Multiple dimensions of biodiversity within and across producer and consumer guilds in the food web affect an ecosystem’s functionality and stability. Tropical and subtropical aquatic ecosystems, which are extremely diverse, have received much less attention than terrestrial ecosystems in regards to the effects of biodiversity on ecosystem functioning. We conducted a field experiment that tested for effects of macrophyte functional diversity on diversity and stability of associated fish assemblages in floodplain lakes of the Upper Paraná River floodplain, Brazil. Three levels of macrophyte functional diversity were maintained through time in five floodplain lakes and response variables included various components of fish taxonomic and functional diversity and stability. Components of functional diversity of fish assemblages were quantified using a suite of ecomorphological traits that relate to foraging and habitat use. Response variables primarily distinguished macrophyte treatments from the control. Macrophyte treatments had, on average, double the number of species and total abundance than the control treatment, but only limited effects on stability. The high diversity treatment was essentially nested within the low diversity for assemblage structure and had similar or even slightly lower levels of species richness and abundance in most cases. Gymnotiformes and young-of-year were diverse and relatively abundant in macrophyte treatments contributing to the large differences in diversity between macrophyte and control treatments. Higher fish diversity in structured habitats compared to more homogenous habitats is likely associated with increased ecomorphological diversity to exploit heterogeneous microhabitats and resources provided by the macrophytes.

Au cours de son assemblage, une communauté végétale va subir de nombreux changements : immigration de nouvelles espèces de plantes possédant de nouveaux traits, disparition de certaines espèces de plantes avec d’autres traits, immigration de nouveaux organismes associés aux plantes (insectes, champignons…), changements de traits chez les espèces présentes… Tous ces changements sont susceptibles de modifier la décomposition de la litière produite par la communauté végétale. En effet, la décomposition de la litière est gouvernée par les traits des espèces végétales, par l’activité des organismes décomposeurs, et par le degré d’adaptation de ces organismes aux traits des espèces végétales. Cependant, les conséquences de l’assemblage de la communauté végétale pour la décomposition de la litière demeurent inconnues. L’objectif de cette thèse est de déterminer les conséquences de l’assemblage des communautés végétales prairiales sur la décomposition de leur litière, et ce à différentes échelles. Tout d’abord, nous avons étudié, très localement, les conséquences des plantes voisines que possèdent un individu pour la décomposition de sa litière (i.e. échelle intraspécifique). Nous avons distingué le cas où la litière de l’individu était seule, du cas où sa litière était mélangée à de la litière provenant d’autres espèces végétales. Puis, nous avons étudié les conséquences de l’assemblage sur la décomposition de la litière au niveau plus global de l’ensemble de la communauté végétale (i.e. échelle interspécifique). Enfin, nous avons exploré la rétroaction de la décomposition sur l’assemblage de la communauté. Deux grandes démarches expérimentales ont été développées, la première utilisant un dispositif de mésocosmes permettant de manipuler le voisinage local des individus, la seconde utilisant un dispositif Long Term Ecological Research (LTER) impliquant un vaste réseau de prairies avec différentes durées d’assemblage. À l’échelle locale, nos résultats indiquent qu’un individu qui possède des plantes voisines fonctionnellement dissemblables produit une litière plus décomposable et peut également abriter des décomposeurs plus efficaces. Lorsque la litière de cet individu est mélangée avec de la litière d’autres espèces, la décomposition du mélange est accélérée par des effets synergiques lorsque les plantes voisines sont évolutivement dissemblables et fonctionnellement éloignées du mélange. À l’échelle globale de l’ensemble de la communauté, nos résultats indiquent que tout au long de l’assemblage, de nombreux changements de traits fonctionnels des espèces végétales ont lieu (ratio C:N foliaire, teneur en matière sèche des feuilles, etc.) ainsi que des