Abstract This study examines the spatial variability of the factors obtained from the application of correspondence analysis to a hydrogeochemical data set. The goal was to synthesize the hydrogeochemical information using this multivariate statistical technique, by setting a series of factors which clarified the main properties of one aquifer. Then, a geostatistical framework to obtain a probabilistic assessment of groundwater quality was established. Experimental and theoretical semivariograms of the selected factors, considered as regionalized variables, were computed. These variographic information and factor values in the experimental sites were used in the ordinary kriging, which provides unbiased and linear estimates of the regionalized variables. These estimates were used to compile maps of the chosen factors, which explain their spatial distribution. The selected case study was the alluvial aquifer of Alto Guadalentı́n which is situated in southeast Spain, in the Internal Zones of Betic Cordilleras. These waters are chiefly SO 4 and Cl types, but HCO 3 facies are common in the central sector of the basin. High temperature, acid pH, problems of overexploitation and pollution by CO 2-gas characterise these waters. Available groundwater quality monitoring data were used to calibrate the numerical model. The present study focused on setting the main physical and chemical attributes and establishing the spatial pattern of groundwater quality and the temporal changes in this pattern.