Main predictors of periphyton species richness depend on adherence strategy and cell size

Abstract

Periphytic algae are important components of aquatic ecosystems. However, the factors driving periphyton species richness variation remain largely unexplored. Here, we used data from a subtropical floodplain (Upper Paraná River floodplain, Brazil) to quantify the influence of environmental variables (total suspended matter, temperature, conductivity, nutrient concentrations, hydrology, phytoplankton biomass, phytoplankton species richness, aquatic macrophyte species richness and zooplankton density) on overall periphytic algal species richness and on the richness of different algal groups defined by morphological traits (cell size and adherence strategy). We expected that the coefficients of determination of the models estimated for different trait-based groups would be higher than the model coefficient of determination of the entire algal community. We also expected that the relative importance of explanatory variables in predicting species richness would differ among algal groups. The coefficient of determination for the model used to predict overall periphytic algal species richness was higher than the ones obtained for models used to predict the species richness of the different groups. Thus, our first prediction was not supported. Species richness of aquatic macrophytes was the main predictor of periphyton species richness of the entire community and a significant predictor of the species richness of small mobile, large mobile and small-loosely attached algae. Abiotic variables, phytoplankton species richness, chlorophyll-a concentration, and hydrology were also significant predictors, depending on the group. These results suggest that habitat heterogeneity (as proxied by aquatic macrophytes richness) is important for maintaining periphyton species richness in floodplain environments. However, other factors played a role, suggesting that the analysis of species richness of different trait-based groups unveils relationships that were not detectable when the entire community was analysed together.