Diagnosis of nutrient composition in fruit crops: Major developments
MetadataShow full item record
Early methods of tissue diagnosis opposed the concept of critical nutrient concentrations supported by the laws of the minimum and the optimum to the more complex one of nutrient balances. But this proved to be a futile debate. Dual interactions and individual nutrient concentrations were later integrated into the Diagnosis and Recommendation Integrated System (DRIS). Methods of compositional data analysis (CoDa) corrected biases in DRIS on strong mathematical basis using centered (clr) and isomeric (ilr) log ratios. Log ratios were designed to adjust each nutrient to the level of others within the closed space or subspace of compositional data. Nutrient imbalance was assessed using a multivariate distance between given composition or ionome and a reference composition. In the near future, machine learning techniques and ionomics will allow predicting crop yield and quality using many more metadata collected in large data sets. Lines of research are proposed to improve models under international collaboration.