Geographia Technica, No. 1/2008, pp. 74 - 81

SELECTION OF THE DETERMINATES TRUSSU RIVER WATER QUALITY FACTORS USING MULTIVARIABLE ANALYSIS

H. de Araújo Queiroz Palácio, Eunice Maia de Andrade, L. Araújo Crisostomo, A. dos Santos Teixeira, Ivam Holanda de Souza

ABSTRACT - Principal component analysis (PCA) was applied to evaluate and interpret a large water quality data set and apportionment of pollution sources factors with a view to get better information about water quality of the Trussu River valley. The investigation was carried out in the part of the valley where several farms with livestock activities and some villages are located. Water quality parameters were sampled from September/2002 to March/2004 at nine stations located along 24 km of the Trussu River for thirteen physical-chemical (2,223 observations). The PCA application resulted in three significant components, explaining 73.78% of the total variance of the data set. The first one, PC₁ (accounting for 48.36% of the variance) was mainly associated with sodium, Electric Conductivity, chloride, magnesium, sulphate and hydrogen-carbonate. It, basically, reflects ionic group of salts (mineralization processes). PC₂ (15.91% of the variance), was dominated by organic contaminations in water (NO₃-N and NH₄-N), suggesting anthropic activities. PC₃ was mainly contributed by pH e PO₄ and than, may be related to the effects caused in the water by non-point sources pollutants, such as agricultural runoff. This study suggests that PCA technique is a useful tool for identification of important surface water quality monitoring parameters.

Keywords: Water quality, Data reduction, Multivariate analysis

Palacio_TRUSSU RIVER WATER QUALITY.pdf