Understanding the effects of pollution on ecological communities and the underlying mechanisms that drive them will helpful for selecting a method to mediate polluted ecosystems. Quantifying the relative importance of deterministic and stochastic processes is a very important issue in ecology. However, little is known about their effects on the succession of microbial communities in different pollution levels rural ponds. Also, the processes that govern bacterial communities in polluted ponds are poorly understood. In this study, the microbial communities in water and sediment from the ponds were investigated by using the 16S rRNA gene high-throughput sequencing technology. Meanwhile, we used null model analyses based on a taxonomic and phylogenetic metrics approach to test the microbial community assembly processes. Pollution levels were found to significantly alter the community composition and diversity of bacteria. In the sediment samples, the bacterial diversity indices decreased with increasing pollutant levels. Between-community analysis revealed that community assembly processes among water and sediment samples stochastic ratio both gradually decreased with the increased pollution levels, indicating a potential deterministic environmental filtering that is elicited by pollution. Our results identified assemblage drivers of bacterial community is important for improving the efficacies of ecological evaluation and remediation for contaminated freshwater systems.

It has been widely recognized that organic fertilizer (OF) application under monoculture and continuous cropping can change the microbial community and increase forage biomass in the Qinghai–Tibet Plateau. However, as a commonly used grassland planting pattern, the way in which grass–legume mixtures respond to OF application remains unclear. To clarify application effects of organic fertilizer in the grass–legume mixtures, we conducted a field experiment at the Qinghai–Tibet Plateau and collected the rhizospheric and bulk soils to reveal their microbial community by using high-throughput sequencing and molecular ecological networks. It was found that OF application changed the microbial community and increased the forage biomass under monoculture. However, in grass–legume mixtures, we found that OF application did not promote the increase of forage (Gramineae) biomass (Student t-test: p > 0.05). By analyzing both prokaryote and fungal communities, it was found that OF application had a greater impact on bulk soil microorganisms than on those of the rhizosphere in grass–legume mixtures. Co-occurrence network analysis showed that the rhizosphere and bulk soil networks of grass–legume mixtures were significantly more vulnerable under OF treatment (vulnerability of prokaryotes in grass: 0.1222; vulnerability of prokaryotes in legumes: 0.1730; fungal vulnerability in grass: 0.0116; fungal vulnerability in legumes: 0.0223) than non-OF treatment (vulnerability of prokaryotes in grass: 0.1015; vulnerability of prokaryotes in legumes: 0.1337; fungal vulnerability in grass: 0.0046; fungal vulnerability in legumes: 0.0126), which indicated that OF application did not provide favorable conditions for microbial interactions in grass–legume mixtures. In addition, structural equation modeling showed that OF application had some significant negative impacts on soil physicochemical properties and the robustness of the prokaryote community. The robustness of fungi had a significant negative (p < 0.001) impact on forage biomass, but OF application had no significant (p > 0.05) direct impact on the forage biomass, which indicated that the OF did not promote forage biomass in grass–legume mixtures. These results suggest that the application of organic fertilizer is unnecessary for grass–legume mixtures, because it does not promote the interactions between rhizospheric microbes and forage.