Disturbance and connectivity control biodiversity, ecosystem functioning and their interactions across connected aquatic and terrestrial ecosystems, that form a meta-ecosystem. In rivers, detrital organic matter (OM) is transported across terrestrial-aquatic boundaries and along the river network and decomposed on the way by diverse communities of organisms, including microorganisms and invertebrates. Drying naturally fragments most river networks and thereby modify organism dispersal and OM transfers across ecosystems. This may prevent organisms from reaching and consuming OM, generating mismatches between community composition and decomposition. However, little evidence of the effects of drying on river network-scale OM cycling exists. Here, we aim to examine the effects of fragmentation by drying on the structure of consumer communities and ecosystem functioning within interacting aquatic-terrestrial river ecosystems. We monitored leaf resource stocks, invertebrate communities and decomposition rates in the instream and riparian habitats of 20 sites in a river network naturally fragmented by drying. Although instream resource quantity and quality increased with drying severity, decomposition decreased due to changes in invertebrate communities and particularly leaf-decomposer abundance. Invertebrate-driven decomposition peaked at intermediate levels of upstream connectivity, suggesting that intermediate levels of fragmentation can promote the functioning of downstream ecosystems. We found that the variability in community composition was unrelated to variability in decomposition at sites with low connectivity and high drying severity, suggesting that such conditions can promote mismatches between community composition and decomposition. Decomposition instream was correlated to decomposition in the riparian area, revealing one of the first network-scale evidence of the links between ecosystem functions across terrestrial-aquatic boundaries. Our river network-scale study thus demonstrates the paramount effect of drying on the dynamics of resources, communities and ecosystem functioning in river networks, with crucial implications for the adaptive management of river networks and preservation of their functional integrity.

In the current context of biodiversity erosion, functional approaches have emerged to study community assembly mechanisms and to better predict the fate of plant species. Assessing patterns of trait variation should be a powerful tool to determine community assembly mechanisms. Yet, studies on trait variations and their consequences on individual performance are usually incomplete as they focus on a single ecological scale or filter, and do not include relationships between traits, resulting in a fragmented view of plant community assembly. We focused on the macrophyte communities living in particular freshwater ecosystems i.e. the ponds of the Iles Kerguelen, in the sub-Antarctic region. We measured different categories of traits (aerial, root, and clonal) on all occurring species to study trait variations across years, sites and phylogeny scale (between species and within species), and in response to multiple habitat abiotic and biotic variables. The consequences of these traits variations and the effects of their correlations for plant individual performance were also explored. Our results first highlighted a filter operating on the overall distribution of trait values within the region, whereas we observed a high amount of intraspecific trait variation allowing individuals to better resist to filters. Second, traits responses to biotic and/or abiotic factors were trait-dependent, and this combination of simultaneous trait responses should allow the plant as a whole to face several simultaneous constraints. Lastly, almost all traits have either direct or indirect effects on individual performance. As a conclusion, partitioning trait variance is a relevant approach to detect at which scale operate the most decisive processes in plant community assembly without scale dependency issues, and then orient further researches. Furhtermore, we plead to consider multi-traits approach, and several biotic and abiotic variables in future studies to better understand the effects of environmental changes on plant communities.