Streambank fencing along stream channels in pastured areas and the exclusion of pasture animals from the channel are best-management practices designed to reduce nutrient and suspended-sediment yields from drainage basins. Establishment of vegetation in the fenced area helps to stabilize streambanks and provides better habitat for wildlife in and near the stream. This study documented the effectiveness of a 5- to 12-foot-wide buffer strip on the quality of surface water and near-stream ground water in a 1.42-mi2 treatment basin in Lancaster County, Pa. Two miles of stream were fenced in the basin in 1997 following a 3- to 4-year pre-treatment period of monitoring surface- and ground-water variables in the treatment and control basins. Changes in surface- and ground-water quality were monitored for about 4 years after fence installation.

To alleviate problems in result interpretation associated with climatic and hydrologic variation over the study period, a nested experimental design including paired-basin and upstream/downstream components was used to study the effects of fencing on surface-water quality and benthic-macroinvertebrate communities. Five surface-water sites, one at the outlet of a 1.77-mi2 control basin (C-1), two sites in the treatment basin (T-3 and T-4) that were above any fence installation, and two sites (one at an upstream tributary site (T-2) and one at the outlet (T-1)) that were treated, were sampled intensively. Low-flow samples were collected at each site (approximately 25-30 per year at each site), and stormflow was sampled with automatic samplers at all sites except T-3. For each site where stormflow was sampled, from 35 to 60 percent of the storm events were sampled over the entire study period. Surface-water sites were sampled for analyses of nutrients, suspended sediment, and fecal streptococcus (only low-flow samples), with field parameters (only low-flow samples) measured during sample collection. Benthic-macroinvertebrate samples were collected in May and September of each year; samples were collected at the outlet of the control and treatment basins and at three upstream sites, two in the treatment basin and one in the control basin. For each benthic-macroinvertebrate sample: Stream riffles and pools were sampled using the kick-net method; habitat was characterized using Rapid Bioassessment Protocols (RBP); water-quality samples were collected for nutrients and suspended sediment; stream field parameters were measured; and multiple biological metrics were calculated.

The experimental design to study the effects of fencing on the quality of near-stream shallow ground water involved a nested well approach. Two well nests were in the treatment basin, one each at surface-water sites T-1 and T-2. Within each well nest, the data from one deep well and three shallow wells (no greater than 12 ft deep) were used for regional characterization of ground-water quality. At each site, two of the shallow wells were inside the eventual fence (treated wells); the other shallow well was outside the eventual fence (control well). The wells were sampled monthly, primarily during periods with little to no recharge, for laboratory analysis of nutrients and fecal streptococcus; field parameters of water quality also were measured.