Variable-rate seeding in soybean according to soil attributes related to grain yield

Defining the ideal soybean stand is essential for obtaining maximum grain yield. As a hypothesis of the work, it was questioned whether it is possible to vary seeding population of soybean based on the physical-chemical attributes of soil related to grain yield. The objectives of this work were (i) to identify soil attributes that help obtain the maximum grain yield grown in the Cerrado region of Brazil; and (ii) to generate a model to define the seeding rate according to these attributes. The experiment was carried out in the 2017/2018 and 2018/2019 crop seasons. In the first year, plots with different rate-seeding were sown on site one and site two. In the second crop season, site three was sown at a variable rate based on the results found in 2017/2018. The pH, calcium, magnesium, phosphorus, potassium, organic matter and clay contents, cation exchange capacity, base saturation and apparent electrical conductivity of the soil were evaluated. The results were subjected to Pearson's correlation network, followed by path analysis and decision tree. The variable-rate seeding suggested by the decision tree was applied on site 3 and the similarity between the soil attributes contained in the decision tree and the grain yield was verified using the relative deviation coefficient technique. The apparent electrical conductivity of the soil and the clay contents are the soil attributes that should be used to determine the soybean seeding rate. Areas with the highest values of electrical conductivity and clay contents should receive the higher soybean rate seeding in the Cerrado region.