Die Nutzung von Beweidung zur Erhaltung und Förderung von Biodiversität wird allgemein als positiv erachtet, da so strukturell diverse Flächen erhalten werden können. Die Auswirkungen unterschiedlicher Managementszenarien auf die Artenvielfalt sind zwar weitreichend untersucht, auf Langzeitexperimenten jedoch sehr rar. Mit unserem ersten Experiment wollten wir die Frage beantworten, welche Nutzungsintensität am besten geeignet ist, um Diversität von Insekten zu fördern. Das dafür genutzte Langzeitexperiment wurde 2002 mit zwei Beweidungsintensitäten etabliert und eine dritte Intensitätsstufe 2005 eingeführt. Zwischen 2002 und 2004 sowie 2010 und 2011 wurde der Einfluss der Beweidungsvarianten auf die Insektendiversität im Sinne des Artenreichtums und der Abundanz von Heuschrecken (Orthoptera) und Tagfaltern (Lepidoptera) aufgenommen. Wie erwartet, zeigte sich bei intensivster Beweidung die geringste Diversität. Trotzdem war die extensivste Beweidung nicht vorteilhafter als die mittel intensive Beweidung. Aus diesem Grund empfehlen wir eine mittelstarke Beweidung, da diese besonders geeignet ist um eine ausreichende Narbenhöhe und eine strukturell diverse Grasnarbe zu schaffen. Dieses Management zeichnet sich durch vielfältige ökologische Nischen aus, die einer Bandbreite von unterschiedlichen Arten Habitat bieten, wovon besonders weniger mobile Herbivoren wie Heuschrecken profitieren können. Wenn Pflanzendiversität im Vordergrund steht, muss auch die Art des Weidetiers betrachtet werden, da sich die Tierarten in ihren Beweidungseigenschaften unterscheiden und demnach auch Vegetation und Narbenstruktur beeinflusst werden können. In einem zweiten Experiment haben wir deshalb den Einfluss von Rindern und Schafen in Monobeweidung als auch in Mischbeweidung untersucht. Eine Hälfte der Parzellen wurde mit Herbiziden gegen Dikotyle behandelt um den Weidetiereffekt in Abhängigkeit der Narbenzusammensetzung zu analysieren. Das Experiment wurde zwischen 2007 und 2011 beweidet und die Vegetationszusammensetzung jeweils vor jeder Beweidungsperiode und zusätzlich im Jahr 2012 aufgenommen. Die beiden Narbentypen zeigten wie erwartet deutliche Unterschiede in Zusammensetzung und Diversität. Dahingegen unterschieden sich die Weidetier-Varianten nur marginal voneinander. Die Rinderparzellen wiesen zwar eine höhere alpha-(Artenreichtum) als auch beta-Diversität auf, allerdings war dieser Effekt nur in einem Jahr auf jeweils einem Narbentyp zu finden. Zusammenfassend können gewöhnliche Weiden in ihrer Vegetationszusammensetzung und -diversität mithilfe unterschiedlicher Tierarten nur geringfügig gesteuert werden. Es konnte lediglich ein marginaler Vorteil von Rinder- gegenüber Schafbeweidung festgestellt werden Auf demselben Experiment waren wir weiterhin daran interessiert, wie pflanzliche Diversität und agrarische Produktivität miteinander verbunden sind. Unsere Ergebnisse verneinen jedoch einen positiven Zusammenhang. Weiden besitzen wahrscheinlich ein genügend großes Arteninventar auch bei artenarmer, herbizidbehandelter Ausprägung, so dass eine komplementäre Ressourcennutzung bereits ausgeschöpft wird. Im Gegensatz dazu enthielten die artenreicheren Futterproben tatsächlich mehr Rohprotein und weniger ADF, so dass von einer gesteigerten Futterqualität gesprochen werden kann. Diese Ergebnisse könnten jedoch auch durch ein erhöhtes Vorkommen von Leguminosen erklärt werden, da auf dem Versuch aufgrund des Designs eine hohe Korrelation von Artenreichtum mit Kräutern und Leguminosen besteht. Hinsichtlich der unterschiedlichen Weidetiere konnte kein signifikanter Effekt weder auf Primärproduktivität noch auf Futterqualität festgestellt werden. Über alle Jahre hinweg zeigte sich vor allem die Phytodiversität als förderlich für die Lebendmassezunahmen der Lämmer. Zudem konnte die Mischbeweidung zu einer gesteigerten Wachstumsintensität von Lämmern und Mutterkühen beitragen. Zusammenfassend kann Beweidung genutzt werden um die Narbenstruktur zu steuern, was vor allem durch eine Anpassung der Beweidungsintensität erreicht werden kann. Nachfolgend wird sich die Intensität auf die Biodiversität aus, die jedoch durch die Wahl des Weidetieres nur recht geringfügig beeinflusst wird. Auch wenn die Primärproduktivität nicht mit dem Artenreichtum verbunden war, so kann durch das Hinzukommen von Kräutern und Leguminosen doch eine Steigerung der Futterqualität erreicht werden, die sich auch in höheren Lebendmassezunahmen der Lämmer wiederspiegelt. Demnach kann durch ein geeignetes Graslandmanagement Einfluss auf die Biodiversität genommen werden, welche wiederum das Potential besitzt, die tierischen Leistungen zu steigern.

