Abstract: A direct-seeding experiment of early rice with 8 sowing dates was carried out in Nanchang county in 2017 and two main conventional early Indica rice varieties in Jiangxi province were was selected in this study. The 8 sowing dates were defined as early sowing (a: March 11, March 16, March 21, March 26) and normal sowing (b: March 31, April 5, April 10, April 15). Growth period, leaf area index, biomass and meteorological data observed in field experiments were applied to the ORYZA_V3 rice model to analyze the suitability of the ORYZA_V3 rice model for direct seeding of early rice in Jiangxi. The results showed that the ORYZA_V3 rice model simulates the growth period of direct-seeded early rice well with (determination coefficient)R2 >0.99, and the fluctuation range of NRMSE（normalized root mean square error） was between 0.61% and 3.12%. The simulation verification results of leaf area index, total biomass, and panicle biomass at the early sowing seeding period showed that the earlier the sowing date, the worse the simulation effected，and the closer to the normal sowing date, the better the simulation effected. The suitability of different direct-seeding early rice varieties of ORYZA_V3 rice model was different. Among them, the simulation value of the growth period of the early sowing and the measured value of Zhongzao 35 was differed from -1 to 1 days, and differed from -1 to 4 days in Zhongzao 39. Studies had shown that when the ORYZA_V3 rice model was used to simulate the double-cropping early direct-seeding rice, it was necessary to calibrate the model parameters of the corresponding rice sowing date and variety based on multi-year observation data to improve the local simulation effect of the ORYZA_V3 rice model.

Key words: Oryza_v3 Rice Model, Direct-seeding early rice, Simulation, Suitability, Jiangxi province