Multivariate statistical technique, factor analysis/principal component analysis (FA/PCA), was applied to identify the human activity effects and natural processes on surface water quality in Low Acaraú basin, Ceará, Brazil. Water samples were collected in four different campaigns from Feb/2003 to Mar/2004, at seven sampling stations sited in low Acaraú watershed. Twenty-five physical, chemical and bacteriological parameters were examined in this study. The FA/PCA allowed the identification of the total variance. The best model was adjusted with four components, explaining 82.1% of the data variance. The first factor was mainly associated with suspended solids defining the erosion process and runoff; that is, a suspended solids component. The second factor was mainly associated with sewage, agricultural runoff and others non-pontual pollution; the third one was defined as an organic component and represents the anthropogenic activites; while the fourth was mainly correlated with soluble salt, expressing the weathering and leaching. Besides, the factorial analysis did not result in data reduction, since 18 of the 25 parameters have been kept to explain 82% of the data variance.
factor analysis; water pollution; watershed