La qualité des services écosystémiques qu'offrent les lacs est liée à la structure et au fonctionnement de leur écosystème. Protéger leur masse d'eau est devenu un objectif global qui requiert une meilleure compréhension de leur fonctionnement, un suivi et une évaluation de leur qualité. Expliquer les hétérogénéités spatio-temporelles des variables physico-chimiques et du phytoplancton est un problème récurrent rencontré en écologie et en hydrobiologie. Comprendre la dynamique de ces hétérogénéités est aussi un prérequis essentiel pour évaluer, protéger et restaurer objectivement les écosystèmes lacustres. En ce qui concerne la surveillance, les hétérogénéités spatio-temporelles introduisent des incertitudes sur la représentativité des mesures par rapport à l'entièreté de la masse d'eau qui est donc discutable et doit être vérifiée. En Europe, la directive cadre sur l'eau (DCE) initiée en 2000 définit un cadre pour la gestion et la protection des eaux. La classification des masses d'eau en fonction de leur état écologique est un point important dans l'implémentation de cette directive. Pour les lacs et les retenues, l'évaluation de cet état écologique est basée sur des paramètres biologiques, physico-chimiques et hydromorphologiques. Les indicateurs liés au phytoplancton et aux paramètres physico-chimiques sont calculés à partir de quatre prélèvements réalisés pendant la période d'activité biologique pour une année sur un plan de gestion de six ans. Dans ce contexte, la modélisation tridimensionnelle (3D) et la prise en compte des forçages qui conduisent aux hétérogénéités spatio-temporelles est une condition préalable nécessaire tant en limnologie appliquée que théorique. Cette thèse aborde la complexité du fonctionnement des lacs et la capacité des modèles 3D à reproduire leur fonctionnement. L'apport de la modélisation 3D est présenté i) pour la compréhension du fonctionnement de lacs de différentes tailles, ii) couplée aux observations satellitaires, pour l'étude de l'influence des forçages par le vent et de l'hydrodynamique sur l'abondance et la distribution spatiale de phytoplancton, iii) dans le cadre de la DCE, pour l'évaluation des incertitudes d'une évaluation de l'état écologique d'un plan d'eau. Pour cela, deux modèles 3D ont été créés et analysés, un pour le lac de Créteil (42~ha) et un autre pour le Léman (580~km$^2$). Celui du lac de Créteil a été validé à partir de données à hautes fréquences acquises en trois points du lac. Il reproduit correctement son hydrodynamique complexe, sa structure thermique, l'alternance entre les périodes de stratification et les épisodes de mélange, ainsi que les ondes internes. Le modèle du Léman a été validé en utilisant des données mensuelles et bimensuelles en deux stations de prélèvement du lac. Il reproduit aussi correctement son hydrodynamique et la variabilité saisonnière de paramètres physico-chimiques et biologiques. Les résultats des simulations mettent en avant les mécanismes physiques et hydrodynamiques responsables de l'apparition de sites où la biomasse de phytoplancton observée est plus élevée. Dans le cadre de la DCE, ces résultats montrent aussi une variabilité spatiale importante des sous-états écologiques basés sur les différents paramètres qui dépendent du choix des dates des campagnes de mesure et du point d'échantillonnage. Ces résultats ont aussi été utilisés pour estimer la représentativité d'une station de prélèvement. Les résultats de cette thèse i) confirment que le fonctionnement des plans d'eau de toute taille est complexe et que les processus physiques génèrent des hétérogénéités spatio-temporelles, ii) suggèrent que le vent et l'hydrodynamique influencent significativement l'abondance et la distribution spatiale du phytoplancton et que iii) ces hétérogénéités peuvent biaiser notre estimation du statut écologique des plans d'eau dans le cadre de la DCE / The quality of ecosystem services provided by lakes is related to the ecosystem structure and functioning. Protecting water bodies is therefore a global goal that requires a better understanding of their function, a monitoring and a water quality assessment. Explaining spatio-temporal heterogeneities of physico-chemical parameters and phytoplankton has been a recurrent ecological and hydrobiological issue. Understanding the dynamics of these heterogeneities is an essential prerequisite for objectively assessing, protecting and restoring freshwater ecosystems. Moreover, three-dimensional (3D) and taking into account and drivers of these heterogeneities are essential prerequisites for theoretical and applied limnology. Concerning the monitoring, spatio-temporal heterogeneities are responsible of uncertainties on the representativeness of the data versus the whole lake which might be questionable and needs to be verified. In Europe, the Water Framework Directive (WFD) initiated in 2000 defines a framework for managing and protecting water bodies in Europe. The classification of water bodies into ecological status