Calibration of CERES3 (Maize) to improve silking date prediction values for South Africa

Abstract

Silking which coincides with the mid-summer drought in South Africa, could have a significant negative effect on maize yield. Errors in the prediction of silking date could contribute to errors in kernel number simulation, and consequently to errors in yield simulation. A field trial, comprising 21 planting dates and three cultivars representing short, medium and long growing season requirements was used to modify the crop growth model CERES3 (Maize) to improve its accuracy in the simulation of silking date by four days. The results were verified against a historical data set and showed an improvement in the simulation of silking date by two days.