We used demographic methods to address one of the main challenges facing ecological science: forecasting the effect of climate change on plant communities. Ecological forecasts will be crucial to inform long-term planning in wildland management and demographic methods are ideal to quantify changes in plant abundance. We carried out our research in the sagebrush steppe, one of the most extensive plant ecosystems of Western North America. Our research intended to inform ecological forecasts on an exotic invader, cheatgrass (Bromus tectorum). Moreover, we investigated the general question asking: to what degree competition among plants influences the outcome of ecological forecasts on the effect of climate change? We carried out two field experiments to test the hypothesis that warming will increase cheatgrass abundance in the sagebrush steppe. This hypothesis was strongly supported by both experiments. Warming increased cheatgrass abundance regardless of elevation, neighboring vegetation or cheatgrass genotype. Moreover, we found cheatgrass was hindered by snow cover. Therefore, warming increases cheatgrass growth directly by increasing temperature, and indirectly by decreasing or removing snow cover. In our last experiment, we tested whether forecasts of climate change effects on rare species can ignore competition from neighbors. This should occur because rare species should have little niche overlap with other species. The lower the niche overlap, the less competition with other species. To test this hypothesis, we used a long-term data set from an Idaho sagebrush steppe. We built population models that reproduced the dynamics of the system by simulating climate and competition. Model simulations supported our hypothesis: rare species have little niche overlap and little competitive interactions with neighbor species.

Les invasions biologiques produisent des assemblages d'espèces souvent inédits et sources de nouvelles interactions spécifiques. L'impact écologique d'une invasion peut être : (i) indirect si l'espèce envahissante entre en compétition avec une ou des espèces résidentes, et (ii) direct si l'espèce envahissante consomme ou parasite une ou des espèces résidentes. Ces deux types d'impacts sont susceptibles de s'appliquer aux insectes envahissants du genre Megastigmus (Hymenoptera: Torymidae), qui exploitent généralement une niche étroite partagée avec d'autres espèces résidentes, et contribuent à d'importantes mortalités dans les populations d'hôtes du fait de leur spécialisation sur les graines. L'objectif de cette thèse est de tester ces prédictions à travers l'étude de l'impact de l'invasion en France de Megastigmus schimitscheki sur : (i) M. pinsapinis, un compétiteur résident, et (ii) les populations en expansion de son hôte obligatoire, le cèdre de l'Atlas (Cedrus atlantica).Un suivi spatio-temporel des populations françaises de Megastigmus spp. révèle que l'expansion rapide de M. schimitscheki est associée à un déclin important de M. pinsapinis. Un modèle mécaniste et statistique a été développé afin de déterminer la contribution relative de la variabilité temporelle de la disponibilité en graines et des divergences de traits d'histoire de vie dans la dynamique de coexistence-exclusion de M. schimitscheki et M. pinsapinis. Nos résultats montrent que la phénologie plus précoce de M. schimitscheki est le principal facteur déterminant le succès de l'invasion de M. schimitscheki et l'exclusion compétitive de M. pinsapinis dans les zones de sympatrie des deux espèces.Nous avons développé une approche théorique de l'impact indirect de M. schimitscheki sur C. atlantica basée sur les modèles de réaction-diffusion. Nos résultats montrent qu'un effet Allee résultant de la prédation des graines sur un front d'expansion peut accroître la contribution des hôtes situés à l'arrière de ce front au pool génétique de la population. Ce phénomène a pour conséquence de freiner l'érosion de la diversité génétique au cours du processus de colonisation.Les approches expérimentales et théoriques développées au cours de cette thèse montrent que l'invasion de M. schimitscheki a un impact indirect négatif sur la démographie d'une espèce résidente occupant la même niche écologique, alors qu'elle pourrait favoriser directement le maintien de la diversité génétique des populations d'hôtes en pleine dynamique de régénération naturelle. / Biological invasions produce novel species assemblages in communities that likely result in novel interspecific interactions. Ecological impacts of invasions may be: (i) indirect, if the invader enters competition with resident species, and (ii) indirect if the invader is a predator or a parasite of resident species. Both indirect and direct ecological invasion impacts are likely to apply to invasive wasps of the Megastigmus genus (Hymenoptera: Torymidae), which exploit narrow ecological niches that overlap with those of resident insect species, and contribute to high mortality levels in host plant populations due to their high specialization on seeds. This thesis aimed at testing such predictions by studying the impacts of the invasion of M. schimitscheki on: (i) its resident competitor for the seed resource M. pinsapinis, and (ii) the expanding populations of its obligatory host the Atlas cedar (Cedrus atlantica) in southeastern France.An extensive spatio-temporal survey of Megastigmus spp. populations indicated that the rapid expansion of M. schimitscheki was associated with a strong decline of M. pinsapinis. A mechanistic-statistical modelling approach allowed us to show that an earlier phenology of M. schimitscheki had a stronger explanatory power of both invasion success and the competitive exclusion of the resident than temporal variation in resource supply.A theoretical approach of the indirect impact of M. schimitscheki on C. atlantica expansion dynamics was based on reaction-diffusion models. We showed that an Allee effect resulting from seed predation at the expansion front of a host population may increase the genetic contribution of host individuals situated in the bulk of the expansion front. Interestingly, this phenomenon results in a limited erosion of genetic diversity during the expansion phase of the host population.The empirical and theoretical approaches developed showed that the invasion of M. schimitscheki in French cedar forests had a strong and negative indirect impact on the demography of a resident species, but, parallely, such invasion may directly favour the maintenance of genetic diversity in expanding host plant populations.

Assembly is a process that shapes the abundance and identity of species in a community. Niche and neutral theory explain assembly processes with mechanisms driven by either species differences, or functional equivalence and stochastic dispersal. In 2017 I sampled vegetation and environmental variables at 59 sites in the Noxubee National Wildlife Refuge and Tombigbee National Forest of Mississippi to explore forest understory community assembly. I developed and assessed a framework of predictions concerning general patterns and underlying mechanism. Evidence of dispersal limitation and functional equivalence were expected under neutral theory. Local environmental characteristics, surrounding landscape variables, and fire were significant determinants of beta diversity. Dispersal was not a strong predictor of beta diversity. I found evidence of both niche complementarity and functional equivalence, as well as niche differences among common vines and an introduced vine (Lonicera japonica). Overall, the results were more congruent with predictions expected under niche theory.