is a key issue for the implementation of that framework. For lakes and reservoirs, the assessment of this status is based on biological, physico-chemical and hydro-morphological indicators. Physico-chemical and phytoplankton indicators are calculated based on four observations at an unique sampling station over the growing season, this evaluation being assessed one year for a six-year management plan. In this context, this thesis focuses on the complexity of lakes functioning and the capability of three-dimensional (3D) models to reproduce their functioning. The contribution of 3D models is presented i) for understanding the functioning of lakes of different sizes, ii) coupled to satellite observations, for studying of the influence of wind forcing and hydrodynamics on phytoplankton abundance and spatial heterogeneities, iii) in the context of the WFD, for assessing uncertainties in the lake ecological status assessment. To do that, two 3D models have been created and analyzed, one for Lake Créteil (42 ha) and another for Lake Geneva (580 km2). Lake Créteil 3D model was validated by using high frequency data recorded at three stations. It reproduces well the complex hydrodynamic functioning of the lake, its thermal structure, the alternation between thermal stratification episodes and mixing events, and internal waves. Lake Geneva 3D model was validated by using monthly and bimonthly data at two stations. It reproduces also properly the hydrodynamic functioning of the lake and the seasonal variability of biological and physico-chemical parameters. Simulation results highlight physical and hydrodynamic mechanisms responsible for the occurrence of seasonal hot-spots in phytoplankton abundance. In the context of the WFD, simulation results show also a strong spatial variability of lake ecological status depending on the timing of the four sampling dates as well as the location of the sampling station. These results were also used to assess to representativeness of sampling stations. The results of this thesis suggest that i) the functioning of lakes of different sizes is complex and physical processes generates spatio-temporal heterogeneities, ii) wind and hydrodynamics influence the abundance and the spatial distribution of phytoplankton et iii) spatio-temporal heterogeneities can bias our evaluation of lake ecological status in the WFD

Submitted by Jos? Henrique Henrique (jose.neves@ufvjm.edu.br) on 2017-06-09T18:42:59Z No. of bitstreams: 2 mayara_ribeiro_lage.pdf: 7630209 bytes, checksum: fb39ab87647bdca7bc30286669fad378 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2017-06-22T14:54:33Z (GMT) No. of bitstreams: 2 mayara_ribeiro_lage.pdf: 7630209 bytes, checksum: fb39ab87647bdca7bc30286669fad378 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-06-22T14:54:33Z (GMT). No. of bitstreams: 2 mayara_ribeiro_lage.pdf: 7630209 bytes, checksum: fb39ab87647bdca7bc30286669fad378 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior (CAPES) / Um trecho da BR- 367, que liga a cidade de Diamantina ao distrito de Guinda, em Minas Gerais, perpassa por ?reas de campo rupestre na Serra do Espinha?o Meridional e apresenta diversas ?reas de empr?stimo, constitu?das por jazidas de material granular consumidas em obra rodovi?ria da pr?pria via entre as d?cadas de 1960 e 1980. Atualmente tais ?reas ainda apresentam-se fortemente degradadas, caracterizando-se pela fraca cobertura vegetal, exposi??o do solo e inexist?ncia de um sistema de drenagem superficial eficiente. Neste contexto, surge a necessidade de alternativas que tenham efic?cia na integra??o dos conhecimentos silviculturais e princ?pios ecol?gicos visando o retorno destas ?reas ?s condi??es pr?ximas das originais. Este trabalho teve como objetivo entender a rela??o entre vari?veis ambientais e a estrutura horizontal e flor?stica das esp?cies colonizadoras destes ambientes degradados. Foram selecionadas quatro ?reas degradadas, analisadas por fotografias a?reas referentes aos anos de 1960 e 1979 nas quais foi poss?vel detectar das ?reas degradadas pontuais. Para o estudo do substrato estratificou-se as quatro ?reas selecionadas para a coleta de amostras deformadas e indeformadas de forma preferencial. Obteve-se as propor??es de areia, silte, argila, dados de resist?ncia ? penetra??o, densidade de part?culas e aparente, porosidade total, micro e macro porosidade. Para a an?lises qu?micas foram coletadas entre 5 subamostras de cada estrato com o intuito de se obter uma amostra composta do substrato superficial (0-20 cm), de forma a representar a heterogeneidade ambiental de cada ?rea. Foram analisados: pH em ?gua; teores de P, K+, Mg2+, Ca2+ e Al3+; complexo sortivo (acidez potencial (H+Al), satura??o e soma por bases (V% e SB), CTC (T), CTC efetiva (t) e satura??o por alum?nio (m%) e mat?ria org?nica (MO). Os dados obtidos foram aferidos e submetidos a an?lises estat?sticas. Realizou-se an?lise de componentes principais (PCA) pelo software PCORD vers?o 6, de modo a sintetizar as vari?veis