The Water Quality data at 104 monitoring stations of 49 rivers in Chiba Prefecture from 1976 to 1982 were analyzed by the Principle Component Analysis (PCA) method.
Before the calculation of PCA, the outliers were examined carefully, and the missing data were replaced by estimate with regression equation.
The correlation matrix was little changed by the replacement. Annual changes of correlation coefficients between the 6 items were found to be fairly little throughout this period (1976-1982).
The first component (z
1), derived from the PCA based on the correlation matrix of six items (DO, BOD, COD, NH
4-N, PO
4-P, SS), indicated the water pollution levels caused by human activities. The z
1 score was defined as water quality index (WQI).
Most of the rivers which flow in the northern west part in Chiba Prefecture were found to be seriously polluted in the estimation using the WQI. The annual changes of the WQI indicated that the water quality was improved at 17, became deterionate at 14 and unchanged at 69 points throughout this period.
View full abstract