The intensification of human activities has led to flow reduction and cut-off in most global rivers, seriously affecting riverine organisms and the biogeochemical processes. As key indicators of river ecosystems' structure and function, benthic biofilms play a critical role in driving primary production and material cycling in rivers. This research aimed to investigate the characteristics of microbial communities' complexity and stability during river flow reduction. Benthic biofilms were grown in artificial channels and subjected to eight gradients of flow reduction (represented by flow velocity from 0.4 to 110 cm/s). Biofilms' biodiversity, ecological networks and community assembly of bacteria, fungi and algae were investigated by high-throughput sequencing. Results showed significant differences in community composition and structure under different flow conditions. The eight flow gradients' microbial communities were divided into three groups: low, medium and high flows. The flow reduction led to significant decreases in bacterial and fungal communities' Chao1 index. Low flow conditions enriched the bacterial phyla Oxyphotobacteria, Alphaproteobacteria and Mollicutes, but significantly decreased the fungal phylum Chytridiomycota. Lowering flow reduced the fungal network's number of nodes and increased the algal network's number of edges. Cross-domain interactions network analysis showed a gradual increase in node and edge numbers with decreasing flow, while decreasing average path length. The neutral model predicted stochastic processes primarily drove biofilm community assembly, and that model's explanations decreased as the flow gradient decreased. The null model analysis revealed diffusion limitation as the most common stochastic ecological process for bacterial and algal communities, with reduced flow reducing heterogeneous selection and increasing diffusion-limited processes. This study provides an in-depth analysis of flow reduction's effects on biofilm communities' ecological networks and community assembly.
Read full abstract