Evaluating crop models to assess rice yield potential in Argentina

Due to the high concentration of the rice production in Asia, the study and comprehension of rice producing chains outside this continent are key to guarantee the global food security. The objectives of this study were to validate the SimulArroz and ORYZA v3 crop simulation models for rice production in Argentina, estimate the yield potential (Yp) for different sowing dates with both models, and estimate how much Argentina can increase rice production in the current agricultural area by reducing the yield gap. Two contrasting rice models (ORYZA v3 and SimulArroz) were used. Besides both are mechanistic models, Simularroz model requires fewer coefficients than ORYZA v3,. which may increase the systematic error. The coefficient of determination between simulated and measured values was at least 0.74 for grain yield and 0.97 for phenology. RMSEn values for phenology ranged from 5.7% to 9.4%. Both models simulated grain yield accurately at high yield levels (>13 ​Mg ​ha⁻¹) with an RMSE <0.5 ​Mg ha^-1. The average irrigated rice Yp for subtropical Argentina estimated by the SimulArroz model ranged from 3.8 to 15.6 ​Mg ​ha⁻¹, and from 7.6 to 14.3 ​Mg ​ha⁻¹ in the ORYZA v3. The differences in the Yp between the models were due to different approaches in the calculation of thermal sum and radiation use efficiency. The increase in the rice production in Argentina if the fields close the exploitable gap (80% of the Yp) would be 42% of the current total rice production or 916,091 ​Mg annually.