RESUMO
Quantifying the relationship between the land use and water nutrient concentrations would help guide land use management and water eutrophication control. Previous studies mainly focus on the quantitative structure of land use and lack a comprehensive characterization of the intensity and spatial distribution of land use. Taking the upper catchment of Miyun Reservoir as the study area, we analyzed the impacts of land use on surface runoff nutrient concentrations. We set 52 water quality monitoring sites within sub-basins of the entire upper catchment of Miyun Reservoir and collected water samples in each month. Using the remote sensing interpretation and spatial analy-sis techniques, land use information was extracted, which included the intensity, slope, distance to the rivers and sampling sites, and spatial arrangement. The land use information was used as independent variables for the multiple regression models to predict the changes of total nitrogen, total phosphorus and chemical oxygen demand concentrations. The consideration of the above four aspects of land use information significantly promoted the explanatory capability of models, with the coefficients of determination of multiple linear regression models changing from 0.294, 0.471 and 0.223 to 0.532, 0.685 and 0.489, respectively. We also compared the mean cumulative contribution rates of land use to nutrients and the corresponding cumulative area percentages with different flow path distances to river. Results showed that the flowing migration path distance to the river in the range of one kilometer was the key area to control water eutrophication in the upper catchment of Miyun Reservoir. Finally, we proposed three measures to control and regulate the water eutrophication, including optimizing the farmland fertilizer management, strengthening the livestock manure handling, building forested filter strips and riparian buffer zones.