Assessing how spatial variations of land use pattern affect water quality across a typical urbanized watershed in Beijing, China

Understanding the relationship between land use patterns and water quality in urbanized areas can provide insight into urban water quality management. In this study, a self-organizing map (SOM) and a linear mixed effects (LME) model were applied to explore the spatiotemporal patterns of nutrients and their relationship with land use characteristics in the highly urbanized Beiyun River watershed. The SOM classified 324 sampling data from 2014 to 2015 at 25 typical river sampling sites into nine clusters, with differences in land use patterns and seasonal (rainy and dry seasons) distributions. In the urbanized upstream reach of the river, concentrations of most pollutants were low, while in downstream sites the concentrations were higher for these pollution parameters. Nitrate-N (NO3--N), however, exhibited the opposite pattern, probably due to the sources of pollutants and biochemical processes. The explanatory ability of the LME model in the established urbanized area and newly developed area in the upstream portion of the watershed was much better than in the downstream area, probably due to the influence of upstream nutrient contributions and the point source inputs in the downstream area. These statistical results indicated that sewage disposal and the main urbanized land use types were the most important drivers of river contamination. Forest land positively or negatively influenced river water quality depending on its spatial location as well as the water quality indicators. Therefore, sewage disposal as well as spatial non-stationarity of land-use impacts should be taken into consideration to reduce water pollution in the Beiyun River watershed.