Land use/land cover and hydrologic effects on North Shore tributary water quality

Abstract

Although they are inextricably linked, the impact of meteorological events and climate variations on water quality have not yet been fully explored. Streams considered for this study are facing increased developmental pressures and have, thus far, remained relatively pristine, although some are already listed as "Impaired". Through use of accumulated water quality data for Duluth and North Shore (Minnesota, USA) streams and GIS analysis of watershed characteristics, empirical models which explain the variability of water quality data, stratified by hydrologic regime, and mediated by landscape characteristics were analyzed. Multivariate statistical found correlations between measures of particulate-related and soluble water quality: sorted by hydrologic regime (i.e. snowmelt, storm events and baseflow); and mediated by landscape metrics. Measures of development at the local scale, including % impervious, were shown to be significantly correlated with increased particulate-related water quality metrics during high flow periods. Wetland cover at the whole watershed scale was negatively correlated with particulate-related water quality metrics, though at the local scale wetlands were positively correlated with soluble water quality metrics. Relationships between water quality metrics and measures of forest cover and stand type were shown to be strongly influenced by the scale at which analyses were performed. Regression analyses indicated that local land use/land cover (LULC) metrics best predicted particulate-related and soluble water quality metrics during high flow periods. During baseflow periods, whole watershed LULC metrics were the best predictors of particulate-related water quality metrics. Results suggested that road salt applied during the winter months may be stored in the soils or groundwater and released into urban trout streams during rain periods. The soluble water quality parameters were less clearly linked to just a few LULC characteristics and a there was no clear pattern for differences between whole vs. local watershed analysis. Soluble nutrients also exhibited seasonal variability. These analyses are an important step towards improving watershed planning and management policies by providing a tool for forecasting the impacts of land use decisions on water quality with known accuracy.