Networks and ordination analyses reveal the stream community structures of fish, macroinvertebrate and benthic algae, and their responses to nutrient enrichment

Rivers are perhaps the most altered ecosystem due to the human disturbances of water contamination, habitat destruction, hydrological regime change, overexploitation and exotic species introduction. Knowledge about the integrated effects of both natural environmental gradients and anthropogenic disturbances on multiple aquatic organism at a watershed scale is limited. In this study, the physical, chemical, hydrological and biological data were collected in May 2009 and 2010 at 281 sampling sites spread out wide stress gradients in the Hun-Tai River. We conducted canonical correspondence analysis (CCA) and network analysis to reveal the aquatic assemblage structures of fish, macroinvertebrates and benthic algae under the steep gradients of nutrient concentration and multiple environmental parameters. The results showed that habitat quality, elevation, water depth, stream width, dissolved oxygen, conductivity, TN, TP, and N:P ratio were identified as the significant parameters on the community compositions of fish, macroinvertebrate and benthic algae. In contrast to the historical effects of organic pollution, the eutrophication became the currently dominant human influencing factor on aquatic organisms in the Hun-Tai River. The nutrient parameters of nitrogen and phosphorus, and stoichiometric of N:P (mainly influenced by TP) were all identified as the significant parameters in the CCA biplots of fish, macroinvertebrate and benthic algae. Fish community was closely related with stream hydromorphological and nutrients parameters, however, macroinvertebrate and periphyton communities were driven by morphometry, chemistry and nutrients parameters. Network analysis was used to reveal the taxa associations and their correlations with nutrient parameters. By calculating the topological parameters of the networks, both the benthic algae and macroinvertebrate assemblages showed higher values of network centralization, heterogeneity, and average numbers of neighbors than fish in the network, those which usually indicated the higher association among taxa and stability of community structures. Huigobio chinssuensis, Cobitis granoei, and Ctenogobius brunneus were the core fish taxa, the subfamily Tanypodiinae, Epeorus and Ephemerella were the core macroinvertebrate taxa, and genera of Diatom, Fragilaria, Achnanthes, Navicula, Cymbella, and Nitzschia were the core benthic algae taxa in the network. A few taxa of fish showed significantly negative correlations with enrichment of nutrients and increasing the N:P ratio, and significantly positive correlations only existed with N:P ratio. In contrast to the fish network, macroinvertebrate and benthic algae taxa showed stronger and broader negative correlations to that of enrichment and N:P ratios in the network. Meanwhile, the positive correlations also broadly existed in the benthic algae network. This finding offers new clues of specific taxa of macroinvertebrate and benthic algae could be used as the indicator of stream eutrophication. This study provided new insights into integrated application of ordination method and network analysis in structuring the assemblage composition. And the network analysis also improve our understanding of individual interactions in the community and their relations to environmental stressors at the species level.