Human activities strengthen the influence of deterministic processes in the mechanisms of fish community assembly in tropical rivers of Yunnan, China

Human-induced global alterations have worsened the severe decrease in fish biodiversity in rivers. To successfully reduce the pace of reduction in fish diversity, it is crucial to prioritize the understanding of how human activities impact the processes that shape and maintain fish diversity. Traditional fish survey methods are based on catch collection and morphological identification, which is often time-consuming and ineffective. Hence, these methods are inadequate for conducting thorough and detailed large-scale surveys of fish ecology. The rapid progress in molecular biology techniques has transformed environmental DNA (eDNA) technique into a highly promising method for studying fish ecology. In this work, we conducted the first systematic study of fish diversity and its formation and maintenance mechanism in the Xishuangbanna section of the Lancang River using eDNA metabarcoding. The eDNA metabarcoding detected a total of 159 species of freshwater fishes spanning 13 orders, 34 families, and 99 genera. The fishes in the order cypriniformes were shown to be overwhelmingly dominant. At different intensities of anthropogenic activity, we found differences in fish community composition and assembly. The analysis of the Sloan's neutral community model fitting revealed that stochastic processes were the dominant factor in the shaping of fish communities in the Xishuangbanna section of the Lancang River. We have further confirmed this result by using the phylogenetic normalized stochasticity ratio. Furthermore, our findings indicate that as human activities get more intense, the influence of stochastic processes on the shaping of fish communities decreases, while the influence of deterministic processes eventually becomes more prominent. Finally, we discovered that salinity positively correlated with fish community changes in the high-intensity anthropogenic sample sites, but all environmental factors had little effect on fish community changes in the low-intensity and moderate-intensity anthropogenic sample sites. Our study not only validated the potential application of eDNA metabarcoding for monitoring fish diversity in tropical rivers, but also revealed how fish communities respond to human activities. This knowledge will serve as a solid foundation for the protection of fish resources in tropical rivers.