Résumé
The effects of light environment, leaf-to-fruit ratio, and 'initial' fruit size on processes involved in mango fruit quality were investigated at the branch level, for a better understanding of the heterogeneity of mango fruit quality within a tree. This approach required analysing (i) the source/sink balance at the branch level, (ii) the water relations between branch and fruit, and (iii) the accumulation of sugars, organic acids and minerals in the flesh. Three models were derived from experimental data. They simulated fruit growth in water and dry matter content and storage of biochemical and mineral compounds in fruit flesh. The changes introduced in the original versions of these models are detailed. These models were integrated in a global model predicting individual fruit size and gustatory quality (i.e. sweetness and sourness). The global model of mango quality has been tested satisfactorily against experimental data of fruit size, sweetness and sourness. The global model at the branch level was then scaled up at the tree level. It predicted accurately the variability of mango quality observed within a tree by taking into account the fruit light environment, the 'initial' fruit size and the leaf-to-fruit ratio. The model was able to predict the distribution of fruit size and harvest date of a tree using as inputs the full bloom date of the orchard and distributions of 'initial' fruit size, light environment branches and competition for assimilate supply. These simulations showed that this model could have practical applications to manage fruit quality and plan harvest and marketing. (Résumé d'auteur)