摘要
通过分析衡阳各县晚稻产量与综合气象干旱指数CI值的关系,利用SPSS软件剔除与预报量关系较弱的自变量,建立了CI值对晚稻产量预报的回归模型。结果表明,从回归模型分析来看,剔除洪涝年份后,晚稻Y和D、M的相关性得到了提高,其回归方程为Y=384.187+0.508M。模型效果检验表明,样本残差服从正态分布;残差散点的分布随机均匀,且大多落在水平直线-2和2之间;标准化残差的正态曲线的均值为0,标准差为0.989,接近标准正态曲线,回归拟合程度高。从模型检验结果来看,利用该模型对2009年各县晚稻产量进行预报,所得结果和实际相比,平均误差为0.428%。
Through analyzing relationship between late rice yield and comprehensive meteorological drought index CI in each county of Hengyang,by using SPSS software to eliminate variables which has weak relationship with predicted value,the regression model about CI value predicting late rice yield was established.Seeing from analysis of regression model,the correlation between Y and D,M increased except flood years,the regression equation is Y=384.187＋0.508M.The model test showed that,sample residual comply with normal distribution;the distribution of residual scatter is random and uniform,and mostly fall in line between-2 and 2;mean of normal curve is 0,standard deviation is 0.989,the regression fitting degree is high.In conclusion,mean deviation between result and actuality is 0.428% by using this model to predict late rice yield in each county in 2009.
出处
《安徽农业科学》
CAS
北大核心
2011年第29期18185-18187,共3页
Journal of Anhui Agricultural Sciences
关键词
晚稻
综合气象干旱指数CI
产量预报
Late rice
Comprehensive meteorological drought index CI
Yield predicting
作者简介
肖媚（1984-），女，湖南湘潭人，助理工程师，在读硕士，从事气候研究，E-mail：xiaomm_1984@126．com。
