HAL Collection UNC (Univ. de la Nouvelle Calédonie)
Not a member yet
7579 research outputs found
Sort by
Pooling de Graph Neural Networks : une approche dense mais adaptative
No full textInternational audienceGraph Neural Networks can process complex, massive and unstructured data which form large graphs. To process these complex data, several pooling methods that reduce their size while preserving essential information have been proposed. Current pooling methods either group nodes into a fixed number of nodes or remove an adaptive number of nodes. In this article, we propose SpaPool, an adaptive pooling method that groups nodes. This method demonstrates competitive performance compared to existing techniques and excels particularly on small-scale graphs.Les réseaux de neurones en graphes (GNN) traitent des données complexes, massives et non structurées formant des graphes volumineux. Différentes méthodes de pooling ont été proposées, permettant de réduire ces graphes, tout en préservant les informations essentielles, soit en regroupant de manière fixe des noeuds, soit en supprimant un nombre adaptatif de noeuds. Cet article propose SpaPool, une méthode adaptative regroupant les noeuds d'un graphe. Les résultats de celle-ci se sont avérés équivalents aux méthodes existantes, voire supérieurs pour les petits graphes
Mesophotic protected habitats as refugia for the most at-risk elasmobranch species
No full textInternational audienceCoastal areas experience among the highest cumulated human pressures across the world's oceans with severe consequences on some fish populations which become scarce and at risk of local extirpation. Marine protected areas (MPAs) are essential for conserving these threatened populations, but they often cover only the most superficial waters and are not three-dimensionally extensive enough, leaving most mobile species with a large home range like elasmobranchs vulnerable to human pressures. Here we explored the role of mesophotic and deep-water habitats as refuges for elasmobranchs and the most at-risk species in the Parc Naturel Marin du Cap Corse et de l'Agriate, a large MPA in the Mediterranean Sea. Using the metabarcoding of 200 environmental DNA samples of seawater filtered between 1 and 835 m depth we revealed that fish species diversity was the highest in the mesophotic habitat (30-149 m). We also found the highest number of endangered species, including Evolutionary Distinct and Globally Endangered (EDGE) species such as Rostroraja alba and Squatina squatina, in this key habitat. Species composition varied significantly with depth, revealing a stratification linked to human pressure and environmental conditions. Protecting both mesophotic and deeper habitats (below 150 m) thus appears crucial to preserving biodiversity, species with unique functional roles but also those with unique phylogenetic diversity, in order to ensure ecosystem resilience in the face of climate change and overexploitation. The full protection of deep-water refuges within the western Mediterranean basin is essential to ensure the preservation of the most at-risk species, with the possibility of rewilding by connectivity other depleted adjacents habitats
Indigenous knowledges, practices and resilience in Pacific university students before and during COVID-19 pandemic: protocol for an interdisciplinary and mixed methods
No full textInternational audienceThese last decades, Pacific islands undergo major climatic transformations and a rapid socio-economic transition that impacted indigenous knowledges and practices in populations across generations. It is also important to identify actions and key points that can be used by the population to be resilient and to be adapted to the world of tomorrow. The COVID-19 pandemic has both reinforced vulnerabilities already present in the Pacific Islands but also revealed a resilience of the population through indigenous knowledges and practices to face this unprecedented situation that has affected the world's population. So, Indigenous Knowledges, Practices and Resilience in the Pacific Islands (SPAR-Pacific) project is an interdisciplinary research initiative that focuses on the indigenous knowledges and practices of Pacific Island populations in a context of high resilience (climate change, socio-economic transition and COVID- 19 pandemic). The objectives of this project are to: 1- ensure the existence and activation of indigenous knowledges and practices among young university students from New Caledonia, Vanuatu, Fiji, Solomon Islands and Papua New Guinea; 2- assess the perceptions associated with these knowledges and practices, particularly in relation to intangible aspects among young people; 3- understand dynamics of development, expression, and transmission of indigenous knowledges after the COVID-19 pandemic in connection with the socio- economic and natural environment among these young people
Utilisation de la télédétection pour la mise à jour des cartes des zones humides méditerranéennes et potentiel pour le contrôle de la nuisance des moustiques
