[Total nitrogen and total phosphorus thresholds for epilithic diatom assemblages in inflow tributaries of the Three...

To study the relationships between in-stream total nitrogen (TN) and phosphorus (TP) loads and responses of epilithic algal metrics, nonparametric change-point analysis (nCPA) and thresholds indicator taxa analysis (TITAN) were used to detect possible TN and TP thresholds for algal chlorophyll a, diatom richness, Shannon diversity index, and diatom community composition for tributary rivers of the Three Gorges Reservoir (TGR). We found that nCPA simulated similar TP thresholds for all the four algal metrics (-0.03 mg·L^-1), with similar TN thresholds (-0.8 mg·L^-1) for all metrics but community composition. In contrast, TITAN provided specific nutrient thresholds for both sensitive and tolerant taxa within the whole community, and the threshold ranges covered thresholds for the other metrics. Additionally, nine (one sensitive and eight tolerant taxa) and ten indicator taxa (five sensitive and five tolerant taxa) for TN and TP were detected by TITAN, respectively. Therefore, we used the results of TITAN to define nutrient thresholds for the study rivers. That is, when in-stream TN concentration exceeded 0.382 mg·L^-1 or TP concentration exceeded 0.016 mg·L^-1(i.e. negative response threshold), sensitive taxa within river epilithic diatom assemblages would respond with decrease in density. When TN exceeded 1.298 mg·L^-1 or TP exceeded 0.065 mg·L^-1(i.e. positive response threshold), even tolerant taxa would be influenced and the whole assemblage would be changed significantly. Based on such criteria, we found that 87% sites had TN concentrations higher than the negative response threshold, with 22% sites being higher than the TN positive response threshold. As for TP, 94% and 14% sites exceeded the negative and positive response threshold, respectively. Our findings suggest that although most of the inflow tributaries of Three Gorges Reservoir were disturbed by human disturbance, yet the disturbance was not so extensive. Our findings provided significant information for managing river ecosystems of the TGR region.