| 春玉米产量动态预报技术的改进方法探索 |
| 点此下载全文 |
| 引用本文：邱美娟，刘布春，刘园，等..春玉米产量动态预报技术的改进方法探索[J].气象与环境科学,2020,43(1):1-8.Qiu Meijuan,Liu Buchun,Liu Yuan,et al..Exploration of an Improved Method for Yield Dynamic Prediction of Spring Maize[J].Meteorological and Environmental Sciences,2020,43(1):1-8. |
| 摘要点击次数: 649 |
| 全文下载次数: 400 |
|
| DOI：doi:10.16765/j.cnki.1673-7148.2019.04.001 |
| 基金项目: |
|
| 中文摘要:为及时、准确地进行玉米产量预报，为吉林省玉米产量预报业务提供参考依据，为政府调控提供科技支撑，利用吉林省19802016年春玉米产量和50个气象站逐日气温、降水量、日照时数等资料，基于相似距离和相关系数构建综合诊断指标筛选气温、降水量、日照时数等各类气象要素历史相似年，根据各类气象要素历史相似年与预报年的玉米产量丰歉气象影响指数之间的关系，建立吉林省春玉米产量动态预报模型。同时，对历史相似年的筛选方法进行改进，利用欧氏距离直接筛选综合气候历史相似年，根据气候历史相似年与预报年的玉米产量丰歉气象影响指数之间的关系，构建春玉米产量预报模型。对比改进前、后的产量预报模型的预报，结果表明：两种方法在吉林省玉米单产预报中，预报准确率均较高，普遍在85%以上。产量预报模型对20022013年的预报检验结果表明，改进方法后20022013年单产预报平均准确率提高了3.9个百分点，均方根误差降低了4个百分点，标准差降低了2。对20142016年的预报检验结果表明，改进方法的玉米产量预报结果优于传统方法的预报结果。改进方法比传统方法准确率更高，稳定性更强，应用价值更高。 |
| 中文关键词:综合诊断指标 产量预报 相似年 春玉米 产量丰歉 |
| |
| Exploration of an Improved Method for Yield Dynamic Prediction of Spring Maize |
|
|
| Abstract:In order to predict the yield of maize in time and accurately,provide reference frame for maize prediction business of Jilin province and scientific support for government regulation,this paper,based on the data of spring maize yield and daily temperature,precipitation,sunshine hours etc.from 50 meteorological stations during 19802016 in Jilin province,constructs a comprehensive diagnostic index to screen out history similar years of temperature,precipitation,sunshine hours and other meteorological factors by using the methods of similar distance and correlation coefficient.Then,the dynamic forecast model of spring maize yield for Jilin is established according to relationship of the meteorological influence index for bumper or poor harvest of maize yield between the historical similar years of each kind of meteorological factors and the forecast year.At the same time,the selection method of historical similar year is improved,and the comprehensive climate history similar years are screened directly based on Euclidean distance.The forecast model of spring maize yield is established based on the meteorological influence indexes for bumper or poor maize yield in historical similar years and forecast years.The prediction effects of the forecast model before and after the improvement are compared.The research reveals that the prediction accuracies of the two methods are both high in the maize yield prediction in Jilin,generally over 85%.However,the results of yield prediction model for 20022013 indicate that the average accuracy rate of the improved method for the 20022013 yield has increased by 3.9% on average,the root mean square error has decreased by 4%,and the standard deviation has decreased by 2.Moreover,the test results of maize yield prediction for 20142016 years are also better than by the traditional methods.Compared with traditional methods,the improved method has high accuracy,strong stability and high application value. |
| Keywords: |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|