A review of stream nutrient criteria development in the United States.

Elevated nutrients and sediments are the main factors contributing to the poor biological condition measured in over 40% of US waters, highlighting the need for criteria that can aid management efforts to protect or restore the quality of US waters. A large amount of literature on nutrient criteria has been generated since the USEPA called for their development in 1998. Our objective was to examine this peer-reviewed literature to evaluate two main approaches for criteria development in lotic ecosystems: percentile rank and bivariate predictive statistical analyses. The 25th percentile approach has been examined broadly across USEPA-aggregate nutrient ecoregions, and we found that USEPA-suggested criteria for these aggregate ecoregions were often more conservative than criteria estimated using more current regionally focused data based on our compiled data set. Furthermore, 25th percentile estimates were often less than 75th percentile estimates based on reference sites, suggesting that 75th percentile estimates were not more conservative than 25th percentile estimates. Predictive approaches have focused on establishing linear and nonlinear relationships between water quality and algae, macroinvertebrate, and fish communities; attributing causation; and determining whether threshold points exist that can aid in nutrient criteria development. Most of the predictive approaches have occurred at the state or watershed level and may not be directly comparable to USEPA aggregate ecoregions. However, percentile method estimates often fell within the confidence interval of biological threshold criteria estimates, suggesting overlap and some consensus between the two main approaches.