Jump to main content.

Landscape Model Development for Estimating Chemical and Biological Conditions of Streams and Potential Sources of Impairment in the Mid-Atlantic Region of the United States


Principal Investigators: K. Bruce Jones1, Anne C. Neale1, Timothy G. Wade2, Maliha S. Nash1, and James D. Wickham2

  1. U.S. Environmental Protection Agency, Office of Research and Development, Las Vegas, Nevada, USA

  2. U.S. Environmental Protection Agency, Office of Research and Development, Research Triangle Park, North Carolina, USA


Scientists and environmental managers alike are concerned about broad-scale changes in land use and landscape pattern and their cumulative impact on hydrological and ecological processes that affect surface water conditions. Of particular concern is the degree to which biophysical and landscape conditions at a range of spatial scales influences chemistry, physical habitat, and biological conditions of surface waters. High levels of nutrients and sediment in water can pose significant human health and ecological risks. A number of studies have shown strong relationships of water quality, water quantity, and run-off to landscape characteristics. A decrease in natural vegetation indicates a potential for future water quality problems. Other studies have shown that the land uses within a watershed can account for a relatively high percentage of the variability in stream and estuary water quality. Empirical studies have established the significant causal relationship between watershed characteristics and nutrient and sediment loads. Changes in landscape conditions in the riparian zone and in areas surrounding water quality sample sites may have a greater influence on water quality than broader scale, watershed conditions, although the importance of near-site, landscape conditions may vary, depending on the biophysical setting. Agriculture on slopes of greater than 3% increases the risk of soil erosion, and this can lead to increases in nutrient and sediment loadings to surface waters. High amounts of impervious surface and roads on watersheds also may result in high loadings of nutrients and sediment to streams, and atmospheric deposition can be a significant source of nitrogen in surface waters.

The objective of this project is to quantify relationships between landscape and biophysical attributes at several scales within watersheds and stream chemical, physical, and biological conditions. Spatially distributed models capturing these quantitative relationships will then be used to estimate stream conditions in areas not sampled by field-based monitoring programs. This type of capability will be critical to programs such as ReVA, and to States attempting to evaluate potential impairments to streams as part of Clean Water Act requirements.


The general approach is to pair stream samples acquired by EMAP, NAWQA, and other monitoring programs with GIS-derived metrics of land-surface conditions, including basic biophysical characteristics, human uses of the landscape, and land cover characteristics and patterns. Biophysical and landscape metrics are calculated for different locations and at different scales within the watershed. Stream variables include chemical measures, such as nitrogen and phosphorus, physical habitat, and biological community structure. A wide range of multivariate techniques are used to establish quantitative relationships.


Two studies have been completed. In one study of the Mid-Atlantic Region, quantitative relationships were established between landscape and biophysical metrics and stream nitrogen, phosphorus, and sediment. This project involved 148 stream samples from the NAWQA program and several region-wide landscape databases, including the National Land Cover Data Base (NLCD) from the early 1990s. Landscape metrics accounted for 83% of the variation in total nitrogen, 86% in total nitrate, 65% in ammonia, 73% in dissolved phosphorus, and 79% in suspended sediment. Results for this study were reported in: Predicting nutrient and sediment loadings to streams from landscape metrics: a multiple watershed study from the United States Mid-Atlantic Region.

Another study was completed using 472 stream samples from EMAP and STORET. A Regression Tree Analysis (RTA) was used to establish quantitative relationships between total stream nitrogen and landscape and biophysical conditions at multiple scales within the 472 watersheds or catchment areas representing each individual stream sample. The results showed strong relationships between total stream nitrogen and the amount of forest at the watershed scale, and the amount of wet nitrate deposition (estimated from a model that uses data from the EPA wet deposition monitoring network and the elevation derived from Digital Elevation Model or DEM data), although wet nitrate deposition was more important in north and northeastern part of the Mid-Atlantic Region than in other areas of the Region. The study also showed that high percentages of forest in the riparian zone can off-set impacts of wet nitrate deposition and relatively poor conditions at the watershed scale. The results of this study will be reported in: Multiple scale relationships of landscape characteristics and nitrogen concentrations in streams across a large geographic area.

Additional studies are underway to quantify relationships between biophysical and landscape attributes and stream biological condition in the Mid-Atlantic region. This involves similar approaches to those given above. In addition to EMAP data, this project will also use biological data collected by the State of Maryland in their state-wide stream survey. In addition to metrics involving composition and pattern, this study involves use of metrics of horizontal interactions and flows among the landscape elements. Data have been assembled for approximately 1800 sites and analysis are currently under way. Results from this study area expected by the end of FY05.


SAS and Excel databases with landscape metrics and stream variables are available for the two completed studies.


The following publications report results of two studies completed for this project:

Jones, K.B., A.C. Neale, M.S. Nash, R.D. Van Remortel, J.D. Wickham, K.H. Riitters, and R.V. O’Neill. 2001. Predicting nutrient and sediment loadings to streams from landscape metrics: a multiple watershed study from the United States Mid-Atlantic Region. Landscape Ecology 16:301-312.

Jones, K.B., Neal, A.C., Wade, T.M., Cross, C.L., Wickham, J.D., Nash, M.S., Edmonds, C.M., Riitters, K.H., Smith, E.R., and Van Remortel, R.D. 2005. Multiple scale relationships of landscape characteristics and nitrogen concentrations in streams across a large geographic area. Chapter 11, in Wu, J, Jones, K.B., Li, H., and Loucks, O. (eds.), Scaling and uncertainty analysis in ecology: methods and applications. Columbia University Press. In Press.

ORD Home | NERL Home
ESD Home | LEB Home
Send questions or comments to ESD Info Desk. (Library-lv@epa.gov)

Local Navigation

Jump to main content.