de maior relev?ncia na correla??o das vari?veis analisadas. Para levantamento da comunidade vegetal foi empregado o m?todo da interse??o na linha no qual em cada ?rea foram alocadas doze linhas com 10,0 m de comprimento, distribu?das em zigzag, seguindo as coletas de solo, para correlacionar caracter?sticas do solo e vegeta??o. Cada linha foi subdividida em dez unidades amostrais (UA) cont?nuas, totalizando 470 UAs e anotado a ocorr?ncia de esp?cies de h?bito herb?ceo, arbustivo e arb?reo. Foram calculadoss par?metros fitossociol?gicos: frequ?ncia absoluta (FA), frequ?ncia relativa (FR), cobertura absoluta (CA), cobertura relativa (CR) e o valor de import?ncia (VI). Para comparar os diferentes setores das ?reas estudadas quanto ao perfil de estrat?gias ecol?gicas de esp?cies herb?ceas, estas foram classificadas em sistemas de guildas de acordo com as estrat?gias de regenera??o, estratifica??o e dispers?o. Nas quatro ?reas analisadas foram amostrados 1.517 indiv?duos, pertencentes a 22 fam?lias e 109 esp?cies. A an?lise dos atributos f?sicos e qu?micos do substrato evidenciou alta limita??o ao desenvolvimento de plantas colonizadoras, apresentando como restri??es a alta resist?ncia a penetra??o e substratos de textura arenosa com baixa fertilidade natural. Verificou-se que houve rela??o entre os gradientes ambientais e a abund?ncia e composi??o flor?stica da vegeta??o colonizadora, ficando a maioria das esp?cies mais fortemente correlacionada com as vari?veis MO, H+Al e V al?m de propor??o de Areia, PT, DMG e RP. / Disserta??o (Mestrado) ? Programa de P?s-Gradua??o em Ci?ncia Florestal, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2016. / A section of the BR-367, connecting the Diamantina town to the Guinda district, Minas Gerais, runs through areas of rupestrian field of Espinha?o Range and presents several loan areas, consisting of deposits of granular material consumed in road construction between the 1960?s and 1980?s, Currently, these areas are still heavily degraded, characterized by poor vegetation cover, soil exposure and the absence of an efficient surface drainage system. In this context, the need arises for alternatives that have effectiveness in the integration of silvicultural knowledge and ecological principles aiming the return of these areas to the conditions close to the originals. This work aimed to understand the relationship between environmental variables and the horizontal and floristic structure of the colonizing species of these degraded environments. Four degraded areas were selected, analyzed by aerial photographs referring to the years 1960 and 1979 in which it was possible to detect the degraded areas. For the study of the substrate, the four areas selected for the collection of deformed and undisturbed samples were stratified in a preferential way. The proportions of sand, silt, clay, penetration resistance data, particle and apparent density, total porosity, micro and macro porosity were obtained. For the chemical analyzes, 5 sub-samples of each stratum were collected with the aim of obtaining a composite sample of the surface substrate (0-20 cm), in order to represent the environmental heterogeneity of each area. The following parameters were analyzed: pH in water; Contents of P, K +, Mg2 +, Ca2 + and Al3 +; (S), saturation (V%), sum of bases (SB), CTC (T), effective CTC (t) and saturation by aluminum (m%) and organic matter. A PCA was performed by software PCORD version 6, in order to synthesize the variables of greater relevance in the correlation of the analyzed variables. Was used the method of the intersection in the line in which in each area were allocated twelve lines with 10.0 m in length, distributed in zigzag, following the soil collections, to correlate soil and vegetation characteristics. Each row was subdivided into ten continuous sample units (AU), totaling 470 POAs and the occurrence of species of herbaceous, shrub and arboreal habit. Phytosociological parameters were calculated: absolute frequency (AF), relative frequency (FR), absolute coverage (AC), relative coverage (CR), and importance value (VI). In order to compare the different sectors of the studied areas with respect to the profile of ecological strategies of herbaceous species, these were classified in guild systems according to the strategies of regeneration, stratification and dispersion. In the four areas analyzed, 1,517 individuals belonging to 22 families and 109 species were sampled. The analysis of the physical and chemical attributes of the substrate showed a high limitation to the development of colonizing plants, presenting as restrictions the high resistance to penetration and substrates of sandy texture with low natural fertility. It was verified that there was a relation between the environmental gradients and the abundance and floristic composition of the colonizing vegetation, being most of the species more strongly correlated with the variables MO, H+Al and V besides the proportion of Sand, PT, DMG and PR.

