基于Logistics概率统计模型的机载气象雷达强风暴快速识别预警算法

Quickly identification and warning strong storm weather method by airborne weather radar based on Logistics probability statistic model

不同经纬度、季节和下垫面条件下,机载气象雷达探测强风暴的特征识别预警参数阈值是不同的,传统单一的阈值门限判定法不具有普适性。为了提高灾害性强风暴天气的识别预警时效和精度,建立一种基于Logistics多元线性对数概率统计模型。利用地基天气雷达和TRMM星载气象雷达观测的大量强风暴天气样本数据,进行Logistics非连续的多元线性回归,获取适用于我国东海和南海地区的机载气象雷达识别预警强风暴天气概率模型。与气象雷达实际观测结果进行对比实验分析表明,在东海和南海地区,该方法建立的模型在识别预警精度上较为精确,有助于飞行员快速、准确判定危险天气,保障飞行安全。

The warning parameter threshold of strong storm detected by airborne weather radar is changed according to different seasons,latitude and longitude and underlying surface.The conventional method of single threshold judgment is not universality.A multivariate linear logarithmic probability statistic model based on Logistics is proposed in this paper in order to improve the warning efficiency and accuracy of severe storm weather.A large number of strong storm weather data sample sounding by ground-based weather radar and TRMM satellite weather radar is used to complete discontinuous multiple linear regression based on Logistics model,which can obtain airborne weather radar warning strong storm weather probability model apply to the east China sea and south China sea.Compared with the actual observation results of weather radar,the simulation experiment result shows that the warning strong storm model established by this method is more accurate than the conventional method in the east China sea and the south China sea.It can help the pilot quickly determine dangerous weather and ensure the flight safety.