Components of the spatial variability in site‑specific management in banana

Abstract

The objective of this work was to evaluate the spatial variability of banana production, in function of physical and chemical variables of soil and of farm physiographic characteristics, in order to select those with the greatest potential for use in site-specific management program. One hundred thirty productive units of banana clone Williams (Cavendish AAA) distributed in four lots and tree soils units of the farm were georeferenced. The bunch weight and root functionality were determined for each plant, as well as 35 physical and chemical soil variables. The spatial variability of the production was evaluated in function of these soil variables, using four different strategies: axes as covariates; both physic and chemical soil properties as covariates; analysis made by individual plot; and analysis made by soil unit. The individual plot study was the best strategy to model the spatial variability of banana production. This analysis model allowed to set soil variable groups which were significantly correlated and precisely explained more than 69% of the bunch weight inside the plots. These groups of variables are the ones with higher potential for the establishment of a site-specific management program.