Relationships between Stream Biotic Integrity and Satellite-derived Land Cover Metrics in the Mid-Atlantic region, USA

The condition of an aquatic system is determined in large part by the condition and use of the surrounding drainage basin. Increased knowledge about the relationship between stream health, land use and landscape metrics is thus useful for both water and land resource management. Multimetric stream health indices of aquatic condition reflect water chemistry, biologic condition, and physical habitat. The impact of land uses immediately adjacent to streams is influenced, and can be mitigated by vegetated stream buffers. We report on a multivariate and regression analysis of the relationship between satellite-derived landscape characteristics and stream indices in Maryland. Whole watershed characteristics and riparian characteristics were obtained from a land use/land cover map developed from Landsat satellite data. Landscape variables were derived and analyzed relative to the benthic and fish indices of biotic integrity. Associations among the water and land variables were analyzed using stepwise regression, factor analysis, and regression trees. Percent impervious surface cover and tree cover (whether in the watershed or the buffer zone) explained most of the variation in watershed health. We also considered metrics accounting for the spatial configuration of the landscape, such as the amount of impervious cover in the flow path from surrounding lands to the stream. These were also found to be significant predictors of stream health, but were second-order predictors. Our results suggest that best management practices designed to improve stream water quality should focus on the amount of impervious area and tree cover in both the watershed and within the buffer zone.