Pedofunções para retenção de água de solos do Rio grande do Sul irrigados por aspersão

Abstract

The determination of some soil physical characteristics, especially those related to water retention is difficult, requiring careful soil sampling, laboratories with sophisticated and high-cost equipment, highly trained personal and high demand of labor time. All of these factors greatly limit the number of specialized labs and make the access to that information very difficult. Alternatively, it is possible to estimate the points of the soil water retention curve using other soil physical attributes, based on pedotransfer functions (PTFs). The main objective of this study was to obtain mathematical models to estimate the soil water retention curve of some irrigated soils of Rio Grande do Sul State. These generated models can be used in soil physics labs to estimate the entire soil water retention curve. The work was carried out using 2235 soil samples, collected in three soil depths (surface, intermediate and lower layer of soil profile) in irrigated areas of Rio Grande do Sul State. The database of soil physical analysis bellongs to the Sistema Irriga lab of the Federal University of Santa Maria. That database consists of information of soil texture (sand, silt and clay content), density (bulk and particles), soil porosity (macro, micro and total), besides the information of soil water content at the matric potential of 0, -1, -6, -33, -100, -500 and -1500 kPa. The equations were obtained through the stepwise option of SAS statistical program. It was used part of the data set (60% of samples) to generate the equations and the remaining (40%) to validate the equations generated for the three soil layers of the soil profile. PTFs were generated to estimate the soil water content at a matrix potential of 0, -1, -33, -100, -500 and -1500 kPa, in the superficial, intermediate and lower layers of the soil profile. All the generated equations presented as variable information of soil textural and structure attributes. The structural attributes showed a water retention at highest matrix potential. At lower matrix potential, the soil textural attributes were better correlated with soil water content. It was observed that increasing the homogeneity of the soil parameters data improve the accuracy in estimating the soil water retention curves using the generated equations. In general, the generated PTFs resulted in excellent predictive capacity of the soil water retention curves and may be used in soil physics labs to reduced labor time and cost, since the measured soil physical attributes are within the tolerated limit of each variable.higher correlation with the soil