TEMPORAL AND SPATIAL VARIATIONS OF WATER QUALITY IN A TRANS-URBAN RIVER IN BEIJING, CHINA

摘要

The deterioration of urban water has inspired wide monitoring and assessment of water quality in urban rivers. In present study, different statistical techniques were used to analyze water quality datasets including 19 water parameters monitored twice a month along an urban river at 10 sites in 2009 and 2010, except the frozen period in Beijing. Composite water quality of upstream and midstream samples fell into II or III class of national surface water standards, but the downstream samples were beyond the worst class (V). The 10 months were assigned to three groups (wet, mean, and dry periods). TEMP, Cl, COD, and DO were discriminant variables of the temporal variation, with 86% correct assignments. All the discriminant parameters (except for Cl) showed a statistical difference (p<0.05) among the three periods. COD concentration was highest in the dry period, while DO was lowest and TEMP was highest during the wet period. The monitoring sites were classified into upstream, midstream, and downstream. TEMP, pH, DO, BOD_5, TOC, NH_4~+-N, Cl, SiO_4~4, Ca~(2+), Mg~(2+), K~+ and Na~+ were discriminant variables of the spatial variation with 92% correct assignment. All discriminant variables of spatial variation (except pH and DO) were highest in downstream, and DO was lowest in downstream and pH was highest in midstream. Principal component analysis/factor analysis identified four latent pollutants, explaining 76% of the total variance in river water quality: inorganic ions, nutrients and organic pollutants, physical contamination, and biological pollution. Source apportionment through absolute principle component score-multiple linear regression revealed that most water variables were primarily influenced by mineral weathering, urban runoff, domestic sewage, and unidentified sources.