Benthic diatoms in USA rivers: distributions along spatial and environmental gradients

Aim To study large-scale patterns of benthic diatom assemblages in rivers, to assess the relative importance of environmental and geographical factors affecting their composition, and to evaluate the implications of these patterns for the use of diatoms as indicators of water quality, particularly nutrient enrichment. Location The United States Geological Survey (USGS) National Water Quality Assessment (NAWQA) Programme data set covers the conterminous United States. Methods We employed gradient analysis to understand the floristic structure of the data set and to discover major ecological gradients underlying variation in species composition at different spatial scales (entire US, Omernik Level 1 and Level 2 ecoregions). We used variance partitioning to separate the effect of environmental and spatial characteristics. Results At the national scale, three major ecological gradients were evident. The first was a complex 'downstream' gradient from fast-flowing, mostly oligotrophic, highland rivers to predominantly eutrophic rivers of high- and low-elevation plains. The second was a gradient integrating water mineral content and pH, which separated the soft and often more acidic waters of the humid eastern part of US from the alkaline waters of and western regions. The third obvious gradient was related to latitudinal and altitudinal variation of temperature. Up to one-third of the total explainable variation in species data was attributed solely to geographical factors not correlated with measured environmental characteristics. We present several examples of species with complex patterns of spatial distribution. Main conclusions Although environment consistently plays the most important role in structuring diatom assemblages in rivers, spatial factors also explain some variation in diatom distribution, especially at the continental scale. Most of the species that are confined to limited geographical areas are not yet described and await future taxonomic work. We show that selection of species which could serve as indicators of nutrient enrichment in US rivers is not a straightforward procedure. The existence of complex environmental gradients, and still poorly understood spatial patterns of species distribution, precludes attempts to develop uniform diatom-based metrics that would be applicable everywhere in the US. We advocate the development and calibration of metrics based on data sets collected from more limited geographical areas, and that include sites having relatively narrow ranges of environmental characteristics other than that which the metrics are designed to indicate.