Le gouvernement français a fixé pour objectif que 23% de la consommation énergétique proviendraitdes énergies renouvelables, y compris les énergies renouvelables marines pour 2020. Huit parcs éoliensen mer sont prévus le long des côtes françaises Atlantique-Manche, parmi lesquels le site de Dieppe-LeTréport dans la partie orientale de la Manche. Afin de mieux connaître la structure et le fonctionnementde cet écosystème avant l'installation du parc éolien, ma thèse est axée sur la construction de modèlesde réseau trophique à l’aide du logiciel Ecopath sur ce site d’implantation en comparaison avec deshabitats benthiques similaires en Manche. La collecte de nouvelles données sur les compartimentsbiologiques (zooplancton, suprabenthos, méiofaune, benthos et poissons démersaux) est essentielle.Ainsi, quatre campagnes (étés 2014/2015, hivers 2015/2016) ont permis d'estimer la contribution dechaque groupe zoologique dans deux habitats benthiques principaux : les graviers ensablés/sablesgraveleux et les sables moyens sur le site (environ 30 km²) correspondant à deux communautésbenthiques largement répandues en Manche. La macrofaune a été échantillonnée avec une benne VanVeen, le suprabenthos avec un traîneau Macer-Giroq modifié avec une station dans chaque habitatbenthique (échantillonnage jour et nuit) et les poissons avec un chalut à perche. Au total, 1 584 poissonsdémersaux ont été recueillis et leurs contenus stomacaux ont été examinés. Les résultats montrent quele site se distingue par la forte contribution des bivalves en termes de biomasses et que les crustacés sontdes proies préférentielles des poissons. Ce lien est discuté à travers l’analyse du réseau trophique avecun focus sur les espèces consommées et non consommées de communautés benthiques, et la notion decul de sac trophique. / The French government has set a target of 23% of energetic consumption derived from renewable energysources including Marine Renewable Energy for 2020. Eight offshore wind farms (OWF) are plannedalong the Atlantic-English Channel French coasts and among them Dieppe-Le Tréport site in the easternpart of the English Channel. In order to gain further knowledge on the ecosystem structure andfunctioning before the OWF installation, my PhD thesis is focused on the construction of an EcopathModel on the wind farm site in comparison with similar benthic habitats from the English Channel. Thecollection of new data on biological compartments (zooplankton, suprabenthos, meiofauna, benthos anddemersal fishes) are essential. Four cruises (summers 2014 /2015 and winters 2015 /2016) allowed toestimate the contribution of each zoological group in two main benthic habitats: i.e. sandy gravels andmedium clean sands, founded on the site (about 30 km²), corresponding to two microbenthiccommunities well represented in the English Channel. Macrofauna was sampled with a Van Veen grab,suprabenthos with a modified Macer-Giroq sledge in one station from each benthic habitat (day andnight sampling) and fish with a beam trawl in ten stations. A total of 1,584 demersal fishes has beencollected and their stomach contents has been examined. The results show the importance of largebivalves in the benthic biomass whereas the crustaceans are preferential preys of fishes. This link isdiscussed within the trophic network between consumed and non-consumed species of benthiccommunities, and the concept trophic ‘cul de sac’.

