【摘要】 利用 T1 0 6格点场资料 ,构造能反映本地天气动力学特征的组合预报因子 ,对预报时效场进行叠加处理 ,对预报关键区格点资料取其均值或极值。预报因子和预报量均采用非线性 0、1化处理。用最优子集回归建立降水预报方程 ,用多因子概率权重回归预测其降水概率。系统从资料采集到输出概率结果在 5 86以上微机上实现全自动化 ,预报结果客观、定量。更多还原

【Abstract】 The nonlinear predictors which reflect the features of local weather dynamics could be developed by the system of T106 products,and it can manipulate time internal with superposing method.The correlation could also be built with the mean values and extremes of grid date within the forecast area togather with the amount of precipitation. The predictors and predicatands were all manipnlated with 0,1 nonlinearization.The prognostic equation of precipitation with optimal subset regression was built,and the precipitation probability with multipal factorial probability weighing regression was forecasted. The system was all automated in 586 computer from data collection to output of probobility results.Its forecast conclusion was objective and quantitative,and it became a major tool of objective precipitation probability forecast.更多还原