Diatoms are better indicators of urban stream conditions: A case study in Beijing, China

- Diatoms were more sensitive than water qualities in assessing stream conditions.
- Diversities were higher in reference and urban upstream than urban downstream.
- Diatom assemblages were very different in reference, urban upstream and downstream.
- Factors affecting diatom species distribution varied by stream sites.
- Diatom index developed for stream monitoring should be based on the local species.

Urbanization dramatically affects hydrology, water quality and aquatic ecosystem composition. Here we characterized changes in diatom assemblages along an urban-to-rural gradient to assess impacts of urbanization on stream conditions in Beijing, China. Diatoms, water chemistry, and physical variables were measured at 22 urban (6 in upstream and 16 in downstream) and 7 rural reference stream sites during July and August of 2013. One-way ANOVA showed that water physical and chemical variables were significantly different (p < 0.05) between urban downstream and both reference and urban upstream sites, but not between reference and urban upstream sites (p > 0.05). Similarly, structural metrics, including species richness (S), Shannon diversity (H′), species evenness (J′) and Simpson diversity (D′), were significantly different (p < 0.05) between urban downstream and both reference and urban upstream sites, but not (p > 0.05) between reference and urban upstream sites. However, diatom assemblages were very different among all sites. Achnanthidium minutissima was a consistent dominant species in reference sites; Staurosira construens var. venter and Pseudostaurosira brevistriata were the dominant species in urban upstream sites; and Nitzschia palea was the dominant species in urban downstream sites. Clustering analyses based on the relative abundance of diatom species, showed all the samples fit into three groups: reference sites, urban upstream sites, and urban downstream sites. Canonical correspondence analysis (CCA) and Monte Carlo permutation tests showed that concentration of K+, EC, TN, Cl− and pH were positively correlated with relative abundance of dominant diatom species in urban downstream samples; WT and F− were correlated with reference and urban stream diatom composition. Our results demonstrate that the composition of diatom species was more sensitive to urbanization than the water physical and chemical parameters, and that diatom assemblage structure metrics more accurately assessed water quality. Some species, such as Amphora pediculus and Cocconeis placentula were among the dominant species in low nutrients stream sites; however, they were considered to be high nutrient indicators in some streams in USA. We suggest using caution in applying indicator indices based on species composition from other regions. It is necessary to build a complete set of diatom species data and their co-ordinate environment data for specific regions.