Les marais littoraux de Charente Maritime représentent des systèmes hydrauliques sur lesquels les données existantes sur la qualité de l’eau sont limitées. Ce déficit d’information représente un frein à la bonne gestion de ces écosystèmes. Leur situation intermédiaire entre continent et océan leur confère un rôle prépondérant au sein de ces entités. Par ailleurs, les marais représentent un patrimoine écologique majeur où le suivi du compartiment aquatique est indispensable en vue d’améliorer la connaissance et d’apprécier l’évolution de ces milieux.Dans ce cadre, un réseau de suivi de la qualité des eaux des marais a été mis en place sur la base du SEQ-Eau (système d’évaluation de la qualité des cours d’eaux) en 2003 par l’UNIMA (Union des Marais de Charente-Maritime) dans le but de créer un référentiel « marais »répondant aux exigences de la DCE (Directive Cadre sur l’Eau européenne). Mais avant d’apprécier la qualité des eaux des marais, il est nécessaire de comprendre leur fonctionnement. Les objectifs de la thèse sont de trois ordres. Le premier était de dégager une typologie de nature d’eau des marais en valorisant les données issues du suivi patrimonial de l’UNIMA. Dans un deuxième temps, le suivi analytique des eaux et du fonctionnement a été approfondi (ajout de paramètre chimique, biologique et géographique) sur un nombre restreint de stations.Une typologie, basée sur une méthode statistique permettant d’appréhender à la fois les variabilités spatiale et temporelle dans le classement des stations en groupes, est composé de 7 groupes en marais doux et de 5 en marais salés. Le suivi des marais pendant prés de 1,5ans par les paramètres de la matière organique, de l’ensemble des compartiments planctoniques constituant les réseaux trophiques planctoniques et la diversité fonctionnelle du phyto- et du méso-zooplancton, ont permis de confirmer un certain nombre d’hypothèses sur le fonctionnement des marais et sur la typologie de nature d’eau, concernant notamment : (i)le rôle écologique des marais, (ii) l’influence des écosystèmes adjacents du méta-écosystème Pertuis-charentais et (iii) leur capacité de résilience.Le fonctionnement des marais salés semble contrôlé par un double mécanisme : la marée et le type d’activité anthropique dans les bassins déterminant le taux de renouvellement des eaux qui conditionne le type de réseaux trophiques en place et leur succession. Dans les marais doux, un schéma ‘classique’ de la succession des réseaux trophiques planctoniques aété dégagé. La dominance du phytoplancton et des réseaux trophiques herbivores et multivores, semblerait cohérent avec les fonctions écologiques d’épuration et d’habitat/nourricerie caractéristiques des zones humides. Ce fonctionnement classique peut être altéré par différents mécanismes : le renouvellement de l’eau ou le développement des macrophytes. D’une part, le faible renouvellement d’eau de certains marais peut conduire à l’’emballement’ du système et provoquer des phénomènes d’eutrophisation, altérant les propriétés de l’écosystème (injection d’azote atmosphérique ou efficacité écotrophique). Enfin, les macrophytes sont plus compétitrices que le phytoplancton dans les milieux peu profond tel que les canaux de marais. Leur capacité a utilisé les nutriments des sédiments permette leur développement et empêche celui du phytoplancton (limitation par la lumière). La dégradation importante de la biomasse macrophytique par les bactéries en période estivale peut engendrer de fortes hypoxies.Quelques indicateurs basés sur des paramètres simples sont proposés pour suivre l’état trophique du système (chlorophylle par classe de taille, biomasse bactérienne), ou bien de repérer un état susceptible de virer sur une eutrophisation (espèces de rotifère spécifiques des périodes pré-eutrophisation). / Charente-Maritime marshes represent hydraulic systems on which available data on thewater quality are limited, preventing a good management of these ecosystems. As transitionalzones between terrestrial and aquatic ecosystems, marshes play a major role in ‘metaecosystems’.Furthermore, marshes represent a major ecological patrimony, thus the samplingof the aquatic compartment seems essential to improve their knowledge. A water qualitymonitoring program has been in place since spring 2003, developed by UNIMA (Union desMarais de Charente-Maritime) in order to create a “marshes standard” conform to the WFDprerogatives. A good understanding of the marshes functioning appears as necessary todevelop this standard. In this context, 3 main objectives rose for this thesis. The first one wasto realize a water body typology presenting different functioning on Charente-Maritimemarshes. The second one was to improve the understanding of marshes functioning throughthe study of the spatio-temporal variation of planktonic food webs. The last one was toevaluate the impact of a storm (Xynthia) on the planktonic resilience of the marshes.The statistical approach applied here had allowed us to define 7 fresh marshes typesand 5 salt marshes types. The sampling of the organic matter, the functional diversity ofplanktons and the food webs succession had confirmed hypothesis on marsh functioningrelated to the typology, particularly: (i) their ecological role, (ii) their influence on ecosystemscomposing the meta-ecosystem of Charente, (iii) their resilience capacity.Salt marshes planktonic food web functioning is controlled by a double mechanism:the tide and the anthropogenic activities in ponds determining the water renewal rate. In freshmarshes, a ‘classical’ pattern of the food web succession had been found. The dominance ofthe phytoplanktonic community and of the herbivorous and multivorous food webs seemed tobe in concordance with the classical ecological functions of wetlands such as their purificationcapacity and their habitat and nursery function. This classical succession can be altered bytwo mechanisms: the renewal rate inside the marsh and the macrophytes development. In onehand, the highly water stagnation of some marshes types can provoke eutrophicationphenomena destabilizing the ecosystem properties. In the other hand, macrophytes are morecompetitive than phytoplankton in shallow marsh channels, because of their capacity to usenutrient of the sediment. The important degradation of the macrophytes’ biomass duringsummer period by bacteria can generate strong hypoxia phenomena.Some indicators based on simple parameters had been proposed to follow the trophicstate of these systems (three-size classes of chlorophyll, bacteria biomass).

