Diatom distributions and species optima for phosphorus and current velocity in rivers from ZhuJiang Watershed within a Karst region of south-central China

The relationships between limestone bedrock geology (with extensive Karst topography), water chemistry and the community structure of benthic diatoms were examined for 35 riffle sites using multivariate analysis. The results showed that physical and chemical environmental variables had a notable effect on the biological diversity of benthic diatoms. Eighteen environmental variables were studied, including velocity (current flow), total phosphorus (TP), conductivity (COND), total nitrogen (TN), oxidation-reduction potential (ORP), chemical oxygen demand (COD), water temperature (WT), pH, silicon dioxide (SiO2), turbidity, major ions (Na+, K,(+) Ca2+, Mg2+, HCO3-+CO32-, Cl-, NO3-, SO42-). Two variables, TP and velocity were found to be the most important factors explaining diatom distributions. With further analysis using weighted average inference models, the results showed good predictive statistics for diatom composition and TP (R-2= 0.654, RMSE = 0.036; R-boot(2) = 0.483; RMSEboot= 0.048) and velocity (R-2= 0.641, RMSE = 0.378; R-boot(2) = 0.416, RMSEboot = 0.500). Indicator diatom species for TP and velocity were identified and assessed using weighted average optima and tolerances. Fifty-seven taxa were identified as biological indicators, which included Fragilaria vaucheriae, Achnanthidium convergens, A. lineare, Kolbesia suchlandtii and Nitzschia frustulum for high TP (> 200 mu g l(-1)) and Reimeria sinuata, Achnanthidium minutissimum var. scoticum, A. exile, Cymbella leptoceris, and Navicula cf. pseudokotschyi for low TP (50-97 mu g l(-1)). These taxa are useful as monitoring tools for eutrophication in ZhuJiang watershed within a Karst region of south-central China. Diatom richness followed a Gaussian distribution with the most taxa observed in waters with moderate levels of TP. This distribution supports the intermediate disturbance hypothesis which predicts that species richness and diversity will be highest under moderate environmental conditions. A Gaussian distribution was also observed for species richness under changing current flow. Measures of diatom diversity, richness, and species optima from the south-central highlands of China will be useful as indicator values for future water-quality assessments.