SUMMARY
This study estimated crop water requirements using CROPWAT while AQUACROP and CERES models evaluated performance in simulating biomass and grain yield of upland rice’s response to different irrigation schedule. NERICA 4 was subjected to five treatment given as: full (100%ET), good (80%ET), medium (60%ET), average (40%ET), and low (20%ET) and three replicates in a randomized block design. It was planted at the farmyard of the International Institute of Tropical Agriculture (IITA) Ibadan, Nigeria for two seasons 2015 and 2016 and the results of the first season was used in model calibration while second season’s result was used in validation. Biomass and grain yields values were obtained and compared with simulated values from AQUACROP and CERES and the results were analyzed using E-VIEWS, R and Minitab 17 statistical tools. Results showed that there were significant differences among the models in simulation of grain and biomass yield with respect to irrigation schedule. CERES slightly underestimated LAI while both CERES and AQUACROP slightly overestimated grains and biomass yields although CERES showed highest precision in grain yield, AQUACROP showed highest accuracy in biomass yield and slightly underestimated canopy cover. CROPWAT performed excellently and could be used to efficiently estimate water requirements and reference evapotranspiration. The models’ showed high performance and accuracy in simulation of crop water requirements, grain and biomass yield of rice respectively.