Using the CSM–CERES–Maize model to assess the gap between actual and potential yields of grain maize

Abstract

SUMMARY Maize in Canada is grown mainly in the south-eastern part of the country. No comprehensive studies on Canadian maize yield levels have been done so far to analyse the barriers of obtaining optimal yields associated with cultivar, environmental stress and agronomic management practices. The objective of the current study was to use a modelling approach to analyse the gaps between actual and potential (determined by cultivar, solar radiation and temperature without any other stresses) maize yields in Eastern Canada. The CSM–CERES–Maize model in DSSAT v4·6 was calibrated and evaluated with measured data of seven cultivars under different nitrogen (N) rates across four sites. The model was then used to simulate grain yield levels defined as: yield potential (Y P ), water-limited (Y W , rainfed), and water- and N-limited yields with N rates 80 kg/ha (Y W, N -80N) and 160 kg/ha (Y W, N -160N). The options were assessed to further increase grain yield by analysing the yield gaps related to water and N deficiencies. The CSM–CERES–Maize model simulated the grain yields in the experiments well with normalized root-mean-squared errors <0·20. The model was able to capture yield variations associated with varying N rates, cultivar, soil type and inter-annual climate variability. The seven calibrated cultivars used in the experiments were divided into three grades according to their simulated Y P : low, medium and high. The simulation results for the 30-year period from 1981 to 2010 showed that the average Y P was 15 000 kg/ha for cultivars with high yield potential. The Y P is generally about 6000 kg/ha greater than the actual yield (Y A ) at each experimental site in Eastern Canada. Two-thirds of this gap between Y P and Y A is probably associated with water stress, as a gap of approximately 4000 kg/ha between the Y W and the Y P was simulated. This gap may be reduced through crop management, such as introducing irrigation to improve the distribution of available water during the growing season. The simulated yields indicated a gap of about 3000 and 1000 kg/ha between Y W and Y W,N -80N for cultivars with high Y P and low Y P , respectively. The gap between Y W and Y W,N -160N decreased to <2000 kg/ha for high Y p cultivars with little difference for the low Y p cultivars. The different yield gaps among cultivars suggest that cultivars with high Y P require high N rates but cultivars with low Y P may need only low N rates.