changements dans la composition de la communauté de décomposeurs (ratio C:N microbien). Ces changements impactent fortement la décomposition de la litière de la communauté prairiale mais s’annulent, maintenant le même taux global de décomposition. Enfin, nos résultats indiquent que plus la litière de couples d’espèces se décompose vite, notamment via des effets synergiques, plus ces espèces coexistent entre elles. Cette thèse met en évidence l’influence majeure de l’assemblage des communautés végétales prairiales sur la décomposition de leur litière, de l’invidu jusqu’à la communauté végétale toute entière. L’assemblage des communautés végétales peut donc influencer les processus écosystémiques d’après-vie tels que la décomposition de la litière. Cette influence se produit via les traits des plantes et l’activité de leurs décomposeurs. En retour, la décomposition de la litière impacte l’assemblage de la communauté végétale. La décomposition de la litière ne semble donc pas une conséquence collatérale des traits des espèces végétales, mais bien un élément important de leur stratégie écologique et de leurs interactions biotiques, situé au coeur d'une boucle de rétroaction avec les processus d'assemblage des communautés. / During its assembly, a plant community will be strongly modified: immigration of new plant species with new traits, disappearance of particular species with other traits, immigration of new plant-associated organisms (insects, fungi…), trait changes in existing species… All these changes are likely to drive the decomposition of litter produced by the plant community. Litter decomposition is indeed controlled by plant traits, activity of decomposer community, and adaptation of decomposer organisms to plant traits. However, the consequences of plant-community assembly on plant litter decomposition remain entirely unknown. This thesis aims at determining the consequences of plant-community assembly on plant litter decomposition, at distinct scales. First of all, we studied, locally, the consequences of neighboring plants on litter decomposition of plant individuals (i.e. intraspecific scale). We distinguished the case where litter of plant individuals was alone from the case where litter of plant individuals was mixed with litter from other species. Then we studied, more globally, the consequences of plant-community assembly on decomposition at the scale of the entire plant community (i.e. interspecific scale). Finally, we investigated whether plant litter decomposition feedbacks on plant-community assembly. We used two experimental approaches, the first one using a long-term mesocosm experiment for manipulating the local plant neighborhood of plant individuals, and the second one using of Long Term Ecological Research network involving grasslands with different time for assembly. At the local scale, our results indicate that plant individuals grown in functionally dissimilar neighborhood produce a more decomposable litter, and can also harbor more efficient decomposers. When the litter of these individuals is mixed with litter from other species, the decomposition of the litter mixture is accelerated by synergistic effects when neighboring plants are phylogenetically diverse, and functionnally dissimilar to the litter mixture. At the scale of whole plant community, our results show that numerous trait changes occur during assembly (leaf C:N ratio, leaf dry matter content…), as well as changes in the composition of the decomposer community (soil microbial C:N ratio). These changes strongly affect litter decomposition but offset each other, maintaining litter decomposition constant. Finally, our result show that the faster the decomposition of mixed-litter from two species is, the more both species coexist. This thesis demonstrates the major influence of plant-community assembly on plant litter decomposition in grassland ecosystems, from the scale of plant individuals to the scale of entire plant community. Plant-community assembly hence affects after-life ecosystem processes like litter decomposition. This influence occurs through plant traits and decomposer activity. In turn, litter decomposition feedbacks on plant-community assembly. Consequently, litter decomposition does not seem to be a collateral consequence of plant traits, but rather an important part of their ecological strategies and biotic interactions, participating to a feedback loop involving community assembly processes.