No full textOnline ISSN: 2054-930XInternational audienceThis study aimed at evaluating the potential of remote sensing technologies for updating Mediterranean wetlands maps from Marseille to the Franco-Spanish border. Previously, these maps were manually produced through in situ surveys. The maps, presenting a typology of 16 wetland classes, can be used for natural area managers to monitor Mediterranean ecosystems as well as to mosquito operators to identify mosquito breeding areas, thus reducing prospecting time and providing guidance on areas requiring treatment. The project developed a supervised classification methodology using Orfeo ToolBox and the Random Forest algorithm applied to Sentinel-2 optical imagery (10m-20m resolution). Training data were derived from updated existing databases. The process involved two stages: a Level 1 classification (e.g., saline, brackish, freshwater wetland, woodland, urban) followed by a Level 2 classification detailing 16 natural habitat classes. The classification achieved a Kappa Index of 0.9, an overall accuracy of 0.913, and an F1-score of 0.87, demonstrating its reliability compared to traditional methods. The resulting maps offers improved details, harmonisation, and precise visualisation of areas requiring monitoring with reduced fieldwork. This study validates remote sensing as a valuable tool to support EID Méditerranée's expertise, optimises resources, and improves the cartographic updates of managed wetlands. Further research is required to integrate additional sensors and environmental variables, validate results, and automate processes for sustainable long-term application. Additionally, future work will be done to refine natural habitats definition to better identify ecological habitats directly linked to mosquito breeding areas in Mediterranean wetlands.✓ Remote sensing (supervised classification), enable us to obtain a harmonised map of wetlands over the entire EID Méditerranée action zone, and to produce a test map of ecological levels that is close to reality on the ground. ✓ Field surveys still have a decisive role to play in obtaining reliable results. ✓ Further research for updating swamp mosquitoes larval habitat map.Cette étude vise à évaluer le potentiel des technologies de télédétection pour la mise à jour des cartes des zones humides méditerranéennes de Marseille à la frontière franco-espagnole. Auparavant, ces cartes étaient produites manuellement à partir de relevés in situ. Les cartes, présentent une typologie de 16 classes de zones humides, et peuvent être utilisées par les gestionnaires d'espaces naturels pour surveiller les écosystèmes méditerranéens ainsi que par les opérateurs de démoustication pour identifier les zones de reproduction des moustiques, réduisant ainsi le temps de prospection et fournissant des conseils sur les zones nécessitant un traitement. Le projet a développé une méthodologie de classification supervisée utilisant Orfeo ToolBox et l'algorithme Random Forest appliqué à l'imagerie optique Sentinel-2 (résolution 10m-20m). Les données d'entraînement ont été dérivées de bases de données existantes mises à jour. Le processus s'est déroulé en deux étapes : une classification de niveau 1 (par exemple, zone humide saline, saumâtre ou d'eau douce, zone boisée, zone urbaine) suivie d'une classification de niveau 2 détaillant 16 classes d'habitats naturels. La classification a atteint un indice Kappa de 0,9, une précision globale de 0,913 et un score F1 de 0,87, ce qui démontre sa fiabilité par rapport aux méthodes traditionnelles. Les cartes qui en résultent offrent de meilleurs détails, une meilleure harmonisation et une visualisation précise des zones nécessitant une surveillance avec un travail de terrain réduit.Cette étude valide la télédétection comme un outil précieux pour soutenir l'expertise de l'EID Méditerranée, optimiser les ressources et améliorer les mises à jour cartographiques des zones humides gérées. D'autres recherches sont nécessaires pour intégrer d'autres capteurs et variables environnementales, valider les résultats et automatiser les processus pour une application durable à long terme. De plus, des travaux futurs permettront d'affiner la définition des habitats naturels afin de mieux identifier les habitats écologiques directement liés aux zones de reproduction des moustiques dans les zones humides méditerranéennes.✓ La télédétection (classification supervisée), permet d'obtenir une carte harmonisée des zones humides sur l'ensemble de la zone d'action de l'EID Méditerranée, et de réaliser une carte test des niveaux écologiques proche de la réalité du terrain. ✓ Les prospections de terrain ont encore un rôle déterminant à jouer pour obtenir des résultats fiables. ✓ Poursuite des recherches pour la mise à jour de la carte des habitats larvaires des moustiques des marais.</p
"Faire de la recherche pour la recherche" : la linguistique est-elle utile aux locuteurs de langues océaniennes?