Predadores de topo desempenham um papel importante na manutenção dos sistemas em que eles ocorrem porque influenciam diretamente a estrutura e a dinâmica de comunidades. Desta forma, este estudo descreveu alguns mecanismos de co-ocorrência espaço-temporal entre a onça-pintada (Panthera onca) e duas espécies de mesopredadores (Puma concolor e Leopardus pardalis), o seu controle top-down na comunidade de mamíferos e os requerimentos ambientais que determinam a sua distribuição no bioma Amazônia. Para as análises de co-ocorrência espaço-temporal foram utilizados o modelo de ocupação e a densidade de Kernel utilizando os registros de armadilhamento fotográfico (2008-2011). Para descrever o controle top-down foi elaborada uma rede trófica utilizando os itens alimentares consumidos pelas espécies e disponíveis na literatura (1983-2014). Para indicar quais as áreas mais adequadas foi utilizado o modelo de distribuição de espécies utilizando as localizações geográficas de ocorrência da espécie que foram compiladas de três diferentes bases de dados (2000-2013). A ocorrência das duas espécies de mesopredadores não foi diferente nos sítios com e sem a onça-pintada nas quatro temporadas de amostragem. A probabilidade de detecção da jaguatirica foi maior nos sítios com a presença da onça-pintada em apenas uma temporada de amostragem. A sobreposição no período de atividade das espécies de mesopredadores com a onça-pintada foi baixa, com cerca de 55% de sobreposição. Também ocorreu uma mudança razoável no período de atividade das espécies entre as temporadas de amostragem, com mínima de 32% e máxima de 56% de sobreposição. As maiores diferenças na detecção e na sobreposição temporal ocorreram na terceira temporada quando teve o maior número de capturas e recapturas da onça-pintada. A rede trófica apresentou seis níveis e grande riqueza de espécies e densidade de interações. Após a remoção da onça-pintada, houve a perda de um nível trófico e de aproximadamente 20% de interações. Quando comparada a distribuição atual da onça-pintada com a histórica, foi possível constatar a perda de espaço ambiental em áreas mais frias (< 6,3 °C), mais secas (< 288 mm), mais úmidas (> 7517 mm) e com maior elevação (> 3597 m). O modelo de distribuição também indicou que muitas áreas potenciais estão disponíveis à ocorrência da onça-pintada e que algumas merecem atenção, especialmente, nas porções leste e sul da Amazônia onde a espécie corre maior risco de desaparecimento local devido à antropização acelerada. Este estudo forneceu significativa contribuição para a compreensão da população de onça-pintada na Amazônia, bioma que ainda possui a maior lacuna de conhecimento sobre a biologia e ecologia da espécie. Demonstrou o efeito da sua presença na detectabilidade e no período de atividade de outras espécies. Também demonstrou seu papel ecológico como predador de topo exclusivo porque foi a única espécie que predou todas as outras de níveis tróficos inferiores. Descreveu o espaço ambiental originalmente ocupado (distribuição histórica) e o perdido (distribuição atual). Indicou as áreas mais adequadas à sua ocorrência e aquelas potenciais ao seu desaparecimento. / Apex predators play an important role in the maintenance of systems in which occur because they directly influence the structure and dynamics of communities Thus, this thesis aims to describe some of the mechanisms of spatio-temporal co-occurrence between the jaguar (Panthera onca) and mesopredators (Puma concolor and Leopardus pardalis), their top-down control on the community and environmental requirements that determine their distribution. To the analyses of spatio-temporal co-occurrence were used the occupancy model and the Kernel density with records derived from camera trapping (2008-2011). To the analysis of trophic networks was used a foodweb model with prey spectra of felids compiled from the available literature (1983-2014). To indicate the most suitable areas to its occurrence was used a species distribution model with geographic location of the species compiled from three different databases. The occupation of the two species of mesopredators was not different among sites with and without jaguar. The probability of detection was different for ocelot, with higher detection on sites with than without jaguar. The overlap in the activity pattern among jaguar and mesopredators was low, with approximately 55% of overlapping. Also there was a reasonable change in the overlap between sampling seasons, with a minimum of 32% and maximum of 56% of the difference in the activity pattern. The differences in the detection and temporal overlap occurred in the third season when occurred the highest number of jaguar captures and recaptures. The trophic network presented six levels and, a high species richness and links density. After removal of the jaguar, a trophic level and approximately 20% of links were lost. When compared historical and current distributions, the loss of environmental space occurred in areas colder (< 6.3 ° C), drier (< 288 mm), wetter (> 7517 mm) and at higher elevation (> 3597 m). The distribution model also indicated that there is many potential suitable areas available to the jaguar occurrence and, some of them deserve attention, particularly, in eastern and southern portions of Amazonia where the species is most at risk of local extinction. This study provided a significant contribution to the knowledge of the population of jaguar in the Amazon biome which still has the largest gap of information on the biology and ecology of the species. It demonstrated the effect of the presence of the species in the detectability and the activity pattern of others. Also, it demonstrated their ecological role as an apex predator exclusive because it was the only one that ate all other species of lower trophic levels. It described the environmental space originally occupied (historical distribution) and lost (current distribution). It indicated potential suitable areas to its occurrence and threatened areas to its disappearance.