In 2011, vertical and longitudinal gradients of temperature, oxygen, pH, conductivity and chlorophyll were measured in the Římov Reservoir. Relative distance model was employed to evaluate spatial heterogeneity of parameters measured along the longitudinal axis of the reservoir. The inflow part comprised 15 % of the overall reservoir length and was characterized with temperature, pH and chlorophyll values lower than the rest of the reservoir. Highest chlorophyll and dissolved oxygen concentrations along with pH were measured in the transition zone constituting 40 % of the reservoir length. Lacustrine zone with low chlorophyll concentration accounted for 45 % of the reservoir. It was characterized with pronounced metalimnetic and hypolimnetic oxygen minima during summer as well as the marked increase in conductivity towards bottom.

Les systèmes agro-sylvo-pastoraux (SASP) d’Afrique de l’Ouest sont des agro-écosystèmes limités en biomasses et en nutriments. Le recyclage des nutriments et les transferts de fertilité sont traditionnellement rythmés par la mobilité des troupeaux de ruminants conduits en extensif. Les agro-éleveurs pratiquent le parcage nocturne de leurs troupeaux pour concentrer la matière organique, dans les champs à proximité des habitations afin de sécuriser une production vivrière suffisante à leurs besoins. Dans un contexte de croissance démographique et de réduction des parcours naturels au profit des zones cultivées, le système d’élevage « traditionnel », basé sur une forte mobilité intra-terroir villageois, est remis en cause. Les stratégies adoptées par les agro-éleveurs sont, soit (i) l’éloignement des troupeaux du terroir villageois pendant des périodes plus ou moins longues par des pratiques de transhumance saisonnière vers des régions moins peuplées et disposant de davantage de ressources fourragères ; soit (ii) plus récemment, des pratiques d’intensification avec des animaux gardés à l’étable au sein du terroir villageois et nourris avec des aliments concentrés, achetés sur le marché local. Ces changements de systèmes d’élevage ont possiblement des conséquences importantes sur les flux de biomasses et les cycles des nutriments au niveau du ménage et du territoire. Il convenait de les évaluer en termes d’impacts sur le fonctionnement et la durabilité des SASP. A cet effet, le modèle multi-agents TERROIR a été développé et implémenté sur la plateforme de modélisation GAMA. Il simule l’effet de changements dans l’organisation du paysage et des systèmes d’élevage sur les flux de biomasse et d’azote aux différents niveaux d’organisation du territoire : la parcelle, le troupeau, le ménage et le terroir villageois. Le modèle simule les échanges de biomasses entre une centaine de ménages comportant des stratégies et des pratiques différentes. Cela inclut les transferts spatiaux de biomasses orchestrés par plusieurs centaines de troupeaux se déplaçant de façon indépendante sur un millier de parcelles. Le modèle synthétise ces flux par un ensemble d’indicateurs issus de deux méthodes d’analyse (« Ecological Network Analysis » et « System Gate Balance ») pour décrire la structure, le fonctionnement et la durabilité de l’agroécosystème, en termes de productivité, d’efficience, d’autonomie, de recyclage, de transferts spatiaux et de bilan de nutriments. Le modèle a été conçu et paramétré à partir des données disponibles sur les agroécosystèmes de savane en Afrique de l’Ouest et il a été évalué à partir des données observées dans deux terroirs villageois du bassin Arachidier au Sénégal où les pratiques des agro-éleveurs sont particulièrement contrastées.Le modèle TERROIR a été utilisé pour explorer les impacts des dynamiques territoriales observées sur la période 1920-2015 dans le bassin Arachidier au Sénégal, une zone agricole à transition agraire rapide et avancée. Les résultats soulignent une réorganisation du cycle de l’azote et une tendance générale à l’intensification des flux et à l’augmentation de la dépendance des agroécosystèmes vis-à-vis de sources extérieures de nutriments. Cependant, le recyclage et les transferts spatiaux de nutriments internes aux agrosystèmes restent à des niveaux élevés. L’intégration sol-plantes-animaux-hommes et l’hétérogénéité spatiale de la répartition des ressources fertilisantes apparaissent comme deux propriétés persistantes des agro-écosystèmes étudiés. Consolider cette intégration et cette organisation spatiale seraient ainsi un gage pour la durabilité des futurs systèmes agricoles qui émergeront dans un contexte de poursuite de la forte croissance démographique et de changement climatique. / Agro-sylvo-pastoral systems (systèmes agro-sylvo-pastoraux - SASP) of West Africa are agroecosystems limited in biomass and nutrients. Nutrient recycling and fertility transfer are traditionally driven by the mobility of ruminant herds led in extensive practices. Agro-pastoralists practice night corralling of their herds to concentrate the organic matter in the fields near the houses, in order to secure a sufficient food production for their needs. In a context of demographic growth and the reduction of natural rangelands in favor of cultivated areas, the "traditional" mobile livestock system, based on high mobility within the village is being called into question. The strategies adopted by the agro-pastoralists are: (i) keeping the herds away from the village for periods of varying lengths, by seasonal transhumance in less populated regions where forage resources