Redundancy analysis for characterizing the groundwater quality in coastal industrial areas

Groundwater quality in coastal area has been an issue of interest because of excessive groundwater extraction for human use, for example, industrialization, irrigation, which can lead to saltwater intrusion. The study develops an integrated data analysis procedure based on multivariate statistics principal component analysis (PCA), hierarchical cluster analysis (HCA) and redundancy analysis (RDA), to determine the effects of key environmental conditions on the formulation of groundwater pollutants. This proposed method was demonstrated by analyzing groundwater quality monitoring data collected between 2011 and 2014 from four coastal industrial areas in Changhua county of Taiwan, namely Chuansing, Xianxi, Lukang and Fangyuan industrial parks. First, different environmental conditions in each industrial region were explored by PCA. The spatial hierarchy and spatial distribution of pollutant categories were then identified using HCA with the kriging method. Finally, the effect of environmental conditions on constitutive pollutants were identified with RDA. The three environmental patterns identified from the analytical results in Chuansing, Lukang and Xianxi were the salination factor (including conductivity and general hardness (GH)), water level and redox condition (including dissolved oxygen and oxidation–reduction potential). Fangyuan industrial park had only two patterns, namely salination (including conductivity and GH) and oxygen content (including DO and depth). The pollutant category indicated high concentrations of all pollutants in Chuansing and Fangyuan, and higher concentration of SO₄^2?, TDS, Cl^? in Xianxi, and of NH₃-N, Mn, Fe and TOC in Lukang. According to RDA results, salination caused the high concentrations of NH₃N, Cl^?, TDS in Chuansing, and of Cl^?, TDS and SO₄^2? in Xianxi and Lukang. Additionally, salination caused high concentrations of Fe in both Lukang and Fangyuan industrial parks in combination with those three pollutants. The redox condition was linked to high content of NO₃^? in Chuansing and Lukang, and of TOC in Xianxi. In Fangyuan industrial park, NO₃^? was also linked to high oxygen concentration. In summary, the combination of PCA, HCA and RDA enables the analysis of monitoring data to support policy decision-making.