CERES-maize模拟模型中品种参数优化方法研究

Method of optimizing maize variety parameters in the CERES-maize simulation model

本文以均匀设计方法为基础 ,建立了CERES maize模拟模型品种参数的优化方法。应用DSSAT35模拟系统提供的与模型品种参数选择相关的评价指标 ,对该方法的有效性进行了数值模拟检验。用 6次迭代后确定的品种参数与模型自身参数的模拟结果作对比 ,花期、生理成熟期、籽粒产量、谷物单粒重和收获期生物量的相对误差分别减少了 1 6 9% ,5 31% ,6 15 % ,3 2 6 %和 18 4 5 % ,但单位面积粒数有所提高 ,为 0 8%。结果表明用本方法确定的品种参数优化效果明显 ,并且该方法可以实现品种参数的自动调整。

Based on uniform design, the method of optimizing maize variety parameters was set up in the CERES maize simulation model. The method was also validated through numerical simulation with the relevant index in DSSAT35. The results showed that the parameters determined by this method were superior to those from DSSAT'S parameters. Comparing with those from DSSAT, the relative errors of flowering date, physiological maturity, grain yield, weight per grain and biomass at havest after 6 times of iteration with this method declined by 1 69 %, 5 31%, 6 15% , 3 26 % and 18 45 % respectively.However, the relative error of grain numbeer increased by 0 8 % . The method could automatically adjust the variety parameters.