How hydrodynamic conditions drive the regime shift towards a bacterial state with lower carbon emissions in river bends.

Hydrodynamic conditions influenced by river sinuosity may alter carbon (e.g., carbon dioxide and methane) emissions and microbial communities responsible for nutrient turnover. However, knowledge is lacking for the linkage between carbon emission and bacterial community in disturbed environments caused by river sinuosity. Here, the alternative states of benthic bacterial communities under the hydrodynamic conditions in river bends and the feedback to carbon emissions were investigated for the first time through the experiment of channels with different sinuosity combining hydrodynamic profiling, high-throughput sequencing and ecological theory. In this study, bimodal distributions combined with potential analysis showed direct evidence of bistability and demonstrated that increasing hydrodynamic conditions over a threshold (i.e., bottom velocity >0.73 cm s-1, turbulence kinetic energy >0.029 cm2 s-2) lead to a transition in benthic bacterial communities. Bacterial communities in high hydrodynamics (sinuosity of 1.4 and 2.2) exhibited lower carbon emissions (with averaged CH4 decreasing 0.04 μmol L-1 and averaged CO2 decreasing 2.48 μmol L-1). The bacterial communities in the high hydrodynamic group had higher α-diversity and more stable network structure based on topological properties of co-occurrence networks than in the low hydrodynamics. Homogeneous selection belonging to deterministic processes affected more on community assemblages of bacteria under conditions with both low hydrodynamics and high hydrodynamics, and a larger effect of deterministic processes on bacterial community assemblage was found in low hydrodynamics. Furthermore, the structural equation model showed hydrodynamic conditions induced by sinuosity regulated carbon emissions by directly and indirectly affecting the bacterial status. This study revealed the existence of alternative bacterial states under hydrodynamic conditions in meandering channels and explored the relationships between the bacterial states and carbon emissions, therefore providing insights into river reconstruction for an appropriate trade-off of urban river channel sinuosity.