(1989) suggested that 200 m represents a maximum sampling reach length for wadeable streams. However, Lyons (1992) reported that although about 200 m represented a median asymptotic distance, in some cases sampling distances as great as 500 m were required to achieve a representative sample of fish species. Therefore, a sampling reach length of 300 m for wadeable streams is estimated as the maximum length of sampling reach necessary to ensure the collection of a representative sample of fish as part of NAWQA, yet minimize unnecessary sampling and reduced sampling efficiency as a result of crew fatigue. However, in relatively wide (greater than about 30 m) wadeable streams, a maximum reach length of 500 m should be considered. The maximum sampling reach length for fish community sampling in nonwadeable streams as part of the NAWQA Program is designated as 1,000 m, which is the same as that recommended by the Ohio Environmental Protection Agency (1987) and Plafkin and others (1989).
Sampling sites are generally chosen to represent the set of environmental conditions deemed important for controlling water quality in the study unit. Sites should represent combinations of natural and human factors thought to collectively influence the physical, chemical, and biological characteristics of water quality in the study unit and to be of local, regional, or national importance. Two distinct types of sampling sites are established as part of NAWQA--fixed sites and synoptic sites.
Fixed sites are typically located at or near USGS gaging stations where continuous discharge measurements are available. At these sites, broad suites of physical and chemical characteristics are measured along with characterizations of fish, benthic invertebrate, and algal communities. Three sampling reaches are established to represent environmental conditions associated with each fixed site. Three sampling reaches are the minimum necessary to establish a meaningful estimate of variability among sampling reaches. Major discontinuities in riparian or instream characteristics within or among sampling reaches should be avoided. The reaches should also include only those habitat features that truly represent the physical conditions of the stream. For example, if a stream is composed predominantly of a sequence of riffles and runs with one large pool, the pool should not be included in a sampling reach. Inclusion of the pool would incorrectly emphasize pool-dwelling species, thereby providing an estimate of the local fish community rather than an estimate of the fish community that is representative of the stream.
The distance between sampling reaches at fixed sites should be equal to the minimum sampling reach length (150 m for wadeable streams and 500 m for nonwadeable streams), to ensure the establishment of distinct sampling units. At a subset of fixed sites, multiple sampling reaches (minimum of three) are sampled in one year to assess the magnitude of sampling reach-to-reach variability. One sampling reach is sampled in each of three successive years to estimate short-term temporal variability.
Synoptic sites are typically nongaged sites where one-time samples of a limited number of physical and chemical characteristics are measured with the objective of answering questions regarding source, occurrence, or spatial distribution. When fish