Performance Assessment of CERES-Rice Model for Predicting Nitrogen Balance Components During Rice Growing Season in Subsurface- Drained Paddy Fields

Crop simulation models are suitable tools for quick- low cost prediction of the effects of structural and managerial changes on agricultural systems. In this research, the capability of CERES- Rice model to predict nitrogen (N) losses was investigated in paddy fields under 3 subsurface drainage systems with 0.9 m drain depth and 30 m drain spacing (D0.9L30), 0.65 m drain depth and 30 m drain spacing (D0.65 30) and 0.65 drain depth and 15 m drain spacing (D0.65L15). During drainage periods of two successive growing seasons of rice (2011- 2012), the losses of N from different drainage systems were measured. The time of phenological changes and rice yield were also determined. The model performance during calibration (2011) and validation 2012) was determined using the model efficiency, index of agreement and normalized root mean square error. The indices were in the acceptable range in both the calibration and validation stages indicating the effectiveness of the model to simulate growth and yield of rice in the studied treatments. During the two growing seasons, the total measured losses of N in the D0.9L30, D0.65L30 and D0.65L15 systems were, respectively, 15.4, 24 and 20.8 kg ha-1 and the corresponding simulated losses were 5.7, 6 and 4.3 kg ha-1, respectively. Measured and simulated N uptakes by rice were in the range of 115.2- 181.4 kg ha-1 and 71.7- 99 kg ha-1, respectively, in the different treatments. Based on the results, accurate prediction of N balance components by CERES- Rice requires basic factors affecting these components in the study condition be considered in the model.