are more important, or (ii) more recently, intensified practices with animals kept in the barn within the village and fed with concentrate feeds, bought on the local markets. These changes in livestock systems may have important consequences for biomass flows and nutrient cycling at the household and village landscape level. There was a need to assess their impact on the functioning and sustainability of SASP.To this end, the TERROIR multi-agent model has been developed and implemented on the GAMA modeling platform. It simulates the effect of changes in the organization of the landscape and livestock systems on biomass and nitrogen flows at different levels of organization in the village: plot, herd, household, village landscape. The model simulates the exchanges of biomasses between dozens of households with different strategies and practices. It includes the spatial transfers of biomasses between several hundred plots orchestrated by dozens of herds moving independently. The model synthesizes these flows with a set of indicators from two methods of analysis (Ecological Network Analysis and System Gate Balance) to describe the structure, functioning and sustainability of the agroecosystem, in terms of productivity, efficiency, autonomy, recycling, spatial transfers and nutrient balance. The model was designed and configured with available data on savannah agroecosystems in West Africa. It was evaluated from data observed in two villages of the Groundnut Basin in Senegal where the practices of agro-pastoralists are particularly contrasted. The TERROIR model was developed and implemented to explore the impacts of the village dynamics observed over the period 1920-2015 in the Groundnut Basin in Senegal, an agricultural zone in fast and advanced agrarian transition. The results highlight a reorganization of the nitrogen cycle and a general trend towards increased flows and increased dependence of agroecosystems on external sources of nutrients. However, the recycling and spatial transfers of nutrients internal to agroecosystems remain at high levels. The soil-plant-animal-human integration and the spatial heterogeneity of the distribution of fertilizing resources appear as two persistent properties of the studied agro-ecosystems. Consolidating this integration and spatial organization could guarantee for the sustainability of future farming systems that will emerge in a context of continued high-population growth and climate change.

Le déclin de la biodiversité lié à l'augmentation des pressions sur les ressources naturelles, fait l'objet d'un large consensus. Cette situation souligne un besoin urgent de développer des outils de diagnostic et de suivi de l'état de la biodiversité qui soient spatialisés, rapides à mettre en œuvre, peu coûteux et qui permettent de réaliser des expertises à l'échelle des territoires. Une des hypothèses centrales en écologie est qu'il existe des relations positives entre l'hétérogénéité spatiale mesurée dans la mosaïque paysagère et la richesse en espèces qui peut y cohabiter à différentes échelles. Dans ce travail, nous nous appuyons sur cette hypothèse pour développer un Indice de Biodiversité Spécifique Potentielle afin de réaliser un diagnostic de l'état de la biodiversité dans la mosaïque paysagère intra-forestière à différentes échelles, en prenant l'exemple du massif du Vercors. Cet indice s'appuie sur des indicateurs et des critères capables de refléter le niveau de biodiversité potentiellement présente dans une zone géographique donnée, et permet d'approfondir les connaissances sur les relations entre la structure des peuplements, les caractéristiques de la mosaïque forestière et la richesse ou la répartition géographique des espèces. Les étapes clés de la méthodologie et de son évaluation sont valorisées sous forme d'articles. La représentation spatiale de la biodiversité potentielle à différentes échelles permet d'envisager de nombreuses applications dans les domaines de la conservation et de la gestion forestière multifonctionnelle. / There is a general consensus that as pressures on natural resources increase, biodiversity is being lost, and the rate of loss is not sustainable. Within the present situation is become a critical need to develop methods and tools for biodiversity monitoring and conservation. In particular, we need spatially-explicit tools and rapid assessment methods that are relatively easy to implement at different scales and at a relatively low cost. One of the basic assumptions in ecology is that there are positive relationships at different spatial scales between landscape heterogeneity and species richness in a given landscape mosaic. In this thesis, we focus on this assumption in order to build a Potential Species Biodiversity Index. Our aim was to assess potential biodiversity within a forest landscape mosaic, considering the Vercors' mountain range as a case study. The index encompasses indirect biodiversity indicators and criteria as a surrogate to biodiversity potential present for a given area. This approach allows a better understanding of the relationships between forests stand structure, forest landscape characteristics and species richness or geographical distribution. The main steps of the method and its evaluation are presented in scientific papers. The analysis of geographical distribution of the index at different spatial scales provides practical applications for conservation planning and multifunctional forest management related fields.

