This study evaluated CERES-Rice, AquaCrop, and ORYZA2000 models performance in simulation of biological and grain yield of rice in response to different irrigation intervals and nitrogen levels. These models were calibrated and validated by using three years (2005 to 2007) field experiments.Three levels of irrigation interval included pond treatment, five days interval, and eight days interval, and consisted of four levels of nitrogen. The study results showed that there were significant differences among study crop models in simulation of grain and biological yield in response to different irrigation intervals. As results showed, study models performed more accurate in estimation of rice yield under irrigation intervals than nitrogen levels. All models illustrated high performance in estimation of rice yield under different irrigation intervals. CERES-Rice and AquaCrop models showed highest accuracy in simulation of grain and biological yield of rice under different levels of nitrogen, respectively. In addition, CERES-Rice model indicated highest performance in simulation of grain yield (rRMSE = 16). However, AquaCrop model estimated biological yield more accurate compared to other models (rRMSE = 15). ORYZA2000 showed less accurate in simulating grain (rRMSE = 23) and biological (rRMSE = 21) yield of rice in comparison with other models.

Keywords