Processes underlying the assembly of plant communities: The role of environmental filtering heterogeneity in a competition–dispersal experiment.

Question: Interspecific competition and seed dispersal are key mechanisms that may influence plant community assembly, but their relative importance remains difficult to disentangle as they can be affected by abiotic conditions (environmental filtering). We experimentally studied the balance between these processes in a Molinia wet meadow plant community in which abiotic constraints are mostly represented by flooding, while agricultural practices such as mowing prevent the dominance of a few competitive species. We hypothesized that (H1) mowing, by decreasing competitive exclusion and (H2) seed addition, by increasing immigration, both enhance plant species diversity, and that (H3) small variations in flooding intensity become secondary to these processes in determining species assembly. Location: French National Nature Reserve "Etang de Cousseau", Nouvelle Aquitaine, France. Methods: In a three‐year experiment, we added seeds of 17 species characteristic of Molinia wet meadows and managed above‐ground competition through mowing. We measured spatial variations in soil moisture as an indicator of the flooding regime. We recorded species relative cover, species richness and evenness annually, and measured light availability, linked with canopy openness, twice. Results: From the beginning of the experiment, small‐scale variations in soil moisture mainly affected species richness, composition and evenness, and canopy openness. Mowing significantly increased light availability and evenness, and affected the relative cover of several species, but not species richness. The addition of seeds did not influence any of the variables monitored during the experiment. Conclusion: Our study helps to disentangle processes underlying community assembly in wet meadows. It demonstrates significant and rapid effects of mowing on canopy opening and on community assembly, while seed addition had only a slight effect, possibly explained by an insufficient decrease in competitive interactions through mowing or unmeasured processes. It also shows that abiotic heterogeneity remains a primary driver of community assembly, even on a small scale. Finally, our results suggest that the respective weight of the different filters (dispersion, environment, competition) is subject to interannual variability or environmental stochasticity. [ABSTRACT FROM AUTHOR]