基于不同空间尺度的作物产量集成预报——以江苏一季稻为例

Comparative analysis of different regional scales integration yield prediction——a case study of single rice in Jiangsu

基于1981—2016年江苏省不同区域一季稻产量序列,计算站点尺度的气温、降水、日照适宜度以及综合气候适宜度,在此基础上构建基于气候适宜指数的作物产量预报模型,开展不同空间尺度的一季稻产量精细化预报。同时,以各主产地市、县一季稻种植面积百分比为权重,加权集成省级产量,开展基于不同空间尺度一季稻产量序列的大区域尺度产量预测。结果表明:(1)江苏省不同区域一季稻气象产量与不同时段气候适宜指数之间存在较高的相关性,基于气候适宜指数的预报方法适用于江苏省不同区域一季稻单产预报。(2)2012—2016年省级尺度模型预报平均准确率高于97.5%,主产地市、县模型平均预报准确率低于省级尺度预报模型,主产县预报准确率年际间波动较大,表明预报区域越小,预报的难度提升。(3)基于气候适宜指数模型的江苏省级、主产地市集成,主产县集成模型预报准确率大部在95%以上,整体上看主产县集成优于主产地市集成,主产地市集成优于省级尺度模型。由此,开展地市级和县级尺度的精细化产量预报可提升省级尺度预报准确率,同时提高县级作物产量预报能力。

Based on the yield series of single rice in different regions of Jiangsu Province from 1981 to 2016, the temperature, precipitation, sunshine hours and comprehensive climate suitability at station scale were calculated. On this basis, the crop yield prediction model based on climate suitability index was constructed to carry out the fine prediction of single rice yield at different spatial scales. At the same time, with the percentage of planting area of single rice in each major producing area city and county as the weight, the provincial yield was weighted and integrated to carry out large regional scale yield prediction based on different spatial scale single rice yield series. Results show that:(1)there is a high correlation between the meteorological yield of single rice in different regions of Jiangsu Province and the climate suitability index of different time periods. The prediction method based on the climate suitability index is suitable for the yield prediction of single rice in different regions of Jiangsu Province.(2)The average forecast accuracy of the provincial scale model from 2012 to 2016 is higher than 97.5%. The average forecast accuracy of the city scale model and the county scale model is lower than the provincial scale model. The accuracy of the county scale model has a large fluctuation in different years, which indicate that the smaller the forecast area, the more difficult the forecast accuracy.(3)Based on the climate suitability index model, the prediction accuracy integrate by province, integrate by city and integrate by county is mostly above 95%.On the whole, the integrated model of main producing area counties is better than the integrated model of main producing area cities, and the integrated model of main producing area cities is better than the provincial scale model. Therefore, to carry out the refined yield forecast at prefecture level and county level can improve the accuracy of provincial scale forecast, and improve the ability of meteorological crop yield forecast at county level.