Examining the impacts of changing land use on biological integrity in streams using Geographical Information Systems and statistical modeling

摘要

Land-use activities and land cover of a watershed influence chemical and physical properties of streams which may impact the biota of the aquatic ecosystem. This study was designed to investigate the impacts of the conversion of forests to urbanized land, including construction and development impacts of land use conversion on water quality and biological integrity of streams. Land use and land cover changes, habitat conditions, water quality, and sediment loadings were studied in eleven subwatersheds of the Reedy River Watershed in the South Carolina Piedmont Ecoregion. Physical and chemical datasets and land use data were used as inputs to build a multivariate model using stepwise multiple regression to predict two in-stream biological dependant metrics, including Biotic Index, Final Bioclassification, and two biotic metrics that measure richness and species composition. Univariate statistical analysis among 25 environmental variables and biological indices indicated that watersheds of stable land use differed significantly from watersheds undergoing conversion of forest to urban land uses. Significant correlations were observed between land use and particle size distribution, land use and in-stream physical parameters, and land use and in-stream chemical variables. Multivariate statistical analysis indicated that the principle environmental factors correlated with biological responses were conductivity, land use, and substrate distribution. Significant correlations between Biotic Index, Final Bioclassification, and compositional measures were found with conductivity, suggesting that conductivity may be a good indicator of land use changes and thus a good predictor of biotic indices in these subwatersheds.