Espacialização e mapeamento da fertilidade em diferentes camadas do solo e densidades amostrais

Abstract

Detailed knowledge of the farm is a fundamental requirement to implement a management system that allowed localized application of inputs, however, poor sampling can reduce the detail levels that cause errors that prevent the adoption of this system. This work investigates the consequences of using different sampling densities by using descriptive statistics and geostatistics in mapping soil fertility chemical attributes of an Oxisol as well as the recommendations of fertilization and liming. Sampling densities were set to collect a sample every 1/4, 1/2, 1, 2 and 4 ha in 0-10 cm and 0-20 cm, where it was possible determining the spatial dependence of the attributes pH, H + Al, K, P, Ca, Mg, V and NC, as well as the range and intensity of these dependencies. It was also possible to determine the coefficients of overall accuracy and Kappa index between attributes, the differences in the amounts of fertilizers and among different sampling densities and traditional method of fertilization and liming. Descriptive statistics showed high soil fertility, normal data distribution and showed the errors caused by using the average instead of amplitude. Geostatistics demonstrated the need for a greater sample density than one sample every 1/4 ha to determine the spatial dependence of the attributes P, K in both layers. At sampling densities smaller and equal to one sample every two hectares was not possible to determine the spatial dependence in any attribute. The coefficients of overall accuracy and Kappa index showed loss of attributes mapping quality with the decrease of sample density, however, the density of a sample every 1 ha showed a sufficiency quality as a mapping for use in precision farming. With the use of precision agriculture, the total need of liming and potassium fertilization on both layers and phosphorus fertilization in the 0-10cm were higher than in the traditional method of recommendation and at 0-20 cm were lower. Precision agriculture has shown favorable results compared to the traditional method of fertilization.