Metodologia para estimativa de zonas de potencial produtivo a partir de dados de produtividade

Abstract

The methodology proposed herein for identifying potentially productive zones from yield data captured by harvester onboard sensors aims to establish a viable and easy-to-implement method for defining management zones by running statistical procedures on data from the harvest monitor. To do this, yield data from maize (2018 winter/second growing season) and soybean (2019 growing season) were converted into ɀ-score values and compared at a 99.8% confidence interval of standard normal distribution ɀ. Simultaneously, the degree of linearity was evaluated and Jackknife resampling, for removing data outside the range (outliers) established by the ɀ table (<-3.09 and >3.09). Next, yield score-ɀ algebraic mapping was performed to obtain a mean crop map, then applying three classes from the probability intervals of a plus and minus deviation, resulting in a map of potentially productive zones (below average, average and above average yield). Using this method, 5.72% of the area exhibited low yield potential, 90.71% average potential and 3.57% high yield potential. This analysis method was easy and quick to perform and provided summarized information, facilitating additional field surveys and providing a basis for decision-making.