Do thermal barriers limit biotic composition and community similarity, potentially helping to shape biodiversity patterns at continental scales? Are environmental variables responsible for broad-scale patterns of species endemism? Are these patterns predictable? And, how can patterns of endemism can inform global conservation strategies? These are some of the questions that I attempted to answer during my doctoral research. In the first chapter, I tested one of the most contentious hypotheses in ecology: Do thermal barriers, which grow stronger along elevational gradients across tropical mountains, create a dispersal barrier to organisms and consequently contribute to the isolation and divergence of species assemblages? If so, do patterns potentially generated by this mechanism detectably relate to dissimilarity of biotic assemblages along altitudinal gradients across the mountains in the Americas? We found that mountain passes are not only higher in tropical realms, as initially thought by Janzen (1967), and extensively popularized and assumed in further research, but they are also present in temperate regions along the western coast of North America. We also found that the stronger the thermal barrier, the higher the dissimilarity between communities. However, the variance explained was low, suggesting thermal barriers play a minor role in creating and maintaining patterns of biodiversity. The second chapter raises the question of why are there more small-ranged species in some places than in others. I tested four macroecological hypotheses (H1: climate velocity; H2: climate seasonality; H3: climate distinctiveness or rarity; and, H4: spatial heterogeneity in contemporary climate, topography or habitat) to predict broad-scale patterns of species endemism, using a cross-continental validation approach. We found that there is no empirical reason, from the standpoint of model fitting, parameter estimates, and model validation, to claim that any of these hypotheses creates and maintains broad-scale patterns of endemism. Although we found statistically significant relationships, they failed stronger tests of a causal relationship, namely accurate prediction. That is, the hypotheses did not survive the test of cross-continental validation, failing to predict observed patterns of endemism. Climate velocity was dropped from some models, suggesting that early correlations in some places probably reflect collinearity with topography. The effect of richness on endemism was in some cases negligible, suggesting that patterns of endemism are not driven by the same variables as total richness. Despite low explained variance, spatial heterogeneity in potential evapotranspiration was the most consistent predictor in all models. The third chapter is aimed to evaluate the extent to which global protected areas (PAs) have included endemic species (species with small range size relative to the median range size). We measure the relative coverage of endemic species by overlapping species geographic ranges for amphibians, mammals, and birds, with the world database of PAs (1990-2016). Then we measure the rate of expansion of the global PA network and the rate of change in endemic species coverage. We found that ~30% of amphibian, ~6% of bird and ~10% of mammal endemic species are completely outside PAs. Most endemic species’ ranges intersect the PA network (amphibian species = 58%; birds = 83%; mammals = 86%), but it usually covers less than 50% of their geographic range. Almost 50% of species outside the PA network are considered threatened (critically endangered, endangered and vulnerable). We identified that ecoregions in tropical Andes, Mesoamerica, Pacific Islands (e.g., New Guinea, Solomon), Dry Chaco, and Atlantic forests are major conservation priorities areas. The historic rates of new PAs added every year to the network is between ~6,000 to ~15,000. In contrast, we found that rates of including endemic species within the PA network have been fairly slow. Historic data shows that every year, the entire geographic range of 3 (amphibians) to 6 (birds and mammals) endemic species is 100% included inside the PA network (amphibians = from 162 to 233; mammals = 10 to 84; and, amphibians = 16 to 99). Based on these trends, it is very unlikely PAs will include all endemic species (14% total endemic species, that is ~1,508 out of 11,274) currently outside the PA network by 2020. It will require five times the effort made in the last two decades. However, projections also showed that is very likely that some portions of the geographic ranges for all endemic birds and mammals, but not for all endemic amphibians, will be covered by the future PA network. I sum, I found that none of the hypotheses tested here can explain broad-scale patterns of total species richness and total species endemism. My main contribution on this research area is clearly rejecting these hypotheses from potential candidates that may explain biodiversity patterns. By removing them, we advance in this field and open possibilities to test new hypotheses and evaluate their mechanisms. I proposed that other drivers and mechanisms (whether biotic and biotic) acting at local scales, and escaping the detection of macroecological approaches, might be responsible for these patterns. Finally, in terms of conservation planning, I proposed that the international community has an opportunity to protect a great number of endemic species and their habitats before 2020, if they strategically create new PAs.

