Evaluation of the CERES-Maize Model for the smallholder farming sector of Zimbabwe: model calibration

The Ceres-Maize model was calibrated using documented historical data from Domboshawa, an agricultural training centre about 30 km north of Harare, Zimbabwe. Agronomic data from long-term tillage trials conducted during the 1990/91 to 1995/96 cropping seasons were used. Theses data were, however, limited to silking, maturity dates and grain yields of the R215 maize variety. The initial genetic coefficients for variety R215 were obtained from previous local studies. Modified coefficients were obtained using the Genotype Coefficient Calculator from the model, based on the results from Domboshawa. In cases where the procedure predicted coefficient values outside the acceptable range, these were determined manually. Results from the model showed that silking and maturity dates were accurately predicted with R2 values of 0.97 and 0.88, respectively. Slopes of lines for both traits were not significantly different from 1 at the 1 % level (p ? 0.001 and p < 0.006, respectively). A highly significant relationship between measured and simulated grain yield was also observed (p ? 0.002; R223 = 0.93), indicating that the model can be used successfully to predict maize yields under similar conditions.