En aquesta tesi presento quatre capítols, en els quals es discuteix com diferents espècies subarbustives interactuen amb plantes coexistents, sota règims variables d’estrès. Aquesta recerca ha estat centrada en ecosistemes de gran valor ambiental, ecològic i de conservació, i alhora sensibles als canvis ambientals, en quatre regions distintes situades en zones de latitud o altitud elevades. Per ordre latitudinal, els ecosistemes estudiats han estat: 1) el límit supraforestal dels Pirineus Centrals, situats en una zona temperada; 2) el gradient de successió primària d’un ecosistema situat en una zona boreal, a Finlàndia,; 3) el límit supraforestal situat a la zona subàrtica de Lapònia; i 4) la tundra situada en una zona de l’alt àrtic, al nord‐est de Grenlàndia . Els dos primers capítols es basen en una aproximació experimental i se centren en investigar com els subarbustos més comuns trobats prop del límit supraforestal interactuen amb plançons d’arbres de les espècies que formen el límit del bosc en dos ambients contrastats, i.e. el límit supraforestal subàrtic, i el límit supraforestal dels Pirineus Centrals. A més, donat que els arbres que viuen prop del límit de distribució són molt sensibles als canvis ambientals , especialment a l’augment de temperatures durant l’estació de creixement (Körner 2003), també s’ha investigat com podrien afectar el creixement dels plançons d’arbre i la seva supervivència al llarg de l’ecotò, diferents escenaris de canvis ambientals. El primer capítol es titula ‘Les interaccions arbre‐arbust i els canvis ambientals dirigeixen la dinàmica supraforestal a la zona subàrtica’, i s’hi exposa l’experiment de tres anys de durada que es va executar al ‘Parc Nacional d’Abisko’, en els Scandes subàrtics, al nord de Suècia; les plantes estudiades varen ser plançons de Betula pubescens i el subarbust Vaccinium myrtillus. En aquest estudi es va certificar la importància de les interaccions arbust‐arbre, tant facilitatives com competitives, com a elements clau en la dinàmica supraforestal subàrtica. A més, vàrem demostrar que la gran sensibilitat dels plançons a l’escalfament tenia fortes implicacions per la dinàmica supraforestal, tenint en compte l’escenari previst d’augment de temperatures en latituds elevades. També es va observar que les interaccions complexes entre arbusts i herbívors són claus per preveure canvis futurs. El segon capítol es titula ‘Els plançons d’arbres situats als límits supraforestals Pirinencs i subàrtics mostren respostes semblants a la presència d'arbustos i a les simulacions de canvis ambientals’. Aquí es presenta una comparació entre els resultats de l’experiment presentat en el primer capítol i els obtinguts en un experiment paral∙lel dut a terme durant un període de temps semblant prop del límit supraforestal en el ‘Parc Natural de l’Alt Pirineu’, als Pirineus Centrals catalans. Aquest experiment es va basar en el mateix disseny factorial que l’anterior estudi, però amb diferents espècies (i.e. plançons de l’arbre Pinus uncinata i l’arbust Rhododendron ferrugineum). Segons ens consta, és el primer estudi que avalua experimentalment les respostes de les plantes vers diferents escenaris ambientals en un ecosistema d’alta muntanya als Pirineus. En aquest capítol es presenten alguns mecanismes que ajudaran a comprendre la variabilitat recentment observada de les respostes locals de límits supraforestals de zones temperades i subàrtiques com a resultat del clima canviant, i també identifiquem alguns paral∙lelismes que poden utilitzar‐se per generalitzar les respostes a gran escala dels límits supraforestals al canvi climàtic. El tercer capítol se centra en els efectes d’un subarbust dominant (i.e. Empetrum nigrum) en plançons de Pinus sylvestris al llarg del gradient de successió primària en un ecosistema boreal en una illa emergent de la badia de Bòtnia, a Finlàndia. Aquest capítol s’anomena ‘Un arbust ericoide exerceix la doble funció de reclutar pins i els seus simbionts fúngics al llarg d’un gradient de successió primària’. Aquí hem mostrat que els efectes facilitadors i competidors dels subarbustos determinen fortament l’establiment de plançons i la seva colonització fúngica al llarg d’aquest gradient de successió. Segons ens consta, aquests són els primers resultats que demostren que un arbust ericoide micorrízic pot millorar tant el desenvolupament de l’arbre hoste ectomicorrízic com el dels simbionts fúngics de l’arbre. L’estudi presentat al quart capítol es va realitzar al llarg d’un gradient de nivositat en un ecosistema extrem de tundra àrtica al ‘Parc Nacional del nord‐est de