Tamaño óptimo de parcela experimental para pasto Mombaça

Abstract

The objective of this study is to use statistical means in the optimal plot sizing for the culture of Mombasa grass. For this, data from a uniformity test will be analyzed in a total area of 100 m², with 400 basic experimental units (BEU), where each sampled plot had a usable area of 0.25 m². The experiment was conducted in the rural area of the municipality of Adelândia, Goiás. The climate of the region is classified as AW-tropical, with a dry season in the winter. In the implementation of the experiment, an application of chicken litter at 18 tons per hectare was made. For the selection of the experimental area, a place where the soil cover was more uniform was taken into account. However, due to the lack of management, the height of the grass within the experimental area was uneven. In view of this, it was necessary to perform a standardization cut throughout the plot. The cut samples collected were stored in paper bags and the production of green mass was measured. For the dry mass, 40 samples were randomly selected, where each one went through the drying process in a microwave. This drying process consists of weighing and marking the initial weight of the sample, putting two containers in the microwave for three minutes, one with water and the other with the sample. The methods used for this design were: maximum curvature model method of the coefficient of variation; linear segmented model method with plateau, and quadratic segmented model method with plateau. The analyzed models presented different plot sizes, with approximate values of 1 m2, 7.4 m2 and 14 m2, respectively. As a methodology for selecting the model that best fits the observed data, the Akaike information criterion (AIC) will be used. This is a relative measure of the quality of a statistical model’s fit. For a set of experimental data, the accepted model will be the one that presents the lowest value for AIC, that is, the lower the value for AIC, the better the model fits the data. We can conclude that the unsegmented method presented optimal plot size, with an area of 1 m2, plots of 4.3, a coefficient of variation of 27.2 and an AIC of 188. The linear segmented model with plateau and quadratic segmented model with plateau presented AIC values of 199 and 198, respectively.