The Indicator

This indicator reports both the yield and load of nitrogen from major rivers to the U.S. coastal ocean. The yield of nitrogen from major watersheds is defined as the pounds of nitrogen per square mile of watershed area that enters rivers and streams through discharges, runoff, and other sources. The load of nitrate, a common form of nitrogen, from major rivers is defined as the tons of nitrate carried to the ocean each year by the four largest U.S. rivers.

Nitrogen can cause significant water-quality problems by stimulating the growth of algae. Two key references provide additional information about how excess nutrients can cause problems in coastal waters. The National Research Council published a study on nutrient pollution in the coastal ocean (NRC 2000) and the National Oceanographic and Atmospheric Administration (NOAA) did a one-time study of actual conditions in the nation’s estuaries (National Estuarine Eutrophication Assessment, http://spo.nos.noaa.gov/projects/cads/nees/Eutro_Report.pdf).

The Data

View Data for Comparison Plot Data Source: U.S. Geological Survey, National Stream Quality Network (NASQAN)

Data Source: Riverine loads of total nitrogen were estimated using streamflow and water-quality data collected by the U.S. Geological Survey (USGS) as part of its National Stream Water Quality Accounting Network (NASQAN), its 1996–1999 National Water Quality Assessment (NAWQA), and its Federal–State Cooperative Program. A few of the stream gauges, most notably those at the mouth of the Mississippi River, are operated by the U.S. Army Corps of Engineers rather than the USGS.

Data Collection Methodology: Stream discharge is estimated by frequent measurement of water depth (stage), which is converted to discharge by use of a rating curve. Data are reported as daily averages. All water-quality samples are representative of the entire river cross-section (depth- and width-integrated) at the time of collection.

At the sites for which data are included in this report, samples were collected at least quarterly over the 4-year period 1996–1999; at most sites, approximately 15 samples were collected each year. A regression model relating concentration to discharge, day-of-year (to capture seasonal effects), and time (to capture any trend over the period) was developed using robust statistical techniques that made no assumption about the underlying statistical distribution of the data. One model was developed for nitrate plus nitrite concentrations (note that nitrite is usually much less abundant than nitrate, so it is normal to discuss the sum of nitrate plus nitrite simply as nitrate); a second model was developed for whole-water organic nitrogen plus ammonia for each station. These models were then used to make daily estimates of concentration, which were multiplied by the daily average discharge to yield the daily load. The daily load of total nitrogen was the sum of predictions of the two models.

Data Manipulation: For the maps, these daily loads were summed over the 4-year period to estimate the load for the entire period and divided by 4 to obtain the average annual load. The coefficient of variation of the average annual load is generally between 20 and 30%. The incremental load was then calculated as the difference between the output load that flowed from the watershed and the input(s) to the watershed. Outputs include the load at the downstream stations and, in the arid western areas, any decrease in runoff, because it was assumed that solutes accompanied any water that was lost to irrigation or transfers to other watersheds (i.e., piping water across watershed boundaries). The incremental yield (shown in the maps) is defined as the incremental load divided by the watershed area. The gray areas of the map are areas for which insufficient USGS data exist to calculate loads.

For the time series plots, the daily loads were summed to determine the annual loads shown in the figure. Note that most of the year-to-year variation in the loads is due to differences in runoff, with wet years having higher loads and dry years having lower loads.

Data Access: All USGS data are available at http://waterdata.usgs.gov/nwis/. This site includes the discharge and nutrient concentration data used in developing the models that produced the load calculations presented here, but the models themselves are not available. Further information on the NASQAN and NAWQA programs can be found at http://water.usgs.gov/nasqan/ and http://water.usgs.gov/nawqa/. The NASQAN Web site contains stream discharge data collected by the U.S. Army Corps of Engineers.

2003 Web Site Update: data for the Mississippi and the Columbia rivers were added in this update. Both of these data sets were obtained from the NASQAN web site (http://water.usgs.gov/nasqan with the guidance from USGS personnel.

2005 Web Site Update:Data on Nitrogen loads in the Columbia (2001-2002), Saint Lawrence (1995-1996), and Susquehanna (1999-2002) rivers were added in this update. These data were obtained directly from USGS. In addition to these new data, this update also provides revisions to previously reported data for all four rivers. Recently, the USGS applied new and improved models and methodology to this dataset; resulting in slight differences and two substantial changes: a new peak in nitrogen load carried by the Mississippi in 1993, and an upward rather than downward trend in the nitrogen load carried by the St. Lawrence since the mid-1980s (see comparison figures). The USGS is in the process of documenting changes in these models and methods; a link to this documentation (cited below) will be provided when available.

References

Aulenbach, B.T., Annual nitrate plus nitrite and total phosphorous loads for the Mississippi, Columbia, St. Lawrence, and Susquehanna rivers, 1971 – 2002, U.S. Geological Survey Open File Report.

National Research Council. 2000. Clean coastal waters: Understanding and reducing the effects of nutrient pollution. Washington, DC: National Academy Press.