基于最优化相关分析的油菜开花期预报模型研究

Establishment of predicting models for oilseed rape florescence using an optimal correlation analysis method

为了最大限度地满足政府举办特色农业旅游的需求,以高淳为例,研究分析了高淳油菜农业气象观测数据,发现油菜开花期与t≥3℃的有效积温呈极显著相关关系;进而利用最优化相关分析法,筛选出影响油菜开花期的最佳建模因子,创建了开花期气象中期预测模型以及基于环流和海面温度的开花期长期预测模型,均达到α=0.001显著水平,预报时效可提前10~30 d;经检验,三种模型的历史拟合结果和三年试报效果较好;在此基础上,试用了三种模型的复相关系数作为加权系数进行加权平均,综合集成预报效果更佳。油菜开花期预报将对公众踏青赏花和政府招商引资具有现实的意义,同时预报思路可为其他作物花期预报研究提供借鉴。

To meet the government s demand of developing characteristic agritourism,the agrometeorological observation data of oilseed rape in Gaochun,Nanjing.It is found there is significant correlation between oilseed rape florescence and effective accumulated temperature of t equal to or higher than 3℃.By using an optimal correlation analysis method,the optimal factors that affect the model performance are then extracted to establish a mid-term meteorological predicting model for florescence and two long-term predicting models for florescence based on atmospheric circulation and SST,the confidence levelαof which both reaches 0.001,and the oilseed rape florescence can be predicted 10 to 30 d in advance.The inspection shows the three models perform well in fitting results and experimental prediction in the past three years.The multiple correlation coefficients of the three models are used as weighting coefficient to calculate weighted mean and the ensemble prediction performs better.The prediction of oilseed rape florescence has practical meanings for people s spring outing and government s investment attraction,and can provide reference for predicting the florescence of other crops.