English

Abstract

There are studies on the current state of understanding productivity models, focusing on the applicability of different estimation models for coffee tree productivity; the majority involving a considerable level of complexity. Thus, when searching for a simple and direct association between phenological characteristics and coffee productivity, doing research on this hypothesis is necessary. In this study, we aimed to validate a phenological model for coffee tree productivity by using phenological indices, under given edaphoclimatic conditions of Southern State of Minas Gerais, Brazil. We used 10 sample plots obtained from the municipalities of Lavras, Varginha, Carmo de Minas, Ijaci and Santo Antônio do Amparo. Plots were chosen based on the existing history about coffee productivity, which is over 40 sacks ha-1. Phenological data were collected in September-October, December-January and March to April of the harvesting seasons 2012/2013 and 2013/2014. The number of flowers and fruits were obtained at the fourth and fifth productive nodes of coffee plagiotropic branches sampled at the middle third of each coffee plant. Forty plants were sampled in each plot for the measurement of plant height and estimates of productivity phenological indices. Data regarding the observed production were obtained for models comparison and validation and, then, statistical tests were run. Results showed that these models are suitable for the coffee crop in the region under study. In addition, productivity phenological indices showed good correlation with the observed productivity. Key words: Coffee phenology, Coffea arabica L., prediction of productivity, agrometeorological modeling.