premiers résultats de ces enquêtes de terrain. Il a exposé comment les agriculteurs de ces territoires, en première ligne face aux dérèglements climatiques, adaptent leurs pratiques pour y faire face
Vers une approche hybride combinant Intégration de données et IA pour la reconstitution d'un réseau d'adduction d'eau potable
No full textInternational audienceMonitoring the drinking water distribution network to anticipate unserved areas during periods of drought and potentially identify locations of leaks or unauthorized connections is a major challenge, especially in the Haute Matsiatra region (Madagascar), where a number of measures aimed at rationalising water distribution are decided on aseasonal basis. Because the water supply network has existed for a very long time and is constantly being updated, its description is often incomplete and sometimes erroneous.In order to monitor water distribution (one of the challenges of the MADATLAS project co-led by ESPACE-DEV), the first objective is to rebuild the graph of the distribution network from the existing erroneous and incomplete data. In this context, we are looking at various approaches that will enable us to verify the location of the actual network.Permettre le suivi de la distribution d'eau potable afin d'identifier les localisations de fuites, de branchements sauvages sur le réseau d'adduction est un enjeu majeur, spécialement dans la région Haute Matsiatra (Madagascar) où plusieurs mesures visant à la rationalisation de la distribution de l'eau sont décidées de façon saisonnière. Du fait de l'existence très ancienne et de la mise à jour continue du réseau d'adduction, la description de ce dernier est souvent incomplète et parfois erronée. Afin de permettre le suivi de la distribution d'eau (un des enjeux du projet MADATLAS co-porté par l'unité ESPACE-DEV), le premier objectif est de pouvoir reconstruire le graphe du réseau de distribution à partir des données parfois imprécises et incomplètes existantes. Dans ce contexte, nous envisageons différentes pistes qui nous permettront de vérifier la localisation du réseau réel. Mots-clésGestion de la distribution d'eau potable, intégration de données hétérogènes, traitement des incertitudes.</div
Atlas des agricultures du monde et des systèmes alimentaires face au changement climatique
No full textInternational audienc
Draw me the sea
No full textDrawings can be used to identify and understand children's ecological knowledge, as well as their representations and emotions, and as a medium for dialogue between science and society