No full textInternational audienceDuring the drafting of a research project in descriptive linguistics, based on international collaboration and aimed at comparing possessive classification systems in some indigenous languages of Vanuatu and New Caledonia, we were criticised by a public partner from New Caledonia who had been asked to join the project for wanting to ‘do research for research's sake’ (email exchanges in July 2017). Seven years later, we are organising a presentation of this research project to one of the language communities with which we worked. Three educational tools were developed with the aim of highlighting the richness of the field of possession grammar in these languages, via pedagogical, free and multilingual resources. These three products were ultimately co-produced and co-signed in partnership with the same institution which, at the outset of the project, did not see any objective in line with its aspirations, nor any social utility for the speakers.As this research project draws to a close and the speakers take ownership of these valorisation productions, we would like to take the opportunity of this paper to collectively question and reflect on the social role of linguistics, and by extension the responsibilities of linguists (Good, 2018). In a twenty-first-century Oceania grappling with complex and multiple issues linked to upheavals in lifestyles and migrations, the erosion of biocultural diversity, identity and cultural claims, and decolonial demands, do linguists necessarily have to choose between theoretical linguistics and applied, or even involved, linguistics (see Burrows & Pradeau, 2022)? since it is still sometimes necessary to point out that the discipline of linguistics (whether descriptive, theoretical or applied) is ‘a science whose aim is not the abduction of language by experimenters located outside the communities concerned’ (de Pury, 2000: 41), how can we (re)instil confidence in social and institutional partners and demonstrate the positive benefits of research projects with a primarily descriptive, theoretical or typological aim (see Moyse-Faurie, 2012)? What types or forms of “rendering to the community” (Grinevald, 2010: 59) should be favoured to ensure that all the stakeholders in such projects find meaning in them? From the point of view of the various parties involved, what ultimately defines a useful linguistic project?Oceanic linguistics, because it focuses in particular on indigenous languages, many of which are in a precarious state of vitality or facing unfavourable sociolinguistic situations to the benefit of the dominant Indo-European languages, has a long tradition of taking account of these crucial issues relating to research ethics and the identification of good practice in “fieldwork” in the Pacific (see, among others, Bowern, 2008; Guérin & Lacrampe, 2010; Thieberger & Musgrave, 2007). These questions are all the more topical in the general context of UNESCO's International Decade of the World's Indigenous Languages (2022-2032), the slogan of which is “nothing about us without us”. This concern for positive collaboration between speakers and social, academic and institutional players is also at the heart of the aspirations arising from the specific context of ‘New Caledonia after the riots of May 2024’. Between global urgency and local common sense, this paper aims to open a space for reflection and exchange on the ethical values and deontological responsibilities on which oceanic linguists can agree today.Bowern, C. (2008). Linguistic Fieldwork: A Practical Guide. Palgrave Macmillan.Burrows, A., & Pradeau, C. (Eds.). (2022). La linguistique appliquée est-elle une discipline de combat ? / Etudes de linguistique appliquée (Vol. 205).de Pury, S. (2000). Entre description et intervention. Mémoires de La Société de Linguistique de Paris / Les Langues En Danger, VIII, 33–41.Good, J. (2018). Ethics in language documentation and revitalization. In K. L. Rehg & L. Campbell (Eds.), The Oxford handbook of endangered languages (pp. 417–440). Oxford University Press.Grinevald, C. (2010). Quarante ans de perspective sur deux langues en danger : le Jakaltek Popti du Guatemala et le rama du Nicaragua. Faits de Langues / Linguistique de Terrain Sur Langues En Danger: Locuteurs et Linguistes, 35–36, 39–78.Guérin, V., & Lacrampe, S. (2010). Trust me, I am a linguist ! Building partnership in the field. Language Documentation and Conservation, 4, 22–33.Moyse-Faurie, C. (2012). Documentation d’une langue ultra-minoritaire: contextes politique et social, réalisation et difficultés rencontrées. In DGLFLF (Ed.), Langues de France, langues en danger: aménagement et rôle des linguistes / Cahiers de l’Observatoire des pratiques linguistiques, n°3 (pp. 147–152).Thieberger, N., & Musgrave, S. (2007). Documentary linguistics and ethical issues. In P. K. Austin (Ed.), Language Documentation and Description (Vol. 4, Issue 4, pp. 26–37). SOAS.Lors