Abstract Plants and soil organisms are closely linked. Plants are the sole source of carbon in the soil and soil organisms are responsible for recycling of nutrients, making them available for plant growth. To understand the function of a system, it is important to understand the interactions between the soil and plants. These interactions have mainly been studied in temperate areas, with few studies in the arctic and subarctic. The aim of this thesis was to investigate the effect of ecological disturbances in sub- and low-arctic grasslands on soil organisms and plant-soil feedback relationships. The effect of removal of vegetation, replanting of a local plant species, and different components of grazing (trampling, defoliation and return of nutrients) on soil decomposer organisms were studied. Whether short term effects of defoliation depended on plant species community was also studied, as well as whether defoliation in the field could create changes in the soil system systems that affect the growth of seedlings. Experiments were conducted under both controlled greenhouse conditions and in field sites. The results showed that physical disturbance (removal of vegetation and trampling) reduced the abundance and diversity of soil biota. Defoliation increased soil decomposer abundance in the short term. Plant species composition did not affect soil biota and only in a few cases did it changes their responses to defoliation. In the long-term, effects of fertilization and defoliation on the soil biota were context-dependent. However, defoliation did create changes in the soil that reduced the growth of seedlings planted into the soil. Furthermore, plant species community and spatial heterogeneity (revealed by blocking) had important effects on the soil communities.

高新技術產業是深圳市產業升級的四大支柱產業之一，其中電子信息產業又佔其一半以上的比重。翻閱探討中國大陸電子資訊產業發展的相關文獻，研究對象多集中在城市內部的科學園區，並以北京和上海的科學園區居多。深圳市作為中國首個國家創新型城市，產業發展不局限於某個科學園區之內，形成了全市範圍的產業聚落。因此，本文以深圳市作為研究對象，探討電子資訊產業的發展狀況和影響因素。 當下，科技創新的重心已經移居深圳，體現在專利數量的優勢、密集的政策出台和資金人才的大量湧入。在產值的增長和知名企業的帶動之下，深圳的電子信息產業聚落尤為突出，已經代表中國在國際範圍施加影響。 然而論及深圳的電子信息產業，經濟特區內部存在先天的空間異質性，各區電子信息產業之傳統、政策和現狀都有很大差異，無法一概而論。本文以深圳市各行政區為分析單位，自主收集相關資料並採用追蹤數據，探討深圳市電子資訊產業在2008-2015年的再結構過程，以及哪些因素對各區產業發展的差異性產生影響。 / High and new technological industry is the most important industry of Shenzhen, in which IT industry is in the majority. Reviewing existing literatures, most of them focus on science parks in Beijing and Shanghai. But Shenzhen is the first official Innovative City in China, where IT industry doesn’t restricted to any science park. After the international financial crisis in 2008, Shenzhen representing China gave influence on the whole world in technology innovation. It has advantages for its IT industry cluster, industry output and large enterprises. Hence, this article will discuss affecting factors of the restructure of Shenzhen IT industry in 2008-2015. However, there is Spatial Heterogeneity in different regions in Shenzhen, such as industry tradition, policies and current situation. Furthermore, this study will base on a sub-regional strategy by panel data.

The South Florida Everglades is a unique ecosystem. Intensive water management in the system has facilitated agricultural, urban, and economic development. The Everglades offers a variety of ecosystem services (ES) to the people living in this region. Nevertheless, the ecosystem is under imminent threat of climate change, which would alter the way water is managed today and ultimately affect the ES offered by the system. On the other hand, substantial restoration is underway that aims to restore the Everglades closer to its historic condition. This research tried to map the public’s preference for Everglades restoration. Using a geocoded discrete-choice survey dataset, the study showed variation in the public’s preference by changing the levels of ES. Additionally, the general public’s attitude toward climate change risk to the Everglades and preference for mitigation were also assessed using the survey data.