A comparative study of single factor and multivariate statistical methods for surface water quality assessment
Water quality assessment is an essential part of water resources management. Different methods, including the single factor (SF) method and multivariate analysis, employed in assessing water quality can give different results. In this study, SF methods and multivariate analysis – including discriminant analysis (DA) and the distance function model (DFM) – were used to assess the water quality of Xinlicheng Reservoir in Northeast China, and the advantages and disadvantages for each method are demonstrated. Results from the SF method show water quality to be Class 4, while results from the DA and DFM methods show it to be Class 3, according to the Surface Water Standard in China. The concentration of total nitrogen is most polluted amongst all the evaluating factors (as a result of human activities), leading to a Class 4 result with the SF method. The alteration of an individual factor has little impact on the results of the DA and DFM methods, since they integrate the features of the all of the evaluating factors. Because anthropogenic pollution generally alters only a few evaluating factors, DA and the DFM are more useful for polluted water quality assessment than SF is. But DA and DFM methods should be used with sufficient consideration of their mathematical requirements to prevent misinterpretation of the results.