![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Spatial Coverage
The spatial array of sampling sites in a given watershed or geographic region will determine to what extent the extrapolation of biological condition and water quality to areas beyond the exact sites can be made. Two primary guidelines can be identified for extrapolating biological assessment data to whole watersheds. First, the structure of aquatic assemblages in aquatic systems naturally and predictably changes with an increase in size of the ecosystem (Vannote et al. 1980). Thresholds in this continuum of change can be established through an analysis of regional databases. The biological condition at any particular site should only be used to represent areas of similar physical dimensions and flow characteristics (for streams and rivers). Likewise, lake or wetland size will influence the number of sites needed to adequately characterize these waterbody ecotypes. For instance, in small lakes, one site will generally be sufficient. In large lakes with multiple watersheds or in reservoirs with various zones (inflow, midsection, outflow, backwaters), a site representative of each basin or zone may be needed.
Second, the change in land use patterns along a stream gradient or lake shoreline should be considered. Changes from agricultural land use to urban centers, forested parkland, etc., would warrant different representative sampling sites. A waterbody with multiple dischargers may also require numerous sampling sites to characterize not only the overall biological condition of the waterbody but the impact of individual sources as well. The spatial density of sites within a watershed and the placement of those sites with respect to watershed area is another important consideration.
Available resources and the desired outcome of the sampling design are key determinants in achieving adequate coverage. Multiple designs are more likely to achieve the multiple objectives of State and Tribal water quality programs and will address needs beyond the determination of status. For instance, most States sample only a tiny percentage of their channel lengths or lake areas. To obtain unbiased estimates of condition in these situations, a probabilistic sampling design can be advantageous for statistical inference of status and trends to all waters from a relatively small sample. Sites or streams that represent the resource population of interest are selected randomly. Thus, one can extrapolate from the survey results to the entire population. Streams are identified by resource type, i.e., intermittent, perennial, etc., and further stratified by size to obtain a framework for randomizing the streams to be sampled in the resource population of interest. This design is cost effective in that the entire resource does not have to be sampled – only a representative number of streams. This sampling design was developed by USEPA's Environmental Monitoring and Assessment Program (EMAP) and has been used to assess the ecological status of waters on basin-wide, statewide, regional, and national scales (http://www.epa.gov/emap). Rotation of sampling among watersheds has also been found to be an effective approach for statewide monitoring over a specified period, which organizes the collection of ambient data in a timely fashion. However, where temporal and climatic variations dictate, this design may require modification.
For additional information on this topics, please visit
- Sampling Design in Developing Biological Indicators: Lessons Learned from Mid-Atlantic Streams
- Sampling Approaches in Statistical Primer
For more information, visit EPA's Aquatic Resources Monitoring - Monitoring Design and Analysis
References
Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell, and C.E. Cushing. 1980. The River Continuum Concept. Can. J. Fish. Aquat. Sci. 37:130-137.