de la rédaction d’un projet de recherche en linguistique descriptive, construit à partir d’une collaboration internationale et ayant pour objet la comparaison des systèmes de classification possessive dans des langues autochtones du Vanuatu et de Nouvelle-Calédonie, il nous avait été reproché de la part d’un partenaire publique de la Nouvelle-Calédonie alors sollicité pour s’associer au projet de vouloir « faire de la recherche pour la recherche » (échanges de courriels de juillet 2017). Sept ans plus tard, nous organisons la restitution de ce projet de recherche auprès de l’une des communautés linguistiques avec qui nous avons collaboré. Trois outils pédagogiques ont été développé avec pour ambition de valoriser en particulier la richesse du domaine de la grammaire de la possession dans ces langues, via des ressources ludiques, gratuites et multilingues. Ces trois productions ont finalement été co-réalisées et co-signées en partenariat avec cette même institution qui, à l’initiale du projet, ne lui envisageait pas d’objectif concordant à ses aspirations ni d’utilité sociale au bénéfice des locuteurs.Alors que ce projet de recherche se clôture et que les locuteurs s’approprient ces productions de valorisation, nous voudrions saisir l’opportunité de cette communication pour interroger et réfléchir collectivement au rôle social de la linguistique, et par extension aux responsabilités des linguistes (Good, 2018). Dans une Océanie du XXIème siècle aux prises avec des problématiques complexes et multiples liées aux bouleversements des modes de vie et migrations, à l’érosion des diversités bioculturelles, aux revendications identitaires et culturelles, aux réclamations décoloniales, les linguistes doivent-ils nécessairement choisir entre une linguistique théorique et une linguistique appliquée, voire impliquée (voir Burrows & Pradeau, 2022) ? puisqu’il est parfois encore nécessaire de rappeler que la discipline de la linguistique (qu’elle soit descriptive, théorique ou appliquée) est « une science qui n’a pas pour but le rapt de la langue par des expérimentateurs situés à l’extérieur des communautés concernées » (de Pury, 2000 : 41), comment (re)donner confiance aux partenaires sociaux et institutionnels et démontrer les bénéfices positifs que constituent des projets de recherche à visée premièrement descriptive, théorique ou typologique (voir Moyse-Faurie, 2012) ? quels types ou formes de « rendu à la communauté » (Grinevald, 2010 : 59) privilégier pour que tous les acteurs de tels projets y trouvent du sens ? faire de la recherche « pour la recherche » est-il foncièrement incompatible avec faire de la recherche « pour la société » ? du point de vue des différentes parties impliquées, qu’est-ce qui définit, finalement, un projet linguistique utile ?La linguistique océanienne, parce qu’elle se consacre en particulier à des langues autochtones pour beaucoup en situation de vitalité précaire ou faisant face à des situations sociolinguistiques défavorables au profit de langues indo-européennes dominantes, a une longue tradition de prise en compte de ces enjeux cruciaux liés à l’éthique de la recherche et à l’identification de bonnes pratiques du « travail de terrain » dans le Pacifique (voir, entre autres, Bowern, 2008; Guérin & Lacrampe, 2010; Thieberger & Musgrave, 2007). Ces questionnements s’imposent comme d’autant plus d’actualité dans le contexte général de la Décennie Internationale des Langues Autochtones décrétée par l’UNESCO (2022-2032) dont le slogan est « nothing about us without us ». Ce souci de faire collaborer positivement les locuteurs, les acteurs sociaux, académiques et institutionnels est également au cœur des aspirations qui résultent du contexte spécifique de la « Nouvelle-Calédonie d’après les émeutes de mai 2024 ». Entre urgence globale et bon sens local, cette communication entend ouvrir un espace de réflexion et d’échanges sur les valeurs éthiques et les responsabilités déontologiques sur lesquelles peuvent s’accorder, aujourd’hui, les linguistes océanistes.Bowern, C. (2008). Linguistic Fieldwork: A Practical Guide. Palgrave Macmillan.Burrows, A., & Pradeau, C. (Eds.). (2022). La linguistique appliquée est-elle une discipline de combat ? / Etudes de linguistique appliquée (Vol. 205).de Pury, S. (2000). Entre description et intervention. Mémoires de La Société de Linguistique de Paris / Les Langues En Danger, VIII, 33–41.Good, J. (2018). Ethics in language documentation and revitalization. In K. L. Rehg & L. Campbell (Eds.), The Oxford handbook of endangered languages (pp. 417–440). Oxford University Press.Grinevald, C. (2010). Quarante ans de perspective sur deux langues en danger : le Jakaltek Popti du Guatemala et le rama du Nicaragua. Faits de Langues / Linguistique de Terrain Sur Langues En Danger: Locuteurs et Linguistes, 35–36, 39–78.Guérin, V., & Lacrampe, S. (2010). Trust me, I am a linguist ! Building partnership in the field. Language Documentation and Conservation, 4, 22–33.Moyse-Faurie, C. (2012). Documentation d’une langue ultra-minoritaire: contextes politique et social, réalisation et difficultés rencontrées. In DGLFLF (Ed.), Langues de France, langues en danger: aménagement et rôle des linguistes / Cahiers de l’Observatoire des pratiques linguistiques, n°3 (pp. 147–152).Thieberger, N., & Musgrave, S. (2007). Documentary linguistics and ethical issues. In P. K. Austin (Ed.), Language Documentation and Description (Vol. 4, Issue 4, pp. 26–37). SOAS
