An Assessment of Regional Water Quality Based on the Time Series Global Principal Component Analysis and Entropy Weight Method: An Example of the Drinking Water Source in Xiamen

In view of the shortcomings of regional water quality health assessment methods, a new method combining time series global principal component analysis and entropy weight method was proposed to construct the water quality ecological health comprehensive index. The new evaluation tool had two advantages. Firstly,the section data was extended to the panel data to ensure the unity, integrity and comparability of the evaluation system. Secondly, the dimension of the original data was reduced to several unrelated principal components and then substituted into the entropy weight method to prevent the repetitive weighting of invalid information. The method was used to assess drinking water quality in Xiamen, The results showed that there was a significant negative correlation between the new method and the nutritional status index method ensured the accuracy of the method. The water quality ecological health comprehensive index of Bantou-Shidou reservoir, Jiulong River and Tingxi reservoir were 0.608, 0.372 and 0.774 respectively, which were rated as good, poor and good. From the perspective of time change trend, Bantou-Shidou reservoir presented a "V-shaped" inversion; Jiulong River had a more obvious seasonal variation showing an overall improvement trend; Tingxi reservoir had always been in the best position with small fluctuation.