Using Trial and Error Method to Estimate the Extinction Coefficient of Rice Varieties for Oryza 2000 with Contrasting Plant Architecture.

Background and Aims: The ecophysiological crop model, ORYZA2000, can be used to simulate the growth, development and water balance of lowland rice. The extinction coefficient (KDF) is one of the most important parameters in ORYZA2000, which has a direct impact on the spatial distribution of solar radiation in canopy and further affects the daily total gross CO₂ assimilation. The present study aims at calibrating KDF and testing the performance of ORYZA2000 on two rice varieties with large differences in plant architecture in Beijing, China.

Methods: Field experimental data, such as dates of key growth periods, leaf area index and dry weight of aboveground biomass, was collected under nonlimited water and nitrogen (N) fertilizer with direct-seeded for Teqing(TQ) and YIL55 between 2012 and 2014 at Shuangzhuang experimental station in Beijing. These data was further used to calibrate crop parameters of ORYZA2000 model. The accuracy of simulated and measured leaf area index (LAI) and organ biomass were evaluated by using root mean square errors (RMSE), and index of agreement (d). Using trial and error method to calibrate KDF with 0.05 step length in a given large number of iterations (it was greater than 5000), and the iterations was stoped when the total RMSE of aboveground biomass (i.e. leaf, stem, panicle) between the calculated KDF and the measured KDF were minimized. Then, the calculated KDF by ORYZA 2000 is compared with the KDF extracted with the spatiotemporal distribution of PAR at different canopy heights, which measured with AccuPAR canopy analysis system at flowering and filling stage. Variety TQ was with erect growth habit, however, YIL55 behaves prostrate growth habit during early growing period and becomes erect after flowering.

Key Results and Conclusion:The calculated results showed that the introgression line YIL55 had lower KDF (KDF=0.25) compared with TQ (KDF=0.4) during vegetative growth period. The KDF value of YIL55 was same with TQ (KDF=0.60) after flowering stage. The trial and errors for calculating KDF showed that RMSE was between 104 and 709 kg ha^-1 for aboveground biomass, and d was between 0.961 and 0.999. Value of RMSE for LAI were between 0.29 and 0.73 ha leaf ha^-1soil. Values of d ranged from 0.958 to 0.987. Goodness-of-fit statistics and the agreement between the KDF calibrated by trial and error method and by AccuPAR (the value of RMSE was 0.017, and NRMSE was 2.9%) indicated that the KDF calibrated by trial and error were reliable and could be able to reflect architecture variation of TQ and YIL55. Further more, the predictive performance of ORYZA2000 was sufficiently accurate in the simulation of plant biomass and LAI for TQ and YIL55 in Beijing.

Keywords: ORYZA2000; crop model; rice; extinction coefficient; trial and error method