Predadores de topo desempenham um papel importante na manutenção dos sistemas em que eles ocorrem porque influenciam diretamente a estrutura e a dinâmica de comunidades. Desta forma, este estudo descreveu alguns mecanismos de co-ocorrência espaço-temporal entre a onça-pintada (Panthera onca) e duas espécies de mesopredadores (Puma concolor e Leopardus pardalis), o seu controle top-down na comunidade de mamíferos e os requerimentos ambientais que determinam a sua distribuição no bioma Amazônia. Para as análises de co-ocorrência espaço-temporal foram utilizados o modelo de ocupação e a densidade de Kernel utilizando os registros de armadilhamento fotográfico (2008-2011). Para descrever o controle top-down foi elaborada uma rede trófica utilizando os itens alimentares consumidos pelas espécies e disponíveis na literatura (1983-2014). Para indicar quais as áreas mais adequadas foi utilizado o modelo de distribuição de espécies utilizando as localizações geográficas de ocorrência da espécie que foram compiladas de três diferentes bases de dados (2000-2013). A ocorrência das duas espécies de mesopredadores não foi diferente nos sítios com e sem a onça-pintada nas quatro temporadas de amostragem. A probabilidade de detecção da jaguatirica foi maior nos sítios com a presença da onça-pintada em apenas uma temporada de amostragem. A sobreposição no período de atividade das espécies de mesopredadores com a onça-pintada foi baixa, com cerca de 55% de sobreposição. Também ocorreu uma mudança razoável no período de atividade das espécies entre as temporadas de amostragem, com mínima de 32% e máxima de 56% de sobreposição. As maiores diferenças na detecção e na sobreposição temporal ocorreram na terceira temporada quando teve o maior número de capturas e recapturas da onça-pintada. A rede trófica apresentou seis níveis e grande riqueza de espécies e densidade de interações. Após a remoção da onça-pintada, houve a perda de um nível trófico e de aproximadamente 20% de interações. Quando comparada a distribuição atual da onça-pintada com a histórica, foi possível constatar a perda de espaço ambiental em áreas mais frias (< 6,3 °C), mais secas (< 288 mm), mais úmidas (> 7517 mm) e com maior elevação (> 3597 m). O modelo de distribuição também indicou que muitas áreas potenciais estão disponíveis à ocorrência da onça-pintada e que algumas merecem atenção, especialmente, nas porções leste e sul da Amazônia onde a espécie corre maior risco de desaparecimento local devido à antropização acelerada. Este estudo forneceu significativa contribuição para a compreensão da população de onça-pintada na Amazônia, bioma que ainda possui a maior lacuna de conhecimento sobre a biologia e ecologia da espécie. Demonstrou o efeito da sua presença na detectabilidade e no período de atividade de outras espécies. Também demonstrou seu papel ecológico como predador de topo exclusivo porque foi a única espécie que predou todas as outras de níveis tróficos inferiores. Descreveu o espaço ambiental originalmente ocupado (distribuição histórica) e o perdido (distribuição atual). Indicou as áreas mais adequadas à sua ocorrência e aquelas potenciais ao seu desaparecimento. / Apex predators play an important role in the maintenance of systems in which occur because they directly influence the structure and dynamics of communities Thus, this thesis aims to describe some of the mechanisms of spatio-temporal co-occurrence between the jaguar (Panthera onca) and mesopredators (Puma concolor and Leopardus pardalis), their top-down control on the community and environmental requirements that determine their distribution. To the analyses of spatio-temporal co-occurrence were used the occupancy model and the Kernel density with records derived from camera trapping (2008-2011). To the analysis of trophic networks was used a foodweb model with prey spectra of felids compiled from the available literature (1983-2014). To indicate the most suitable areas to its occurrence was used a species distribution model with geographic location of the species compiled from three different databases. The occupation of the two species of mesopredators was not different among sites with and without jaguar. The probability of detection was different for ocelot, with higher detection on sites with than without jaguar. The overlap in the activity pattern among jaguar and mesopredators was low, with approximately 55% of overlapping. Also there was a reasonable change in the overlap between sampling seasons, with a minimum of 32% and maximum of 56% of the difference in the activity pattern. The differences in the detection and temporal overlap occurred in the third season when occurred the highest number of jaguar captures and recaptures. The trophic