Evaluation of DSSAT model (CERES rice) on rice production: A review

Rice is the staple food for over half of the world population. Crop models potentially offer a means to readily explore management options to increase yield, and to determine trade-off between yield, resource-use efficiency and environmental outcomes. This paper reviews the performance of CERES-Rice model in different regions of the world in relation to their potential application towards increasing resource use efficiency and yield of rice. In this article, the CERES-Rice model evaluation by using the simulated and observed values on crop phenology (anthesis, physiological maturity) and final grain yield mainly over Asian countries by different authors has been compiled and described. Mainly the model was evaluated based on different statistical measures such as RMSE and D-index. Several datasets for the prediction of grain yield and phenological period across different parts of Asia were examined. This particular model predicted those with high-accuracy (nRMSE1-5% for anthesis and 1-4% for physiological maturity days). For various data sets for grain yield, the nRMSE varied between 0.05–5.00 percent with error percentage of 2-5%. The model sometimes over-estimated or under-estimated the values of grain yield, especially under water stress conditions.