Sistema integrado de diagnose e recomendação e lógica fuzzy para análise do estado nutricional de Coffea canephora Pierre ex Froehner

Abstract

This work carried out to use a fuzzy classification (FC) as a methodology in the standardization of leaf nutrient based on the sufficiency ranges (FS) (low, adequate and high) for the analysis of the spatial distribution of the nutritional status of Coffea canephora. The study was carried out in a commercial crop of Coffea canephora Pierre ex Froehner, in Cachoeiro de Itapemirim-ES, in an area with an irregular mesh of 80 points for the sampling of leaves of plants to determine the contents the Nitrogen (N), Calcium ( Ca), Potassium (K), Magnesium (Mg), Sulfur (S), Phosphorus (P), Iron (Fe), Manganese (Mn), Boron (B), Zinc (Zn) and Copper (Cu). With the nutrient contents, descriptive analysis was performed and the nutritional diagnosis of plants was subsequently performed with the integrated diagnosis and recommendation system (IDRS) and the sufficiency range (SR). In FC, the input variables (INPUT) were the nutrient contents and the SR were transformed into linguistic variables and analyzed by mathematical models: triangular for Low and High and the trapezoidal model for suitable. The method used in defuzzification after the establishment of the rules was that of the centroid. The inverse interpolation method of the square of the distance (IDW) was used to build the maps. Pearson's correlation coefficient, the agreement index and the confidence index between the maps was determined in order to compare the original and defuzzied data. The nutrients S, Fe and Cu showed adequate levels, N, K and Mg levels in low sufficiency, Ca, P, Mn, B and Zn high levels. The correlation coefficients of Ca, Mg and Fe with productivity showed significant and positive correlation. Data from the input and output macronutrients showed a high correlation. The mathematical models chosen and the rules of inferences adopted in the fuzzy classification standardized on the same scale the values of the lower and upper limits of the adequate sufficiency range for macros and micronutrients, allowing the construction of the average map of the nutritional status of the crop. N and Zn were the limiting nutrients due to lack and excess, respectively, by IDRS. The average nutritional balance index (NBIm) showed a significant correlation (r = 0,65) wip crop yield (bags/hectare).