Application of correspondence analysis in the assessment of groundwater chemistry

Correspondence Analysis (CA) was used to determine the sources and processes that may explain the variation observed in datasets of groundwater analyses. The following method was adopted: (1) based on the observation of "sympathies" and "antipathies" between loadings signs, correspondence factors were represented by parameters with some pertinent geochemical meaning: (2) the relation between factors and parameters then was checked by Multiple Linear Regression (MLR) where factors acted as independent variables and parameters as dependent variables. Sample scores of the selected parameters gave the ability to separate polluted from nonpolluted waters, identify areas where pollution is dominated by agriculture and areas where contamination is controlled by domestic effluents, and describe weathering and agricultural activities in the sampled area. The practical example presented here anticipates the impact of agriculture and urban pollution on the chemistry of 160 water sa,samples collected in a granitoid area at central Portugal (Fundao).