Item talk:Q67412

Chapter B. Physical, Chemical, and Biological Responses of Streams to Increasing Watershed Urbanization in the Piedmont Ecoregion of Georgia and Alabama, 2003

As part of the U.S. Geological Survey National Water-Quality Assessment Program?s effort to assess the physical, chemical, and biological responses of streams to urbanization, 30 wadable streams were sampled near Atlanta, Ga., during 2002?2003. Watersheds were selected to minimize natural factors such as geology, altitude, and climate while representing a range of urban development. A multimetric urban intensity index was calculated using watershed land use, land cover, infrastructure, and socioeconomic variables that are highly correlated with population density. The index was used to select sites along a gradient from low to high urban intensity. Response variables measured include stream hydrology and water temperature, instream habitat, field properties (pH, conductivity, dissolved oxygen, turbidity), nutrients, pesticides, suspended sediment, sulfate, chloride, Escherichia coli (E. coli) concentrations, and characterization of algal, invertebrate and fish communities. In addition, semipermeablemembrane devices (SPMDs)?passive samplers that concentrate hydrophobic organic contaminants such as polycyclicaromatic hydrocarbons (PAHs)?were used to evaluate water-quality conditions during the 4 weeks prior to biological sampling. Changes in physical, chemical, and biological conditions were evaluated using both nonparametric correlation analysis and nonmetric multidimensional scaling (MDS) ordinations and associated comparisons of dataset similarity matrices. Many of the commonly reported effects of watershed urbanization on streams were observed in this study, such as altered hydrology and increases in some chemical constituent levels. Analysis of water-chemistry data showed that specific conductance, chloride, sulfate, and pesticides increased as urbanization increased. Nutrient concentrations were not directly correlated to increases in development, but were inversely correlated to percent forest in the watershed. Analyses of SPMD-derived data showed that bioassays and certain chemical constituents such as pyrene and benzophenanthrene, both PAHs found in coal tar, were strongly correlated with measures of watershed urbanization. Hydrologic variability metrics indicated that as urban development increased, streams became flashier, with characteristic high flows having shorter duration. The hydrologic effects associated with urbanization were greatest during the fall and least apparent during the winter. No correlations were observed between increasing urbanization and stream temperature or changes in stream habitat. Algal, invertebrate, and fish communities exhibited statistically significant changes as watersheds became increasingly urban, with the strongest responses observed in the invertebrate community followed by fishes, then algal diatom communities. Invertebrate communities were the most responsive to increasing urbanization with Ephemeroptera, Plecoptera, and Tricoptera taxa, especially Plecoptera (stoneflies) responding negatively and most strongly to increasing urbanization. Invertebrate communities were influenced more significantly by water quality, although significant responses to altered hydrology also were noted. In terms of the fish community, the percentage of cyprinids present in the stream was the only Index of Biotic Integrity metric that responded negatively to increases in watershed urbanization. Fish community response to urbanization was intermediate relative to algae and invertebrates with respect to significant metric responses as well as the overall community response to increasing urbanization. Measures of hydrologic variability were the most influential environmental variables affecting the algal community. Although sites were originally chosen to represent a gradient of increasing urbanization, a cluster analysis performed on the component metrics of the urban index categorized sites into four distinct groups. Multivariate analysis based on nonmetric MDS and related analyses of data ma