Grenlàndia’, el Parc Nacional més gran del món. El capítol es titula ‘Interaccions vegetals i composició de la vegetació àrtica al llarg d’un gradient de nivositat al nord‐est de Grenlàndia’. Aquest ecosistema és probablement el més sensible i fràgil de tots els ecosistemes estudiats en aquesta tesi, donat que s’espera que a la costa est de Grenlàndia s’hi esdevinguin canvis substancials en el clima com a resultat de canvis destacables en els règims de precipitació de neu i de les temperatures (Brown i Mote 2009). Aquí es va avaluar la riquesa d’espècies de plantes, així com els patrons d’establiment i composició de diverses formes de creixement existents en comunitats vegetals àrtiques associades a una variació de la cobertura de neu durant els mesos d’hivern. Aquest estudi ajudarà a preveure la diversitat potencial i els canvis en la vegetació a la zona de l’alt Àrtic, si els règims de precipitació de neu canvien en el futur com es preveu. / In this thesis I present four chapters, and in all of them I discuss how dwarf shrubs interact with co‐occurring plants under varying regimes of stress. This research involved ecosystems of great environmental, nature conservation and ecological value, yet highly sensitive to environmental changes, in four contrasting cold regions at high altitude or high latitude. Following a latitudinal order, the selected ecosystems were: 1) a temperate alpine treeline in the Central Pyrenees; 2) a primary succession gradient in a boreal ecosystem in Finland; 3) a subarctic alpine treeline in Lapland; and 4) a high‐arctic tundra in north‐eastern Greenland. The first two chapters are based on an experimental approach and focus on how shrubs commonly found near the treeline interact with tree seedlings of treelineforming species in two contrasting environments, i.e. in a subarctic forest‐tundra ecotone in Lapland, northern Sweden, and in a more southern, temperate forestalpine pasture ecotone in the Central Pyrenees. In addition, since trees living near their limit of distribution are very sensitive to environmental changes, especially to increased temperature during the growing season (Körner 2003), we also assessed how distinct environmental change scenarios may affect tree seedling growth and survival across the ecotone. The first chapter is entitled ‘Shrub‐tree interactions and environmental changes drive treeline dynamics in the Subarctic’, where we explain the three‐year‐long experiment performed in the Abisko National Park, in the subarctic Scandes, Northern Sweden; the species studied were Betula pubescens tree seedlings and the shrub Vaccinium myrtillus. In this study we showed the importance of facilitative and competitive shrub‐tree interactions as drivers of subarctic treeline dynamics. Furthermore, we demonstrated that the great sensitivity of tree seedlings to warming had strong implications for treeline dynamics under the predicted warmer scenario at high latitudes, and we identified that complex interactions between shrubs and herbivores are critical to predicting future changes. The second chapter is entitled ‘Similar tree seedling responses to shrubs and to simulated environmental changes at Pyrenean and subarctic treelines’. Here we presented a comparison between the results obtained in the experiment presented in the first chapter and those obtained in a parallel experiment performed during a similar period near the treeline in the Alt Pirineu Natural Park, in the Central Pyrenees, Catalonia. This experiment was based on the same factorial design but with different species (i.e. Pinus uncinata tree seedlings and the shrub Rhododendron ferrugineum). To our knowledge, it is the first study which experimentally tests the responses of plants to distinct environmental scenarios in a high mountain ecosystem in the Pyrenees. In this chapter we presented some mechanisms for understanding the recently observed variability of local responses of both subarctic and alpine treelines to currently changing climate while identifying some commonalities that can be used to generalise large scale response of treelines to climate warming. The third chapter focuses on the effects of a dominant dwarf shrub (i.e. Empetrum nigrum) on Pinus sylvestris tree seedlings along a primary succession within a boreal ecosystem on an uplifting island in Bothnian Bay, Finland. This chapter is called ‘An ericoid shrub plays a dual role in recruiting both pines and their fungal symbionts along primary succession gradients’. Here we showed that facilitative and competitive effects of shrubs markedly determined tree seedling establishment and their