宿州市秋冬连旱的灰色BP神经网络预测模型  

Gray BP Neural Network Prediction Model for the Persistent Drought in Autumn and Winter in Suzhou City

作　　者：孙学浩[1] 孙惠合 

机构地区：[1]南京信息工程大学,江苏南京210044 [2]安徽省宿州市气象局,安微宿州234000

出　　处：《安徽农业科学》2009年第24期11649-11650,共2页Journal of Anhui Agricultural Sciences

摘　　要：秋冬连旱是影响宿州冬小麦、油菜等越冬作物生长发育的重要因素。以Z指数≤-0.8为标准,确定宿州市秋冬持续重旱年份序列,建立GM(1,1)预测模型,并应用BP人工神经网络(BP-ANN)对残差进行拟合,对GM(1,1)预测模型进行修正。结果表明,拟合结果较单一的GM(1,1)模型有一定提高。预测2008年后的下一个宿州市秋冬(10月～2月)持续重旱年度发生在2017～2018年,对当地农业生产和防灾减灾有一定的参考价值。The persistent drought in autumn and winter, is an important impact factor of growth and development of winter wheat,rape and other winter crops. Z index of ≤- 0. 8 as the standard , year sequence of autumn and winter continuing severe drought in Suzhou City were determined, and GM （1,1）prediction model was established, and BP artificial neural network （BP-ANN） fit residuals was used to amend the GM（ 1,1 ） prediction model. The results showed that the combination of model fitting results had a certain improvement. The next persistent drought in autumn and winter in Suzhou will occur in 2017 -2018 （Oct -Feb） after 2008, the conclusions of the forecast had a certain reference value for agricultural production, disaster prevention and reduction.

关 键 词：宿州 秋冬连旱 GM(1 1)模型 BP人工神经网络 

分 类 号：S165.2[农业科学—农业气象学]