Note de synthèse de la thèse doctorale « Repenser la justice transitionnelle à partir du cas colombien : le vivant comme victime des conflits armés »
No full textInternational audienceThis article is a summary of a doctoral thesis, Rethinking Transitional Justice from the Colombian Case: The Living as a Victim of Armed Conflicts, defended on Junuary 24, 2024 at Limoges’ University. Its objective is to highlight the main aspects to doctoral research on recognizing non-human living entities as victims of armed conflicts, using the Colombian cas as a study example.Cet article est présenté en forme de note de synthèse d’une thèse doctorale Repenser la justice transitionnelle à partir du cas colombien : le vivant comme victime des conflits armés, soutenue le 24 janvier 2024 à l’Université de Limoges. Il a pour objectif de présenter les principaux aspects des recherches doctorales sur la reconnaissance du vivant non-humain comme victime des conflits armés en prenant l’exemple colombien comme étude de cas
The Glacial Paleolandscapes of Southern Africa: the Legacy of the Late Paleozoic Ice Age
No full textInternational audienceThe modern relief of Southern Africa is characterised by stepped plateaus bordered by escarpments. This morphology is thought to result from stepwise uplift and ensuing continental-scale erosion of the region as it rode over Africa’s mantle ‘superplume’ following the break-up of Gondwana, i.e. since the mid-Mesozoic. We demonstrate in this contribution that this modern morphology of Southern Africa is in fact largely inherited from glacial erosion associated to the Late Paleozoic Ice Age (LPIA) that occurred between 370 and 260 Myr ago, during which Gondwana – which included Southern Africa – was covered in thick ice masses. Southern Africa hosts vast (up to 106 km²) and thick (up to 5 km) sedimentary basins ranging from the Carboniferous, represented by glaciogenic sediments tied to the LPIA, to the Jurassic-Cretaceous. These basins are separated by intervening regions largely underlain by Archean to Paleoproterozoic cratonic areas that correspond to paleohighlands that preserve much of the morphology that existed when sedimentary basins formed, and particularly glacial landforms. In this contribution, we review published field and remote data and provide new large-scale interpretation of the geomorphology of these paleohighlands of Southern Africa. Our foremost finding is that over Southern Africa, vast surfaces, tens to hundreds of thousands km² (71.000–360.000 km²) are exhumed glacial landscapes tied to the LPIA. These glacial landscapes manifest in the form of cm-scale striated pavements, m-scale fields of roches moutonnées, whalebacks and crag-and-tails, narrow gorges cut into high-standing mountain ranges, and km-scale planation surfaces and large U-shaped valleys, overdeepenings, fjords and troughs up to 200 km in length. Many modern savannahs and desertic landscapes of Southern Africa are therefore relict glacial landscapes and relief ca. 300 Myr old. These exhumed glacial relief moreover exerts a strong control on the modern-day aspect of the geomorphology of Southern Africa as (1) some escarpments that delineate high-standing plateaux from valleys and coastal plains are inherited glacial relief in which glacial valleys are carved, (2) some hill or mountain ranges already existed by LPIA times and were likely modelled by glacial erosion, and (3) the drainage network of many of the main rivers of Southern Africa is funnelled through ancient glacial valleys. This remarkable preservation allowed us to reconstruct the paleogeography of Southern Africa in the aftermath of the LPIA, consisting of highlands over which ice masses nucleated and from which they flowed through the escarpments and toward lowlands that now correspond to sedimentary basins. Our findings therefore indicate that glacial landforms and relief of continental-scale can survive over tens to hundreds of million years. This preservation and modern exposure of the glacial paleolandscapes were achieved through burial under piles of Karoo sediments and lavas over ca. 120 to 170 million years and a subsequent exhumation since the middle Mesozoic owing to the uplift of Southern Africa. Owing to strong erodibility contrasts between resistant Precambrian bedrock and softer sedimentary infill, the glacial landscapes have been exhumed and rejuvenated. We therefore emphasise the need of considering