Calibration and Validation of the DSSAT Model with Experimental Data for Three Varieties of Wheat on Different Planting Dates

Abstract

The Crop Simulation Model (CERES-wheat module) of the Decision Support System for Agrotechnolgy Transfer (DSSAT) was used in this study to simulate the effect of planting date (D) on growth, development and yield of three varieties of durum wheat (Triticum turgidum L subsp. durum). The studied varieties were Cham1 (V1), Deir Alla6 (V2) and Haurani (V3). Data were obtained from a field experiment conducted for this purpose on the growing season 2015/2016 at Mushagger Agricultural Research Station Southwest Amman (31° 46' 24.7'' N, 35° 47' 47.3'' E, 800 m above sea level). First planting date (D1) was used for model calibration and the GenCalc software was applied to achieve acceptable genetic coefficient values. Calibration results showed an excellent estimation for days to anthesis, number of grains m-2, grain yield and days to physiological maturity with normalized root mean square error (nRMSE) ranged from 0 to 5.79%. Tops weight, harvest index, unit grain weight (except for V3) and number of tillers m-2 were poorly simulated. Validation of the CERES-Wheat model of the DSSAT using means of observed results of D2 and D3 showed excellent simulation (nRMSE < 10%) for anthesis and maturity dates for the three examined varieties. Good prediction (10 ≤ nRMSE < 20%) were attained for grain yield (for V2 and V3) and for grain nitrogen percent (except for V3 which was excellent). Fair predictions (20 ≤ nRMSE < 30%) were recorded for grain unit weight of (V1 and V2) and tops weight of V1. Conversely, grain unit weight was poorly simulated for V3. From these results, it can be concluded that the DSSAT model can be considered as an effective tool for predicting wheat growth and yield.