Several studies have been accomplished lately to evaluate irrigation water quality in the semi-arid region of the Northeast Brazil. In some of these studies, the authors have adjusted some chemical characteristics such as calcium, magnesium, sodium, chloride and sum of cations as a function of electrical conductivity (EC) through empirical equations, however attention should be given to temporal and spatial variations. In this paper, the influence of water source, time of sampling, and soil type in the linear regression determination of electrical conductivity as a function of the ions present in the irrigation water was evaluated. A 562-analysis database originated from 55 rural properties was utilized. The water samples were analyzed for pH, EC, Ca2+, Mg2+, Na+, K+, Cl-, HCO3-, CO3(2-) , and SO4(2-). From January 1988 samplings in the properties were made during 411 days. The database was divided into 14 sampling times, three sources (well, river, and dam), and 10 soil types. For comparing the adjusted equations an identity model test was utilized. The adjusted linear equations with electrical conductivity as a function of the contents of calcium, magnesium, potassium, sodium, chloride, bicarbonate, carbonate, and sulfate varied significantly with time of sampling, water source, and soil type.
water quality; salinity; model identity test