fungal colonisation along this succession gradient, but in this chapter we did not relate these findings to any environmental changes. As far as we know, we presented the first finding that an ericoid mycorrhizal shrub may enhance both the performance of the ectomycorrhizal host tree and the tree’s fungal symbionts. The study presented in the fourth chapter was performed along a snow‐depth gradient in an extreme arctic tundra ecosystem in the Northeast Greenland National Park, the largest national park in the world. The chapter is entitled ‘Plant interactions and higharctic vegetation composition along a snow‐depth gradient in NE Greenland’. This ecosystem is probably the most sensitive and fragile among the ecosystems studied in this thesis as the eastern coast of Greenland is expected to experience substantial changes in climate due to marked changes in snow precipitation and temperature regimes (Brown and Mote 2009). Here we assessed plant species richness, establishment and composition patterns in distinct growth forms occurring in common arctic plant communities associated with varying snow‐depth during the winter season. This study will help to predict potential diversity and vegetation changes in the high Arctic if snow precipitation regime changes in the future as anticipated.

In recent decades, it was shown that belowground competition for some plants may take form of the tragedy of the commons (TOC). In these plants, the competing neighbours invest more in root systems than would be appropriate for optimal nutrient uptake for the group and also more than they do when grown alone. However, there is also strong evidence that other species do not follow TOC, and tailor their root system to best nutrient exploitation irrespectively of competitor presence. The root investment strategy of these plants should correspond to the ideal free distribution (IFD). In my thesis I focus on two aspects:  I use game theoretical models to explore, whether those strategies can coexist within species and also whether different species can coexist if they have different strategy. From this model I draw predictions, which I test by meta-analysis.  Using Agrostis stolonifera as a model, I test assumptions on nutrient and neighbour perception, which underlie TOC and IFD models. I show that according to mathematical models, those two strategies can coexist in different species in a community, but cannot coexist within a species. Within a species, the TOC strategy should always dominate, once it appears. This can be extrapolated to macroevolutionary scale - once TOC occurs in certain clade, it should...

Prairie dogs (Cynomys sp.) are considered keystone species and ecosystem engineers for their grazing and burrowing activities (summarized here as disturbances). As climate changes and its variability increases, the mechanisms underlying organisms' interactions with their habitat will likely shift. Understanding the mediating role of prairie dog disturbance on vegetation structure, and its interaction with environmental conditions through time, will increase knowledge on the risks and vulnerability of grasslands. Here, we compared how plant taxonomical diversity, functional diversity metrics, and community-weighted trait means (CWM) respond to prairie dog C. mexicanus disturbance across grassland types and seasons (dry and wet) in a priority conservation semiarid grassland of Northeast Mexico. Our findings suggest that functional metrics and CWM analyses responded to interactions between prairie dog disturbance, grassland type and season, whilst species diversity and cover measures were less sensitive to the role of prairie dog disturbance. We found weak evidence that prairie dog disturbance has a negative effect on vegetation structure, except for minimal effects on C4 and graminoid cover, but which depended mainly on season. Grassland type and season explained most of the effects on plant functional and taxonomic diversity as well as CWM traits. Furthermore, we found that leaf area as well as forb and annual cover increased during the wet season, independent of prairie dog disturbance. Our results provide evidence that grassland type and season have a stronger effect than prairie dog disturbance on the vegetation of this short-grass, water-restricted grassland ecosystem. We argue that focusing solely on disturbance and grazing effects is misleading, and attention is needed on the relationships between vegetation and environmental conditions which will be critical to understand semiarid grassland dynamics under future climate change conditions in the region.