Unbiased Approach to Diagnose the Nutrient Status of Red Guava (Psidium guajava)

Abstract

Fertilization of guava relies on soil and tissue testing. The interpretation of tissue test is currently conducted by comparing nutrient concentrations or dual ratios with critical values or ranges. The critical value approach is affected by nutrient interactions. Nutrient interactions can be described by dual ratios where two nutrients are compressed into a single expression or a ternary diagrams where one redundant proportion can be computed by difference between 100% and the sum of the other two. There are D(D-1) possible dual ratios in a D-parts composition and most of them are thus redundant. Nutrients are components of a mixture that convey relative, not absolute information on the composition. There are D-1 balances between components or ingredients in any mixture. Compositional data are intrinsically redundant, scale dependent and non-normally distributed. Based on the principles of equilibrium and orthogonality, the nutrient balance concept projects D-1 isometric log ratio (ilr) coordinates into the Euclidean space. The D-1 balances between groups of nutrients are ordered to reflect knowledge in plant physiology, soil fertility and crop management. Our objective was to evaluate the ilr approach using nutrient data from a guava orchard survey and fertilizer trials across the state of Sao Paulo, Brazil. Cationic balances varied widely between orchards. We found that the Redfield N/P ratio of 13 was critical for high guava yield. We present guava yield maps in ternary diagrams. Although the ratio between nutrients changing in the same direction with time is often assumed to be stationary, most guava nutrient balances and dual ratios were found to be non-stationary. The ilr model provided an unbiased nutrient diagnosis of guava.