Calibration and Validation of the CERES-Wheat Crop Simulation Model for Phenology, Leaf Area Dynamics and Yield of Wheat (Triticum aestivum L.) under Varying Sowing Dates, Irrigation Schedules and Nitrogen Levels in a Subtropical Agro-Ecosystem of the Vin
Wasim Khan *
Department of Agronomy, Naini Agricultural Institute, Sam Higginbottom University of Agriculture, Technology and Sciences, Prayagraj–211007, Uttar Pradesh, India.
Rajesh Singh
Department of Agronomy, Naini Agricultural Institute, Sam Higginbottom University of Agriculture, Technology and Sciences, Prayagraj–211007, Uttar Pradesh, India.
Shraddha Rawat
Department of Agronomy, Naini Agricultural Institute, Sam Higginbottom University of Agriculture, Technology and Sciences, Prayagraj–211007, Uttar Pradesh, India.
*Author to whom correspondence should be addressed.
Abstract
Aims: To calibrate and validate the CERES-Wheat model of the Decision Support System for Agrotechnology Transfer (DSSAT v4.7) to simulate phenology, leaf area index (LAI), grain yield and biological yield of wheat (Triticum aestivum L.) cultivar HD 2967 under varying sowing dates, irrigation schedules and nitrogen levels.
Study Design: Split-plot field experiment with three replications and 18 treatment combinations of sowing date (main plot), irrigation and nitrogen.
Place and Duration of Study: Crop Research Farm, Sam Higginbottom University of Agriculture, Technology and Sciences, Prayagraj, Uttar Pradesh, India, during the rabi (winter) seasons of 2019–20 and 2020–21.
Methodology: Wheat was sown on 20 November and 20 December, with two, three or four irrigations and nitrogen rates of 80, 120 or 160 kg N ha⁻¹. Genetic coefficients were calibrated using 2019–20 data and validated against data from an independent season (2020–21) using mean bias error (MBE), root mean square error (RMSE), normalised RMSE (nRMSE), and Willmott’s index of agreement (d-stat).
Results: Anthesis and maturity were simulated with nRMSE below 6% in both seasons, although maturity was predicted 3.5–7.5 days early in the warmer 2020–21 season. Grain yield was simulated with acceptable accuracy in calibration (nRMSE 1.9–20.4%) but less accurately in validation (6.7–27.6%), when it was overpredicted under timely sowing. The model underestimated biological yield in both seasons (nRMSE 6.7–31.0% and 14.2–28.6% in calibration and validation, respectively), and harvest index was overestimated. Maximum LAI was underestimated in calibration (nRMSE 21.2–40.8%) but was simulated accurately in validation (11.1–14.0%).
Conclusion: The calibrated genetic coefficients captured the phenology and grain-yield levels of HD 2967 reasonably well, but the underestimation of biomass and canopy growth and the early maturity in the warmer season show that they need refinement with multi-season data before the model is used for management optimisation or yield forecasting in the Vindhyan zone.
Keywords: Crop simulation model, CERES-Wheat, calibration, validation, phenology, leaf area index, grain yield, Vindhyan zone