基于卫星遥感AOD的华北地区2003-2014年PM2.5浓度时空分布特征

THE CHARACTERISTICS OF SPATIOTEMPORAL DISTRIBUTION OF HAZE WEATHER IN NORTH CHINA DURING 2003—2014

关键词:  AOD  PM2.5污染  mixed effect model(混合效应模型)  华北地区  EOF分解

Key words:  MODIS AOD  PM2.5 pollution  mixed effect model  North China  EOF

摘要:

华北及周边地区PM2.5造成的污染，近十年来引起了社会的广泛关注，也是科学研究的重要领域。利用2003—2014年的卫星遥感MODIS AOD数据和2014—2015年的地面观测PM2.5浓度数据，采用聚类分析、混合效应模型、EOF分解等统计分析方法，反演了2003—2014年华北及周边地区PM2.5浓度，分析其时空分布特征。主要结论如下：(1) 卫星遥感MODIS AOD与地面观测PM2.5值有较高的相关系数，可利用MODIS卫星遥感AOD对地面观测的PM2.5浓度进行反演；(2) 华北地区PM2.5浓度呈现出明显的空间分布特征：太行山脉是污染强弱明确的分界线，山脉东南部的污染显著高于西部，且在地势变化的地方出现明显的突变；河北南部、河南北部和山东西北部分区域是污染最严重的地区；(3)2004年、2009年以及2013年后都是污染浓度比较低的年份。

Abstract:

In north China, haze pollution has aroused wide concern of the society in the past ten years. Based on satellite remote sensing of aerosol optical thickness (AOD) and ground-based observation of PM2.5 data, linear regression, cluster analysis, and other statistical analysis methods were used to study the spatiotemporal distribution characteristics. The results showed that: (1) The AOD had higher correlation coefficient with the ground-observed PM2.5 value, and can be used for the inversion of PM2.5 in the study area. (2) In north China and the surrounding areas, PM2.5 presented obvious spatial distribution, the Taihang Mountains was the clear dividing line of the intensity of pollution, contamination of the southeastern mountains was significantly higher than that of the northwest and there were obvious abrupt changes where the terrain varied. Small parts of the southern Hebei Province, northern Henan Province and southern and northwestern parts of Shandong Province were the most polluted areas. (3) Relatively low pollution concentrations happened in 2004, 2009 and the years after 2013.