"Metacommunity theory and its empirical tests have yielded novel insights into how ecological communities are structured at multiple spatial scales. However, much still remains to be understood regarding the spatial structure of food webs. On the other hand, a well-known effect of chained predator-prey interactions is the existence of trophic cascades, which have almost never been studied in a spatialized context. Recently, it has been shown (by the outgoing host group) that an intermediate level of dispersal among communities tends to buffer the effects of trophic cascades when top predators periodically colonize communities. In this context, the proposed two-year project aims at (i) building a comprehensive model capable of capturing the main effects of dispersal on ecosystem functioning and stability when trophic cascades are present, and (ii) conducting mesocosm experiments on freshwater plankton food webs to confirm theoretical results on the determinants of the buffering effect of dispersal. The model will be built in two phases, first as a purely deterministic description of spatial food webs, and extended in a second time to account for demographic stochasticity of the upper trophic levels. The model will be developed in collaboration with researchers from both host groups, and will span both years. Mesocosm experiments will be conducted in an already existing experimental setting, during the first year. This part of the project will be accomplished in the outgoing host group, in collaboration with researchers skilled in experimental biology and statistical analyses. The project will help the fellow acquire new skills, notably in experimental ecology and hone other proficiencies (writing, presentation, management,…). The project will also allow the fellow to establish a long-lasting collaboration between the two host groups and to produce novel important results that will improve his prospects as a researcher in Europe."