Different criteria for determining DRIS standards influencing the nutritional diagnosis and potential fertilization response of sugarcane

Integrated System of Diagnostic and Recommendation (DRIS) require establishment of norms for calculation of dual relationships between nutrients. Depending on the criteria used to establish the norms, nutritional diagnosis may vary. The objective of this study was to establish DRIS standards for sugarcane by different criteria and to evaluate the influence of these standards on the nutritional diagnoses. Four criteria were used: Nutrient relations with higher variance ratio between population of low and high productivity (C1); Nutrient relations with higher variance ratio and lower coefficient of asymmetry with partial transformation of Box and Cox (C2) and with total transformation of Box and Cox (C3); Nutrient relations with logarithmic neperian transformation (C4). The database consisted of 183 samples, in which 31 were in areas with high productivity (≥ 80 Mg ha-1) and 152 in areas with low productivity ( < 80 Mg ha-1). Sugarcane leaves in posicion (+3) were collected and contents of N, P, K, Ca, Mg, S, Fe, Zn, Cu, Mn and B were determined, according to the Kuijper system. The results implied that criteria for choosing nutritional relations with high ratios of variance for establishment of the DRIS norms were not adequate because the data were not standardized and presented a high probability of diagnosing nutritional imbalance. Criteria of the nutritional relations with high ratios of variance with a lowest asymmetry coefficient were more adequate because the data were normalized, providing similar nutritional diagnoses. Nutritional diagnoses were influenced by the criteria used to generate DRIS standards, diagnosing differences in positive response to fertilization.