Modelling the Effects of Flowering, Drought and Fruit Load on Mango Leaf Photosynthesis L. Urban and F. Normand P. Montpied INRA/CIRAD-Flhor UMR INRA UHP Ecophysiologie forestières Station de Bassin-Plat Centre de Nancy, Champenoux Saint-Pierre, La Réunion France France Keywords: Mangifera indica L., nitrogen, source-sink relationships, stomata Abstract Our general objective is to design carbon-based models of flowering and fruit quality. The first step of our procedure consisted in modelling carbon gains in mango, cv. Cogshall, using the biochemical model of C3 leaf photosynthesis of Farquhar et al. (1980). Data are presented about the temperature-response of the key parameters of photosynthetic capacity which was assessed in 17-month old mango plants. The temperature-corrected photosynthesis model was then coupled to the model of leaf diffusive conductance (g s ) of Ball et al. (1987). Simulation results obtained over one year on recently matured leaves from 13 to 13-year old trees show that our model is apparently robust for leaves from vegetative shoots, in orchard conditions characterized by mild seasonal changes in temperature and non-limiting water supply. However, our model failed to simulate A net accurately during the flowering period and periods of drought, as well as when the source-sink balance is modified as, for instance, in girdled branches at different fruit loads. Using results obtained over the 5 last years about the effects of flowering, fruiting and water- stress on photosynthesis in mango, we tested three modifications of our model: 1) the fixed parameters of the model of g s were replaced by adapted values for leaves from water-stressed trees and leaves close to inflorescences; 2) the total light-driven photosynthetic electron flux (J T ) was corrected to account for the existence of electron fluxes towards alternative sinks in leaves close to inflorescences; and 3) J T was corrected to account for the inhibiting effect of starch accumulation in leaves from girdled branches. Simulation results are presented and discussed in the perspective of our modelling approach. INTRODUCTION Mango is well-known for being a poor and irregular bearer. Moreover fruit quality and size seem to vary a lot as a function of time and space, making prediction and scheduling of production very difficult. The carbon status plays apparently a key-role in flowering and fruiting of mango trees (Whiley et al., 1989). Thus, developing a process- based model of carbon fluxes for mango trees is a crucial step in improving control of flowering and fruiting as well as fruit quality management through cultural practices. Of all the processes controlling carbon fluxes, photosynthesis is of prime importance since it represents the starting point of the supply and demand approach of reproductive and vegetative growth modelling. Biochemical models of leaf photosynthesis have been extensively used to compare photosynthetic performance among plant species and to analyse photosynthetic acclimation to high CO 2 concentrations or growth irradiance. Coupled to radiation transfer models, these models can be used to simulate photosynthesis at the individual plant or the canopy level. Unfortunately, such models do not satisfy the specific needs of fruit tree modelling, mainly because they do not integrate the effects of phenology and cultural practices, like irrigation. In tree crops, the effects of flowering and fruiting, which modify source-sink relationships, cannot be overlooked. This is even truer in tropical fruit tree crops, where the flowering and fruiting phases may exceed 6 months at the individual tree scale, depending on climatic conditions. In mango, under the conditions in Réunion Island, flowering usually extends over two months and fruiting 271