摘要
本文以北京市8个行政区(东城区、西城区、石景山区、海淀区、朝阳区、昌平区、顺义区、怀柔区)的PM2.5指数计算各区逐月雾霞天气过程计数频数为研究对象,选择考虑包括地表温度、相对湿度、平均风速、SO_2质量浓度和NO_2质量浓度在内的5个影响因素。本文定义雾霾天气过程,构建分层贝叶斯时空模型,在一个统计模型中对诸多影响因素进行分析,并从计数分析的角度对北京市雾霾天气现象的时空分布、影响因素进行深入讨论。通过分析得出,温度、湿度、污染物浓度对于雾霾天气过程发生具有促进作用,平均风速对于雾霾天气过程发生具有抑制作用。从时空角度分析,从时间维度上看雾霾天气过程的发生具有明显的季节性特征,冬季(1月、2月)以及3月雾霾天气过程发生次数最高,春季(4月、5月)发生次数最低,秋季发生次数略高于夏季。从空间维度上来看,中心城区(东城区、西城区、石景山、海淀区、朝阳区)雾霾天气过程发生次数明显高于郊区(顺义、昌平、怀柔),以东城区、西城区和朝阳区最为严重。
Summary In recent years, the fog and haze phenomenon have rapidly increased, enhanced, and expanded sphere of influence in China. It had great effect on the quality of the environment, the health of residents, etc., and to some extent triggered panic among people. This article analyzes the process of fog and haze and did a in-depth discussion concerning the spatial and temporal distribution and influencing factors about atmospheric pollution problems from count analytic point. Because of weather data have spatial and temporal characteristics, and its influencing factors are relatively high, and there are complex interactions. This article build a Hierarchical Bayesian space-time models to unified the different influencing factors to analysis, and solving complicated model by hierarchy simplify every phases. In this paper, the data from eight administrative regions Dongcheng District, Xicheng District, Shijingshan District, Haidian District, Chaoyang District, Changping District, Shunyi District, and Huairou District in Beijing. Firstly, according to the PM2.5 index calculated how many times the fog and haze happen in those districts, and analyzed the spatial and temporal characteristics of temperature, humidity, wind speed, visibility levels and SO2, NO2, CO, 03 mass concentration. Then compare with the frequency of fog and haze. Finally, selected surface temperature, relative humidity, average wind speed, and SO2, No2 mass concentration including five factors. By analyzing the results, temperature, humidity, fog and haze for pollutant concentration process occurs with the promotion of the role of the average wind speed for the occurrence of fog and haze process inhibited. Analysis from the point of spatial and temporal, from the time dimension, fog and haze happens has obvious seasonal characteristics. In winter, January, February, and March the number of fog and haze process occurs highest. In spring, April, May the minimum number of occurrences. It happens more times in autumn than summer; from the spatial dimension, the central city of fog and haze process occurs significantly higher than the number of suburbs east of the city; Xicheng District, Chaoyang District are the most serious, and the situation is getting better from the northern part to the southern.
出处
《数理统计与管理》
CSSCI
北大核心
2017年第6期970-982,共13页
Journal of Applied Statistics and Management
基金
中国人民大学科学研究基金(中央高校基本科研业务费专项资金资助)项目成果(15XNL008)
关键词
分层贝叶斯
时空模型
雾霾
计数分析
Hierarchical Bayesian, spatiotemporal model, haze, counts analysis