network presented six levels and, a high species richness and links density. After removal of the jaguar, a trophic level and approximately 20% of links were lost. When compared historical and current distributions, the loss of environmental space occurred in areas colder (< 6.3 ° C), drier (< 288 mm), wetter (> 7517 mm) and at higher elevation (> 3597 m). The distribution model also indicated that there is many potential suitable areas available to the jaguar occurrence and, some of them deserve attention, particularly, in eastern and southern portions of Amazonia where the species is most at risk of local extinction. This study provided a significant contribution to the knowledge of the population of jaguar in the Amazon biome which still has the largest gap of information on the biology and ecology of the species. It demonstrated the effect of the presence of the species in the detectability and the activity pattern of others. Also, it demonstrated their ecological role as an apex predator exclusive because it was the only one that ate all other species of lower trophic levels. It described the environmental space originally occupied (historical distribution) and lost (current distribution). It indicated potential suitable areas to its occurrence and threatened areas to its disappearance.

L'objectif principal de cette thèse était d'étudier la dynamique spatiale et temporelle de la structure des communautés de poissons du fleuve Mékong en aval et du grand lac Tonlé Sap au Cambodge. Les deux systèmes sont fortement liés et caractérisés par les systèmes d'inondation pulsés. À l'aide d'approches multivariées sur des bases de données piscicoles, environnementales et des traits fonctionnels des taxons étudiés, j'ai pu mettre en évidence : * les gradients longitudinaux des caractéristiques physicochimiques des eaux et l'importance relative des facteurs environnementaux dans la structuration des communautés de poissons à large échelle ; * la structure temporelle des communautés de poissons qui est fortement liée aux traits fonctionnels des taxons et déclenchée par les crues occasionnées du système * la structure de base du réseau trophique et l'état écologique du système d'inondation pulsé qui est perturbé par la pression anthropique. Ces résultats ont démontré leurs valeurs quant à leurs potentiels de transferts vers la compréhension écologique et la gestion durable des ressources de pêches dans le système d'inondation pulsé en Asie tropicale en relevant que : * la connectivité entre le fleuve Mékong et le lac Tonlé Sap est indispensable pour les poissons pour compléter leurs cycles de vie, et la conservation devrait prioritairement s'effectuer au milieu du Mékong où se trouve la plus forte diversité. * dans le système d'inondation pulsé, la structure temporelle des communautés dépend des dynamiques des crues, donc l'aménagement des cours d'eau est à éviter au maximum. * la compréhension bioécologique complète des communautés et leurs relations trophiques sont primordiales pour la mise en place des plans de gestion et de conservation à long terme. / The main objective of the thesis was to study the spatial and temporal dynamics of fish community structure in Lower Mekong river basin and Tonle Sap Great Lake in Cambodia. These two systems are strongly connected to each other and characterised by flood pulse system. By using the multivariate approaches on the fish, environmental and fish functional traits data sets, I am able to highlight: * the longitudinal gradients of physicochemical characteristics of water and the relative importance of environmental factors in explaining the large scale patterns of fish community; * the temporal structure of fish community, which is strongly linked to functional traits of fish and trigged by the pulse dynamics of the system; * the baseline trophic structure model and ecological health assessment of flood pulse system, which is disturbed by anthropogenic pressure. These results have demonstrated their values as well as their potentials to transfers toward an ecological understanding and sustainable fisheries resource management of the flood pulse system in tropical Asia by suggesting: * the connectivity between Mekong river and Tonle Sap lek is essential for fish to complete their life cycle and also the conservation should be done in the middle part of Lower Mekong river where containing the highest diversity; * in flood pulse system, the temporal structure of fish community strongly depend on the dynamics of pulses; therefore any river development must be strictly prohibited; * complete bio-ecological understanding of communities and their trophic relationships are crucial to the establishment of long-term management and conservation plans.