During August 2000, responses of biological communities (invertebrates, fish, and algae), physical habitat, and water chemistry to urban intensity were compared among 30 streams within 80 miles of Boston, Massachusetts. Sites chosen for sampling represented a gradient of the intensity of urban development (urban intensity) among drainage basins that had minimal natural variability. In this study, spatial differences were used as surrogates for temporal changes to represent the effects of urbanization over time. The degree of urban intensity for each drainage basin was characterized with a standardized urban index (0-100, lowest to highest) derived from land cover, infrastructure, and socioeconomic variables. Multivariate and multimetric analyses were used to compare urban index values with biological, physical, and chemical data to determine how the data indicated responses to urbanization. Multivariate ordinations were derived for the invertebrate-, fish-, and algae-community data by use of correspondence analysis, and ordinations were derived for the chemical and physical data by use of principal-component analysis. Site scores from each of the ordinations were plotted in relation to the urban index to test for a response. In all cases, the primary axis scores showed the strongest response to the urban index, indicating that urbanization was a primary factor affecting the data ordination.

For the multimetric analyses, each of the biological data sets was used to calculate a series of community metrics. For the sets of chemical and physical data, the individual variables and various combinations of individual variables were used as measured and derived metrics, respectively. Metrics that were generally most responsive to the urban index for each data set included: EPT (Ephemeroptera, Plecoptera, Trichoptera) taxa for invertebrates; cyprinid taxa for fish; diatom taxa for algae; bicarbonate, conductivity, and nitrogen for chemistry; and water depth and temperature for physical habitat. The slopes of the responses generally were higher between the urban index values of 0 to 35, indicating that the greatest change in aquatic health may occur between low and moderate levels of urban intensity. Additionally, many of the responses showed that at urban index values greater than 35, there was a threshold effect where the response variable no longer changed with respect to urban intensity. Recognizing and understanding this type of response is important in management and monitoring programs that rely on decisive interpretations of variable responses. Any biological, physical, or chemical variable that is used to characterize stream health over a gradient of disturbance would not be a reliable indicator when a level of disturbance is reached where the variable does not respond in a predictable manner.