Planktonic and periphytic diatoms as indicators of stress on great rivers of the United States: Testing water quality and disturbance models

Planktonic and periphytic diatom assemblages were investigated as potential biological indicators of ecological conditions in the Ohio, Missouri and Upper Mississippi rivers. Diatoms were analyzed to determine which of the two assemblage types was most appropriate for monitoring environmental conditions. The comparative effectiveness of diatom data format was similarly evaluated, testing relative and absolute abundance data. Stressor gradients for the rivers were created using principal components analysis for suites of water quality variables, landscape disturbance variables, and combined variables. Diatom-based weighted averaging transfer functions were developed for these integrated stressor gradients using each habitat and data format. Representing the diatoms as relative abundance data provided optimal tracking of environmental stress, and both periphyton and phytoplankton showed potential as indicators of disturbance. These results reinforce the use of relative abundance data, often used for diatom water quality monitoring, over more labor intensive data formats. Diatom-based models were also created to infer water quality measures and to compare measured and diatom-inferred water quality to landscape stressors. In most instances, diatom-inferred water quality data were more strongly related to watershed-based stressors. This investigation supports the value of diatoms in assessing landscape stressors in large rivers as well as the more traditionally applied water quality conditions. The findings further suggest both planktonic and periphytic diatom assemblages may be used as bioindicators of river condition and may provide unique stressor response information.

• We studied diatom periphyton and phytoplankton great river environmental indicators. • Diatom-inferred water quality was significantly related to landscape stressors. • Relative abundance species data developed optimal disturbance models. • We tested inference models for integrated water quality and landscape disturbances. • Both phytoplankton and periphyton are suitable for great river indicators.