基于BP神经网络模型的广西月降水量降尺度预报

DOWNSCALING FORECAST FOR THE MONTHLY PRECIPITATION OVER GUANGXI BASED ON THE BP NEURAL NETWORK MODEL

利用500hPa月平均高度距平场派生出涡度变化、经向风切变、纬向风切变等变量场。从1958～2001年6月500hPa月平均高度距平场及其派生变量场中选取预报因子,并将各个场中的因子分别作EOF分解,得到浓缩了初选因子变量大部分信息的综合预报因子,用以建立同月的广西月降水量的BP神经网络预报模型。进而利用2002～2005年月动力延伸集合预报产品及其派生变量,对广西6月降水量作BP神经网络降尺度释用预报。作为对比试验,以相同的预报量,从1957～2000年5～12月及1958～2001年1～4月500hPa月平均高度距平场中选取预报因子,并作相同处理,建立前期综合因子的广西6月降水量BP神经网络预报模型。独立样本试验结果表明,利用同期综合因子建立的BP神经网络降尺度预报模型的拟合精度优于利用前期综合因子建立的预报模型,但预报效果依赖于月动力延伸集合预报产品。

Variables fields such as enstrophy, meridional-wind vary, and zonal-wind vary are derived from monthly 500 hPa geopotential height abnormal fields. In this work, we elect original predictors from monthly 500 hPa geopotential height abnormal fields and their variables in June of 1958-2001, and make comprehensive predictors by way of making empirical orthogonal function（EOF） respectively with the original predictors. A downscaling forecast model based on the back propagation（BP） neural network is built by use of the comprehensive predictors to predict the monthly precipitations in June over Guangxi with the monthly dynamic extended rang forecast products. To contrasted, we also build another BP neural network model with the same predictands by using the former comprehensive predictors elected from 500 hPa geopotential height abnormal fields in May to December of 1957-2000 and January to April of 1958-2001. The two models are tested and results show that the precision of superposition of the downscaling model is better than that of the one based on former comprehensive predictors, but the prediction accuracy of the downscaling model depends on the output of monthly dynamic extended rang forecast.