Eficiência do modelo Ceres-Maize na simulação do desempenho de híbridos de milho

Abstract

The CERES-Maize model was developed for simulating the development and performance of corn and has been used as a tool for planning and decision making by farmers in several countries. The objective of this study was to evaluate the efficiency of the CERES-Maize model in the simulation of agronomic characteristics of corn hybrids on tropical conditions. A field experiment was conducted with five hybrids (AG7000, AG8060, DKB199, GNZ2004 and P30F90) and three sowing times (24/11/2006, 19/12/2006 and 13/01/2007) in an experimental area of the Agriculture Department of the Federal University of Lavras, Lavras, Minas Gerais State, Brazil. The treatments were arranged in a randomized block design with three replicates. Data were collected on flowering dates and physiological maturity, constituígrain number per square meter, grain mass and grain yields. These data were compared with the data simulated by the model using the mean square error (RSME), percentage deviation (PD) and agreement index (d). The results showed that corn sown in January, independent of the hybrid, had lower grain number per square meter, grain mass and grain yield than sowings in November and December. CERES-Maize proved very efficient to simulate the flowering and physiological maturity dates as the RSME values were below 10%, the ‘d’values were above 0.80, and the highest PD value was -11%. For grain number per square meter, grain mass and grain yield, the simulation was considered good, with RSME below 20%. These variables had the highest PD values, mainly in the last sowing date, indicating that environmental conditions unfavorable to the good performance of the crop affect efficiency of simulation. The Ceres- Maize model was proven a good tool for simulation of agronomic characteristics in corn hybrids.