the legacy of glacial erosion processes and the resulting presence of glacial landscapes when assessing the post-Gondwana-breakup evolution of Southern African topography and its resulting modern-day aspect, as well as inferences about climate changes and tectonic processes. Finally, we explore the potential pre-LPIA origin for some of the landscapes. In the Kaoko region of northern Namibia, the escarpments into which glacial valleys are carved may correspond to a reminiscence of the Kaoko Pan-African Belt, whose crustal structures were either reactivated or where relief persisted since then. In South Africa, the escarpment bordering the paleohighland corresponds to crustal-scale faults that might have been reactivated during LPIA by subsidence processes. These inherited morphological or crustal features may have been re-exploited and enhanced by glacial erosion during the LPIA, as it is the case for some Quaternary glacial morphology
Socio-economic impacts and responses of the fishing industry and fishery managers to changes in small pelagic fish distribution and abundance
No full textInternational audienceSmall Pelagic Fish (SPF) exhibit large fluctuations in abundance and distribution in response to environmental variability. To maintain the resilience of fishing communities and develop effective and equitable climate adaptation strategies, improved understanding of how the fishing industry responds to spatio-temporal shifts within and across SPF populations is of critical importance. In this paper, we examine the responses of the fishing industry and resource managers to shifts in SPF availability worldwide and identify the resulting socioeconomic impacts. Leveraging SPF case studies from around the globe, we synthesize and compare the social-ecological linkages and feedbacks mediating how SPF fisheries respond to changes in marine ecosystem structure and function associated with (1) spatial shifts in species distribution and habitat availability, (2) ‘boom and bust’ population dynamics, or (3) changes in fish size and quality. Our case studies illustrate multiple paths towards the resilience of small pelagic fisheries and the fishing industry dependent upon them while emphasizing the need for increased coordination and cooperation across sectors and scales as climate change progresses. Drawing from the lessons offered by historical responses, as environmental variability increases, efforts to increase the flexibility and dynamism of SPF harvest portfolios and management strategies, licensing regimes, and international catch and allocation agreements may be required to ensure resource sustainability and human well-being
Local knowledge and farmers’ capacity of experimentation as a lever of adaptation to climate change. How ethnography can be articulate within an action research program (CLIPSSA)?: How ethnography can be articulate within an action research program (CLIPSSA)?
No full textInternational audienceThe islands of the South Pacific are among the most exposed territories to climate change, but they are also areas of experimentation at the forefront of the fight against this phenomenon. These territories have resources embedded to their geographical, socio-cultural and ecological diversity, which can feed local, regional and national strategies for adapting to climate change.We'll be looking at localized resources such as local knowledge linked to agricultural practices: how can this knowledge be a resource for coping with climate change? How are they built, adapted, transmitted and how do they evolve? These questions are at the heart of the CLIPSSA project (Pacific Climate, Local Knowledge and Adaptation Strategies), deployed over 5 years (2022-2026) in four pacific territories: New Caledonia, French Polynesia, Vanuatu and Wallis and Futuna.More specifically, we will look at empiricism as a fundamental resource for adapting local knowledge and farming practices to cope with climate change in French Polynesia. There are many learning channels (family, professional, specialized groups, trainings, IT tools, etc.) mobilized in different ways by farmers (Richez, 2024). Among them, empiricism, a method that relies on concrete experience, has been identified as a method for testing new practices, adjusting them, or even developing them within the farm. In some cases, new practices implemented through empiricism have even been disseminated, through informal observation and transmission, to other farmers.From qualitative surveys conducted with farmers on the Tahiti peninsula and in Moorea, we will examine the following questions: to what extent is empiricism a resource mobilized by farmers? Is it widespread, or confined to a few expert figures? Does it facilitate the hybridization of knowledge, specifically the hybridation of scientific knowledge with local knowledge systems? And at least, how can they be supported by public policies to help them adapt to climate change
