Arranjo amostral para mapeamento de atributos do solo

• Main Author: Gelain, Evandro
• Publication Date: 2016
• Format: Doctoral thesis
• Language: por
• Source: Repositório Institucional da UFGD
• Download full: http://repositorio.ufgd.edu.br/jspui/handle/prefix/457
• topic: Distribuição espacial; Soja; Glycine max; Propriedade físico-química do solo; Spatial distribution; Soybeans; Soil chemicophysical properties; CNPQ::CIENCIAS AGRARIAS::AGRONOMIA

The conquest of the Cerrado for agriculture made it possible for Brazil to occupy the position of second largest soybean producer. This was only possible due to soil correction and fertilization practices, and the adoption of no-tillage. Since the 90s, the advent of precision farming tools made possible to handle the horizontal variability in cultivated soils. However, for the success of this tool, it is necessary to model this variability, and predict it reliably in non-sampled locations. This study was conducted on a field of Planalto farm, located in the city of Maracaju - MS, with the aim of studying different sampling arrangements for soil collection and its effects on geostatistical analysis and quality of soil attributes prediction maps: phosphorus, potassium, base saturation, pH, soil organic matter, and clay content. For this study we used different numbers of single samples to form the composite sample (1, 5, 10, 17 and 25) combined with a thinning out of sampling points to form different resolutions of sampling grids (187, 95, 47 and 23). Subsequently, the correlation between the chemical and physical soil properties and corn and soybean yield was evaluated using data collected in the sampling grid of 187 points made up of 10 single samples per point. The use of only a single sample per point affected the modeling of soil chemical properties, especially phosphorus and potassium. For a good estimate at sites not sampled it is necessary to use 8 single samples per point. The best estimates are obtained for the attributes suffering less influence of management practices, fertilization and soil correction. It was not detected spatial dependence for calcium, aluminum, potential acidity, sum of bases, iron and zinc, for the remainder of the attributes of the evaluated soil, moderate to strong spatial dependence was detected. Soybean yield showed a strong spatial dependence and correlated positively but moderately with the soil organic matter, phosphorus, magnesium and capacity of cation exchange.