A special attention should be devoted to the water quality determinant factors, since the water supply has been decreased each year. The aim of this research was to define the similarity of determinant variables of water quality in the perennialized part of the Curu watershed, Ceará, Brazil. Multivariate analysis/Cluster Analysis was used for this investigation and it was proceeded during the wet and dry season. The samples were took in seven stations, spreading out along the rivers during four campaigns of collection (January, March, May and August of 2005), totalizing 28 samples. The water attributes analyzed were: pH, Ce, Ca2+, Mg2+, Na+, K+, Cl-, CO3-, HCO3-, SO4-, PO4(3-), NH4- and NO3-. Four homogeneous groups were identified. They were independent of their geographic position. They were identified expressing the influence of weather seasonality. The values of Sodium Adsorption Ratio (RAS) defined the dissimilarity of group 1 in relation to the others. The values of nitrate and phosphorus defined the similarity of group 2; while the Electric Conductivity (CE), RAS, sodium and chloride ions determined the non homogeneity between groups 3 and 4. Although the salinity level of water (CE>0,5d Sm-1) was not suitable to human consume, the Na+ and Cl- concentration do not present risks to human health. In relation to irrigation, only one group presented salinity risk from low to moderate, while SAR showed the same risk level for all groups.
water quality; sazonality; cluster analysis; multivariate analysis
