ENVIRONMENTAL HETEROGENEOUS DATA ANALYSIS OF WATER QUALITY TO TRACE POTENTIAL FACTORS AND IDENTIFY SOURCES IN AN URBANIZED RIVER

Environmental heterogeneity analysis was employed to characterize spatial variations of water quality, to trace potential factors, to identify sources, and to verify sampling sites in Puhe River, a typical urbanized river, in Shengyang City of northeast China. According to the hydrological systems and economic and social characteristics of Puhe River Basin, 34 sampling sites were selected, where 102 water samples were collected to measure water quality parameters. The level of water quality was the highest in the upstream, followed by the midstream and the downstream. The sampling sites were divided into three clusters by the hierarchical cluster analysis, i.e., the sites #1-#13, the sites #14-#23, and the sites #24-#34. It indirectly proved that the decreasing order of the water quality level could be upstream > midstream > downstream. Based on the principal component analysis, the potential factors included TN, NH3-N and TP in the whole river, and NH3-N had a significant correlation relationship with TP (r = 0.57, P < 0.01). It indicated that an amount of domestic sewage, fishery and livestock wastewater could be loaded into the river. In the upstream, TN, NH3-N, TP, and CODCr, were traced as potential factors, and NH3-N had not only significant correlations with CODCr, (r = 0.79, P < 0.01), TP (r = 0.57, P < 0.05) and TN (r = 0.57, P < 0.05) but negative correlation with oxidation reduction potential (r = -0.56, P < 0.05), which could be associated with domestic/industrial wastewater from wastewater treatment plants. In the midstream, TN, NH3-N, TP were sought as the potential factors, a strong significant correlation relationship between NH3-N and TN (r = 0.96, P < 0.01) was observed, which could represent fishery and industrial wastewater. In the downstream, the potential factors were TN and NH3N, and TN showed correlation (r = 0.63, P < 0.05) with NH3-N. It indicated that pollutants mainly derived from livestock and domestic wastewater. The discriminant analysis showed that the site #15 in the midstream was predicted into the downstream, but the site #27 in the downstream was predicted into the midstream. Hence, locations of the sites #15 and #27 should be adjusted in the midstream and the downstream, respectively. The heterogeneity analysis method was a useful tool to interpret multiscale datasets, to detect potential factors and to identify source, which could be available for monitoring and management of aquatic environmental system.