摘要
利用2002～2004年1～12月EC、T213天气数值产品的温度格点实时分析资料、预报资料以及贵州省黔西南州8个测站观测资料,从160个样本中筛选出相关较好的因子,利用多元回归统计模型,以及根据地理环境,天气形势分型造成的温度差值(ΔT)为逐步订正值,建立黔西南州分县温度预报方程。在2005年11月～2006年2月应用中取得较好效果,预报误差小于±3.0℃,平均误差1.3℃。
The relative preferable factor in selected from 160 samples by use of Ec、the real-time analytic temperature data of T213、the forecast data and the data of eight observation stations in Qianxi nan Prefecture of Guizhou Province from 2002 to 2004,the forecast equation of every county in Qianxi nan Prefecture are created by using of multiple regression of statistical model and the stepwise of revisory number which is affected by geographical conditions and the weather types,the equation is useful for forecast ...
出处
《云南地理环境研究》
2007年第S1期67-70,共4页
Yunnan Geographic Environment Research
关键词
数值预报产品
温度
预报方程
numerical forecast of product
temperature
forecast of equation
