PUBLICATIONS

• Publications by topic
  • Groundwater
  • Surface Water
  • Water Quality
  • Water Use

• Publications by date
  • Recent publications

Annual Water Data Report

USGS Publications Warehouse

USGS IN YOUR STATE

USGS Water Science Centers are located in each state.

A regression-based approach to assessing stream nitrogen impairment probabilities

Project Abstract
By Gerard McMahon, Song Qian, and Chris Roessler

Abstract

A recently completed National Research Council study of the Total Maximum Daily Load (TMDL) program of the U.S. Environmental Protection Agency recommends an increased use of models to assess the conditions of waters for which nutrient load limits may need to be developed. Models can synthesize data to fill gaps associated with limited monitoring networks and estimate impairment probabilities for contaminants of interest. The U.S. Geological Survey, as part of the National Water-Quality Assessment Program, the North Carolina Division of Water Quality, has developed a nonlinear regression model to estimate impairment probabilities for all river segments, or reaches, in North Carolina's Neuse River. In this study, a reach is considered impaired if the annual mean concentration of total nitrogen is greater than 1.5 milligrams per liter (mg/L), a concentration associated with stream eutrophication. A SPARROW (Spatially Referenced Regressions on Watershed attributes) total nitrogen model was calibrated using data from three large basins in eastern North Carolina, including the Neuse River. The model specifies that in-stream nitrogen flux is a function of a nonlinear relation of nitrogen sources, including point sources, atmospheric deposition, inputs from agricultural and developed land, and terrestrial and aquatic nutrient processing. Because data are managed in a geographic information system, the SPARROW model uses information that can be derived from the stream reach network about the spatial relations among nitrogen fluxes, sources, landscape characteristics, and stream characteristics. This presentation describes a process for estimating the proportion (and 90-percent confidence interval) of Neuse River reaches with a total nitrogen concentration less than 1.5 mg/L and discusses the incorporation of prediction errors into the analysis.