基于ARIMA模型的我国油菜单产预测分析

CHINESE RAPESEED YIELD PROJECTED ON ARIMA MODEL BASIS

关键词:  ARIMA模型 AEZ模型 油菜 单产 潜力

Key words:  ARIMA model  AEZ model  rapeseed  yield  yield potential

摘要:

［目的］探索我国油菜单产潜力极限、2020年前单产潜力、提高单产潜力的空间及耕地利用布局。［方法］在我国油菜单产光合潜力和“农业生态区划”（AEZ）潜力基础上，基于1961~2010年我国油菜单产统计值，分析“时间序列”数据平稳性，逐步建立ARMA（1, 2）模型、ARMA（1, 1）模型、AR（1）模型、MA（2）模型和MA（1）模型，根据赤池信息准则（AIC）选择ARMA（1, 1）模型建立ARIMA（自回归单整移动平均）模型，拟合2011~2015年我国油菜单产并将其与统计值比较，判断拟合效果并（调整）预测2016~2020年我国油菜单产，将预测值与单产潜力极限对比，得出提高空间。［结果］我国油菜单产最大潜力为3700kg/hm2，而2017年、2018年、2019年和2020年单产将分别为2291、2353、2471和2483kg/hm2，分别是最大潜力的61.92%、63.59%、66.78%和67.11%。［结论］文章的结果意味着提高我国未来油菜单产潜力尚有较大空间，应改良中低产耕地与保持高产耕地生产力并重。研究结果旨在为我国油菜生产提供决策参考信息。

Abstract:

The objective of this research was to explore maximum potential of Chinese rapeseed yield in long term, Chinese rapeseed yield potential by 2020, how far present Chinese rapeseed yield is away from its maximum potential, and how China should make layout of differently productive arable lands with regard to increasing rapeseed yield in the future. Compared to Chinese rapeseed yield potentials of photosynthesis and maximum potential estimated by AEZ (Agricultural Ecological Zoning) model, the stationary of time series data of statistical Chinese yield from 1961 to 2010 was examined, or historic data of Chinese rapeseed yield during this period was not stationary while its lnyield is stationary, based on which ARMA(1,2) model, ARMA(1,1) model, AR(1) model, MA(2) model and MA(1) model were established step by step, and then ARIMA (Auto-regression Integrated Moving Average) model was established on ARMA(1,1) model according to its AIC (Akaike info criterion) value. Chinese rapeseed yield from 2011 to 2015 was fitted by ARIMA model and compared with statistical ones to assess the fitness, namely fitted yields is higher than statistical ones by 6.32% to 9.59% and 7.66% on the average, and eventually Chinese rapeseed yield from 2016 to 2020 were predicted by ARIMA model. The results showed that maximum potential of Chinese rapeseed yield in the future estimated by AEZ model was 3 700kg/hm2, while its yields in 2017, 2018, 2019 and 2020, projected by ARIMA model basis, were 2 291 kg/hm2, 2 353 kg/hm2, 2 417 kg/hm2 and 2 483 kg/hm2, or 61.92%, 63.59%, 66.78% and 67.11% of its maximum potential, respectively. There is still a considerable space for China to increase rapeseed yield in the future, and higher improvement opportunities for Chinese rapeseed production from ameliorating middle & low yield fields and maintaining high yield one.