基于机器学习算法的冬小麦始花期预报方法  

Method for forecasting winter wheat first flowering stage based on machine learning algorithm

作　　者：徐敏 赵艳霞[2] 张顾 高苹 杨荣明 Xu Min;Zhao Yanxia;Zhang Gu;Gao Ping;Yang Rongming(Climate Center of Jiangsu Province,Nanjing 210008,China;Chinese Academy of Meteorological Sciences,Beijing 100081,China;Meteorological Service Center of Jiangsu Province,Nanjing 210008,China;Plant Protection and Plant Quarantine Station of Jiangsu Province,Nanjing 210036,China)

机构地区：[1]江苏省气候中心,南京210008 [2]中国气象科学研究院,北京100081 [3]江苏省气象服务中心,南京210008 [4]江苏省植物保护植物检疫站,南京210036

出　　处：《农业工程学报》2021年第11期162-171,共10页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家重点研发计划项目(2019YFD1002201);国家自然科学基金项目(42075118)。

摘　　要：该研究采用机器学习算法,明确冬小麦始花期的主要气象影响因子,并建立始花期预报模型。基于1980-2019年江苏省10个观测点冬小麦生育期观测资料和逐日气象数据,应用随机森林(Random Forest,RF)、反向神经网络(Back Propagation,BP)、多元线性回归(Multiple Linear Regression,MLR)3种算法分别建立始花期预报模型,以决定系数、均方根误差、预报准确率为评判指标,对模型模拟精度进行比较分析。结果表明,温度类因子对始花期影响的重要性明显大于降水类和日照类。基于筛选出的重要特征变量,3种算法建立的始花期预报模型均可在4月初对始花期进行预报,最迟可提前5 d预报,最早可提前32 d预报;RF算法模拟精度最高,BP算法次之,MLR算法相对低一些;RF算法能准确模拟出始花期波动趋势,大部分站点的始花日期预报准确率都在85.0%以上,表明RF算法在始花期预报中有较高的可靠性和业务应用潜力。The first flowering stage of winter wheat depends strongly on meteorological factors,particularly on climate-sensitive elements.Therefore,it is remarkably significant to develop prediction models using machine learning for precise control of wheat scab.In this study,Random Forest(RF),Back Propagation neural network(BP),and Multiple Linear Regression(MLR)were integrated to establish precise prediction models for the first flowering period.The winter wheat phenology and daily meteorology were collected in 10 observation points of Jiangsu Province,China,during the period from 1980 to 2019.Five kinds of parameters were set in the RF and BP.The optimal model of each parameter was achieved after hundreds of times of automatic learning.The determination coefficient,the root mean square error,and the prediction accuracy were applied as the evaluation indicators to evaluate the forecast ability of the models.The results showed that there were obviously interannual fluctuations in the first flowering period of winter wheat,where most regions tended to be ahead of time.There were also great differences in the first flowering period among different regions.The specific difference was more than 21 d between the latest and the earliest days for the first flowering stage of winter wheat.The influence of temperature factors on the first flowering stage was more important than that of precipitation and sunshine factors.The five most important factors were ranked as follows,the active accumulated temperature of daily average temperature≥0℃from December of last year to March of that year,the average temperature from December of last year to March of that year,the accumulated days of daily minimum temperature≤0℃from December of last year to March of that year,the average temperature in March of that year,and the active accumulated temperature of daily average temperature≥0℃from December of last year to February of that year.Furthermore,the important characteristic variables were selected to predict the first flowering st

关 键 词：作物 气象 冬小麦 始花期 随机森林算法 神经网络算法 

分 类 号：P49[天文地球—大气科学及气象学]