基于ARIMA模型的中国水稻单产预测分析

Chinese Rice Yield Predictive Parsing Based on ARIMA Model

在中国水稻单产的光合潜力和"农业生态区划"(AEZ)模型计算的最大潜力的基础上,运用ARIMA(自回归单整移动平均)模型预测了2016—2020年中国水稻单产水平。结果表明,2016、2017、2018、2019和2020年中国水稻单产分别为7 070、7 189、7 311、7 435和7 561 kg/hm^2,分别是其最大潜力7 800 kg/hm^2的90.64%、92.17%、93.73%、95.32%和96.94%,说明提高中国水稻单产将变得越来越困难,未来总产的提高应主要依靠改良中低产稻田。

Based on the estimation by light use efficiency（LUE） and the Agricultural Ecological Zoning（AEZ） model,the potentials of Chinese rice yield for 2016—2020 were projected by the ARIMA（Auto-regression Integrated Moving Average） model.The results showed that maximum potential of Chinese rice yield in the future estimated by AEZ model was 7 800 kg/hm-2,while its yields in 2016,2017,2018,2019 and 2020,projected on ARIMA model basis,would be 7 070,7 189,7 311,7 435 and 7 561 kg/hm-2,respectively,or 90.64%,92.17%,93.73%,95.32% and 96.94% of the maximum of 7 800 kg/hm-2.It is indicated that higher improvement opportunities for Chinese rice production would mainly come from raising the potential of middle and low yield fields as it is increasingly difficult to raise rice yield in the future.