Adubação nitrogenada em soqueira da cana-de-açúcar e estimativa de produtividade por análise de imagem multiespectral

Abstract

The planning of a sugarcane plant has as main variable the production of the crop, which together with the planted area allows estimate productivity.Among the factors that make up productivity, fertilization plays an important role, and nitrogen is considered an essential component, recommendations are made based on expected productivity. In this sense, the objective of this study was to estimate sugarcane productivity by multispectral images in experiments with variations in liquid nitrogen (N) sources. Three experiments were installed in the areas of sugarcane mills, COFCO International (Sebastianópolis/SP), Guaíra Sugar Mill (Guaíra/SP) and São Francisco Power Plant (Quirinópolis/GO).I n three experiments were tested five sources of N, being they: Ucan (23% N + 4% Ca), Uran (32% N), Nitran (20% N) 20-00-00 + 4% Mg + 1% of Ca and liquid urea (20% N), at the dose of 100 kg ha-1 N. The fertilizers Ucan and Uran were also tested at doses of 33, 66 and 133 kg ha-1 of N, beyond the control treatment without N. The experimental plots were composed by 5 rows of sugar cane with 10 m in length and spaced 1.5 m apart from each other. The application of fertilizers was performed with costal sprayer pressurized cylinder and to CO2, being driven application to soil, on the line of the stump sprouting. Approximately 90 days after the application, we evaluated the number of tiller per meter and leaf samples were collected for the determination of the levels of calcium (Ca) and N, as well as flight was performed with manned aircraft remotely (ARP - Sensix SX2) loaded with multiespecral (Micasense sensor). Approximately 250 days after application, the harvest of the experimental plots, in order to determine the productivity of stalks (TCH) to evaluate the sources and model data. Were also carried out analyzes of raw material quality, determining the inch (%) of sugar cane and the total recoverable sugar (kg t-1), this latter being used to determine the produtivade of sugar per hectare (TAH). With the results of TCH and TAH, the agronomic efficiency index (AEI) of nitran sources was calculated, 20-00-00 + 4% Mg + 1% Ca and liquid Urea it was observed that the liquid Urea source presented the lowest index in all experiments, regardless of the variable analyzed (TCH and TAH). In general, liquid N sources tested resulted in agronomic efficiency índices higher than Uran sources applied in the dose of 100 kg ha-1 of N. The modeling of the estimate presented correlation of 60% (R²).