南京禄口机场低能见度天气统计分析及判断模型

基于2014—2019年基于南京禄口机场能见度观测资料,统计分析了机场低能见度天气,发现低能见度天气大多发生在11月至次年1月04—07时(北京时)。通过分析雨强—能见度数据发现降水强度和能见度呈幂函数关系,建立了降水时段基于降水强度的能见度拟合模型。通过K-means聚类中心分析,发现无降水时能见度和风速以及温度的关系为正相关,和相对湿度以及修正海压的关系为负相关。最后使用随机森林算法建立低能见度天气等级判断模型,对比发现随机森林在准确度和计算效率方面较其他模型均有一定优势。

Analysis and judgement model of low-visibility weather in Nanjing airport

Based on the visibility observation data of Nanjing Lukou Airport from 2014 to 2019, this paper makes a statistical analysis of the low visibility weather at the airport. It was found that low visibility weather mostly occurred between 04:00 BST and 07:00 BST from November to January. The relationship between rainfall intensity and visibility was found to be a power function by analyzing the rainfall intensity-visibility data, and the visibility fitting model based on rainfall intensity is established. Through K-means cluster analysis, it was found that the relationship between visibility and wind speed, as well as visibility and temperature were positive, and the relationship between visibility and relative humidity, as well as visibility and sea level pressure were negative in the absence of precipitation. Finally, the low visibility weather level judgment model was established by using the random forest algorithm. Through comparative analysis, we also found that the model established based on random forest algorithm had certain advantages over other models with respect of accuracy and computational efficiency.