基于污染物和气象要素的北京市雾霾影响因素分析

Analysis of the influencing factors of haze in Beijing based on pollutants and meteorological elements

根据2014年-2017年间北京市每日空气质量和气象观测站观测的每日气象数据资料,在时间变化、气象条件、大气污染物组分等方面分析了北京市雾霾的影响因素。结果表明:2014年-2017年间北京的雾霾天数总体上呈现下降趋势,严重和重度雾霾天比例减少,空气质量优良天气比例提高,按月份和季节分析,北京市的雾霾多发生在10月-次年3月份之间,冬季、秋季和春季雾霾发生天数显著高于夏季,其主要原因与大气污染物的浓度密切相关;气象条件方面,雾霾的发生与近地面风速、最大风速的风向和大气相对湿度等气象要素有较强的相关性,近地面风速小、最大风速的风向来自西南、大气相对湿度高都是发生雾霾的有利因素;在污染物方面,雾霾天出现频次较多的首要污染物是PM 2.5和O 3,其中PM 2.5与全年不同时段发生的雾霾都有相关性,O 3则主要与夏季雾霾相关;文章最后部分通过统计学中ANN模型和二分类logistic回归模型分析筛选出雾霾的主要影响因素,并建立预测雾霾的简化数学模型,该数学模型的预测正确率达到95.4%。

According to the daily air quality and meteorological monitoring data in Beijing from 2014 to 2017,the influencing factors of haze in Beijing were analyzed in terms of time variation,meteorological conditions and atmospheric pollutant composition.The results show that the number of haze days in Beijing shows a decreasing trend in 2014 to 2017,and the ratio of severely and heavily polluted days decreases while that of good and moderate days increases.The haze in Beijing mostly occurred between October and March of the next year,and the number of haze days in winter,autumn and spring was significantly higher than that in summer with the pattern analysis of months and seasons,which is closely related to the concentration of air pollutants.The occurrence of haze is strongly correlated with meteorological factors such as near ground wind speed,wind direction of maximum wind speed and atmospheric relative humidity,small near surface wind speed,maximum wind speed from southwest,and high atmospheric relative humidity are all favorable factors for the occurrence of haze.In terms of pollutants,the primary pollutants with more frequent occurrence in haze days are PM 2.5 and O 3.PM 2.5 is correlated with haze that occurs at different times of the year,while O 3 is mainly related to haze in summer.At the end of this article,the main influencing factors of haze were screened out through the statistics analysis of ANN model and binary logistic regression model,and a simplified mathematical model for haze prediction was established and that of the prediction accuracy rate reached 95.4%.