用事件概率回归方法预报咸阳机场辐射雾消散

在常规的天气分析基础上，尝试用事件概率回归方法来预报咸阳机场辐射雾的消散，以实现雾消预报的定量化。选取咸阳机场1992-2002年95个辐射雾日样本资料，进行事件概率回归分析。入型条件为：早晨06时(北京时)成阳机场有辐射雾(能见度≤700m)，考虑4小时内辐射雾消散与否的概率。因子的选取根据因子间及因子与预报量间的相关系数决定，所求回归方程经F检验是显著的。预报效果经回报与试报，准确率达到78％以上。

A Regression Estimation of Event Probabilities for ForecastingScatter of Radiation Fog at Xianyang Airport

Based on the synoptic routine analysis, a regression estimation of event probabilities to forecast scatter of radiation fog at Xianyang Airport is developed. An analysis is made by selecting 95 events of radiation fog at 6:00 a.m. during the period from 1992 to 2002 and thinking over the probabilities of scatter of the radiation fog in 4 hours. The regression equation is significant on the statistic F-test, and the accuracy of returning-forecast and testing-forecast are over 78%.