Assessing the options to improve regional wheat yield in eastern canada using the csm–ceres–wheat model
Citation
Jing, Q., Qian, B., Shang, J., Huffman, T., Liu, J., Pattey, E., Dong, T., Tremblay, N., Drury, C.F., Ma, B.L., Jégo, G., Jiao, X., Kovacs, J., Walters, D., Wang, J. (2017). Assessing the options to improve regional wheat yield in eastern canada using the csm–ceres–wheat model. Agronomy Journal, [online] 109(2), 510-523. http://dx.doi.org/10.2134/agronj2016.06.0364
Plain language summary
Wheat yield is relatively low in eastern Canada. This study aimed to assess fertilizer N management options to improve the regional yield of wheat using a modelling approach. A crop model was adapted to simulate winter wheat by replacing air temperatures with estimated temperatures under snow cover, and then the model was evaluated for simulating winter wheat using experimental data collected at two sites and spring wheat at three sites in eastern Canada. Across all the experimental years and sites, the model simulated the measured yields with a deviation of 14%. Regional yield under rainfed conditions in the Eastern Ontario Region was simulated with 0, 1, 1.5, and 2 times the recommended N rate (around 50 kg N ha–1) and unlimited N, i.e., no N stress all the time during the growing period. The simulated average regional yield (in dry matter) with the recommended N rate ranged from 2180 kg ha–1 to 2502 kg ha–1 depending on the cultivars. They were close to the reported yield of 2440 kg ha–1. The simulated regional yields with unlimited N were two times that with the recommended N rate, showing a considerable yield gap. Our simulations indicate that regional yield could increase to 3600 kg ha–1 in the Eastern Ontario Region if the N rate was increased to around 100 kg N ha–1, although other abiotic factors such as lodging and environmental and economic implications should be evaluated.
Abstract
Wheat (Triticum aestivum L.) yield is relatively low in eastern Canada. This study aimed to assess fertilizer N management options to improve the regional yield of wheat using the CSM– CERES–Wheat model. The model was adapted to simulate winter wheat by replacing air temperatures with estimated temperatures under snow cover, and then the model was evaluated for simulating winter wheat using experimental data collected at two sites and spring wheat at three sites in eastern Canada. Across all the experimental years and sites, the normalized root mean squared error (nRMSE) between simulated and measured yields was 14%. Regional yield under rainfed conditions in the Eastern Ontario Region (a Census of Agriculture unit as a case study) was simulated with 0, 1, 1.5, and 2 times the recommended N rate (around 50 kg N ha–1) and unlimited N for the calibrated cultivars of spring wheat from 1981 to 1999. The simulated average regional yield (in dry matter) with the recommended N rate ranged from 2180 kg ha–1 for cultivar Hoffman to 2502 kg ha–1 for AC Brio. Both were close to the reported yield of 2440 kg ha–1, with nRMSE values ranging between 20.3 and 16.6%. The simulated regional yields with unlimited N were two times that with the recommended N rate, showing a considerable yield gap. Our simulations indicate that regional yield could increase to 3600 kg ha–1 in the Eastern Ontario Region if the N rate was increased to around 100 kg N ha–1, although a slight decrease in N use efficiency would occur. In addition, with such increases in the N fertilization rate, other abiotic factors such as lodging should be evaluated.