Abstract: With reference to the number of grades and threshold for each grade in a certain standard for water quality evaluation, a water quality evaluation method was established combining multivariate analysis with multi-factor evaluation. Through the regression analysis a corresponding function equation was established, and then by substituting sample data into corresponding relational expressions to build up a table of water quality evaluation indices. At the same time factor analysis was performed to allocate weight to each type of numeric data. Finally, by combining the use of the table of water quality evaluation indices with the weight for each type of numeric data, water quality evaluation indices were worked out for the sampling sites, and then the evaluation was done. With reference to the thresholds in the standard specific to lakes and reservoirs in the “Standard for Surface Water Environment Quality” (GB3838-2002), case studies were done of the Dianshan Lake and the Yuandang Lake. TN,TP, Chl-a and SD were monitored for four seasons consecutively through sampling at 29 sampling points in the studied area. Based on the monitoring data and the analysis with the established model, evaluation was done. Results show that the annual overall evaluation index of the Dianshan Lake is 5.727 and the Yuandang Lake is 5.418, both falling into Category V in water quality. The latter is a little better than the former. Among the four seasons, spring is the best with water quality generally in Category IV, and summer the worst with water quality in Category V minus.  

Key words: regression analysis, factor analysis, water quality assessment