Variabilidade espacial da produtividade de soja e dos atributos do solo em sistema de plantio direto sob rotação de culturas

• Autor(a) principal: Bottega, Eduardo Leonel
• Data de Publicação: 2011
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/3576
• topic: Agricultura de precisão; Variabilidade espacial; Condutividade elétrica aparente do solo; Geoestatística; Precision agriculture; Spatial variability; Apparent soil electrical conductivity; Geostatistics; CNPQ::CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA

The objectives of this study were analyze the spatial variability of soil attributes in a system of agricultural production, analyze the spatial and temporal variability of productivity and its correlation with soil properties and study the influence of different sampling grids in modeling the spatial dependency of soil attributes. Data collection was performed on a farm located in the Brazilian cerrado, which is important in cereal production. The sample used in the study area has approximately 90 hectares. The sampling grid used consisted of two points per hectare. We collected 181 soil samples composed of four single samples each, representing the 0 - 0.20 m. At each sampling point were three readings of soil apparent electrical conductivity using a portable device. For the analysis of temporal variability in productivity yield maps were used representing three crops of soybean. In analyzing the influence of different configurations of sampling grids, the area was divided into cells of 2.9 and 4.7 hectares. We used a GPS receiver with differential correction for topographic geo-referencing of sampling points. The yield maps were acquired by a grain harvester CaseIH, model 8010, equipped with the yield mapping system AFS Pro® model 600. Statistical analysis and geostatistics were used the softwares Statistica version 7 and GS+ 9. Analysis of variability and temporal filtering of yield maps were performed using analysis tools of software ArcGIS 9.3. The apparent soil electrical conductivity showed significant positive correlation with the chemical phosphorus, phosphorus and zinc balance. Was not detected spatial dependence for variables apparent soil electrical conductivity, active acidity in water, aluminum, potential acidity, aluminium saturation and organic matter. The best estimates of values for unsampled locations were found for soil physical properties. Geostatistics techniques used made it possible to fit the theoretical models that best represented the experimental semivariance, thus enabling the construction of thematic maps of spatial distribution of attribute values of these soils. The apparent soil electrical conductivity was not significantly correlated with soybean harvest 2009/2010. Significant positive correlation between the soybean harvest in 2009/2010 was observed with the chemical attributes of soil potassium, calcium, total bases, effective capacity cation exchange and base saturation. Significant negative correlation between the soybean harvest in 2009/2010 was observed with the micronutrients iron, manganese and copper. Spatial dependence was detected soybean yields for the three crop seasons. The best estimate value of soybeans in areas not sampled was observed for the 2007/2008 crop productivity. In the analysis of temporal variability in productivity was found that 12.7% of the area showed unstable productivity. Using a sampling grid of one point for each 2.9 hectares was possible to detect the spatial variability of chemical and physical properties of soil, under the conditions studied. Using the sampling grid of one point for every 4.7 acres, could only detect the spatial variability of soil physical properties, within the conditions studied. All thematic maps of spatial distribution of soil attributes were significantly correlated with each other, regardless of the sampling grid used, however, have been softened as the number of sampling points was decreasing.