利用非线性回归技术建立洛阳机场大风预报方法

利用2005-2008年洛阳机场200个大风天气个例,从经验和物理意义出发选取地面风速预报因子,然后对初选因子进行多种组合并同预报量作相关分析,确定了东、西风型各10个因子。根据各个因子在引入的线性和5种非线性函数形式下同预报量的相关程度,确定每个因子同预报量的最优函数关系,进而分别建立了东、西风型风速预报的多元非线性回归方程,再应用双重检验的逐步回归方案对方程进行优化,并将优化后的方程进行了检验,最后分析了影响风速主要因子的作用。结果表明:利用非线性回归方程对风速进行客观定量化预报,预报效果达到了民航气象对大风预报的质量要求,能够为实际工作提供参考;预报东、西风型风速的首要因子分别是气压梯度和变压梯度。

Prediction Method Establishment of Strong Wind at the Luoyang Airport Using the Nonlinear Regression

Some predictable factors affecting surface wind speed are selected based on the experience and physical significance from the 200 strong wind days during the period of 2005-2008 at the Luoyang airport. Then these factors are analyzed in different company and make correlation analysis with the predictor, and eastern and western wind type each has ten factors according to the magnitude of coefficient. By introducing the linear and five kinds of nonlinear functions, according to the magnitude of coefficient between the predictor and each factor in the different functional form, the most major functional relation of each factor is determined. With multi-variable and nonlinear regression, two equations including eastern and western wind type are constructed by using the factors; then the equations are optimized using the double test successive regression and tested. Finally, the functions of main factors affecting surface wind speed are analyzed. The results indicated that the wind speed prediction using the nonlinear regression is objective and quantitative, the capability of the equations has met the requirements of the commercial aviation meteorology and can provide reference for the practical work. The most important factor for eastern wind type is barometric gradient, but that for western wind type is allobaric gradient.