Topossequência e drenagem ambiental de solos : relações com a salinidade em perímetro irrigado do semiárido brasileiro

Abstract

During successive years the relationship between irrigation and drainage needs has been adopted as a rule for controlling soil degradation in irrigated crops. In order to evaluate the influence of toposequence and the presence and absence of subsurface drainage in the prevention of salinization of an irrigated Argissolo in the Senador Nilo Coelho Public Irrigation Project, Petrolina in Pernambuco State, Brazil, observation wells were installed to monitor weekly the groundwater fluctuation and monthly determine its chemical characteristics during the dry and rainy period for 18 months. For soil physicochemical characterization, samples were collected at depths 0-30, 30-60 and 60-90 cm, at the beginning of the experiment, at the end of the dry season and at six and eighteen months later, corresponding to the end of the rainy season. The obtained data were interpreted by analysis of variance using the F test of Snedecor e student t, by General Linear Models (MLG), having as response variable the hydraulic conductivity, and a set of fixed covariates (soil texture, soil density, macroporosity and saturated moisture), the akaike criterion (AIC) allowed to select the models. The variables electrical conductivity, exchangeable sodium percentage (PST), sodium adsorption ratio (RAS), cation exchange capacity (CTC) and hydrogen potential (pH) were analyzed by Tukey test of multiple comparisons between collection period, depth and the topossequence. The toposequence influenced the electrical conductivity of the saturation extract, the pH in water, the sodium adsorption ratio and the exchangeable sodium percentage, showed a tendency to accumulate in the low lot. Therefore, there was no tendency of salinization and / or sodification in the lots located in the highest toposequence quotas. The proposed models for estimating hydraulic conductivity satured (Ksat) can be applied with reasonable accuracy from sand content, silt content, soil density, macroporosity and saturation moisture, and the exponential model provided better results when compared to linear and quadratic models.