Résumé
Coffee leaf rust (Hemileia vastatrix) caused major epidemics since 2012 in Central America. In order to prevent future epidemics, PROCAGICA program (Programa Centroamericano de Gestión Integral de la Roya del Café) aims to the creation of a warning system with an epidemic forecast component based on meteorological variables. Currently, recommendations to control this disease are based on costly treatment schedules. We hypothesize that it is possible to predict the outbreaks of coffee leaf rust and that dynamic of epidemics is the result of complex combinations of microclimatic variables acting at different times (times and durations). Our goal is to build three models: appearance of infections, first produced spores, intensification of sporulation. For this purpose, a trial was set up in Costa Rica in three sites at different altitudes and under different oceanic influences in order to cover important fluctuations of the mesoclimate. The microclimate is measured continuously using weather stations and a weekly monitoring of rust lesions is performed to know the dates of onset of the different symptoms corresponding to different stages of development of the epidemic. Preliminary results on the microclimate variables (nature and timing) that determine the onset of first symptom (emergence of non-sporulating lesions) will be presented and discussed.