Validation of a quantitative method for estimating the indicator power of, diatoms for ecoregional river water quality assessment

The primary step in the development of diatom-based surface water quality monitoring tools is to estimate the indicator power of many diatom species for a range of ecological conditions. It is widely believed that the use of quantitative criteria greatly improves the accuracy of selecting the best indicator species. Nevertheless, the estimation of the indicator power of a species is currently only based on qualitative criteria. Therefore, the objective of this study was to develop a quantitative method for characterizing diatom traits that could be used to diagnose stressors of streams. We designed a quantitative method to classify environmental descriptors into three major ecological conditions. This quantitative method was combined with the numerical statistics developed by Dufrêne and Legendre to estimate the indicator power of diatom species using two datasets. In addition, using the nonparametric multiplicative regression (NPMR) model, we verified the fitness of the estimated indicator value of a taxon with its response to the particular environmental variable(s). The test dataset contained samples from the river systems of Ethiopia, and these constitute small portions of the three major watersheds of the Upper Awash, Blue Nile, and Omo-Gibe river basins, which are ideal regions for sources of variation in diatom indicators. The training dataset included samples from three Kenyan river basins. The pollution criterion of central tendency and variation was proven to be a useful quantitative approach for delineating ecological boundaries from the environmental data. We found that the method of estimating indicator power was sufficiently robust to quantitatively estimate the indicator value of abundant diatom species. It was also proven that the NPMR model can complement the method to obtain a clear visualization of the species optima to environmental variables and to identify rare diatom taxa that could serve as powerful indicators for the ecological categories.