多种气象要素及其变化对AOD与PM_(2.5)关联模型的影响研究  

Influence of multiple meteorological parameters and their variations on the association model of AOD and PM_(2.5)

作　　者：周云云 张德英 施润和[1,2,3,4] ZHOU Yunyun;ZHANG Deying;SHI Runhe(School of Geographic Sciences,East China Normal University,Shanghai 200241;Key Laboratory of Geographic Information Science,Ministry of Education,East China Normal University,Shanghai 200241;Joint Laboratory for Environmental Remote Sensing and Data Assimilation,East China Normal University,Shanghai 200241;Joint Institute of New Energy and the Environment,East China Normal University and Colorado State University,Shanghai 200062)

机构地区：[1]华东师范大学地理科学学院,上海200241 [2]华东师范大学地理信息科学教育部重点实验室,上海200241 [3]华东师范大学环境遥感与数据同化联合实验室,上海200241 [4]华东师范大学科罗拉多州立大学中美新能源与环境联合研究院,上海200062

出　　处：《环境科学学报》2019年第1期204-211,共8页Acta Scientiae Circumstantiae

基　　金：国家重点研发计划项目(No.2016YFC1302602);上海市卫计委重点学科建设项目(No.15GWZK0201);中央高校基本科研业务费项目

摘　　要：利用卫星遥感反演气溶胶光学厚度(AOD)已成为获取宏观、连续空气污染信息的一种有效手段.通过构建AOD-PM_(2.5)的关联模型是实现空间范围内PM_(2.5)监测的主要方法,而气象要素是该模型中的重要输入参数,直接影响到模型模拟的精度.当前诸多模型多采用地面气象要素,缺乏对于不同高度气象要素及其变化对构建AOD-PM_(2.5)关联模型的影响研究.本文以淮河流域五省为例,在实测地面气象资料的基础上,利用再分析气象资料,考虑了从地面至高空不同高度处的气象要素及其垂直变化,运用多元逐步回归方法,对比了地面与不同高度气象要素及其变化量对AOD-PM_(2.5)关联模型的贡献程度.结果表明:①AOD-PM_(2.5)关联模型在不同站点、不同季节的差异仍较为明显,不同高度及随高度变化的气象要素对提高春季AOD-PM_(2.5)关联模型的精度有较显著影响;②考虑了不同高度气象要素及垂直变化的多元逐步回归线性模型的表现优于仅考虑地面气象要素的模型,尤其是春季的改善较明显,RMSE降幅达到近43%;③基于地理加权回归方法的AOD-PM_(2.5)关联模型的估算结果略优于多元逐步回归线性模型.The use of aerosol optical depth(AOD)data derived from the satellite remote sensing retrievals has become an effective means to obtain large-scale and continuous air pollution information.By combining with particulate matters(PM2.5)observations,various AOD-PM2.5 regression models have been established and become an important method to monitor the changes of PM2.5 concentrations over a wider space.In these models,specific meteorological elements are required as important input parameters,which directly affect the model accuracy.However,most current AOD-PM2.5 models only incorporate surface meteorological elements,and the impact of meteorological parameters at other different heights and their vertical changes are generally missed.Therefore,this study uses the multivariate stepwise linear regression method to construct AOD-PM2.5 models in different sites over the Huaihe River Basin on the basis of surface observations and reanalysis data,with a focus on comparing the contributions of various meteorological parameters at different heights as well as their vertical changes to the model.The main results are as follows:①the AOD-PM2.5 regression model exhibit significant differences at different sites and seasons.In particular,the incorporation of meteorological parameters at different heights and their vertical variations can greatly improve the AOD-PM2.5 model accuracy in spring.②The AOD-PM2.5 regression model by including key meteorological parameters at various heights as well as their vertical changes shows superiority over the models only considering the surface meteorological parameters,especially in spring with the root mean square errors reduced by more than 43%.③The estimation of PM2.5 concentrations from the geographically weighted regression model is slightly better than that from multivariate stepwise linear regression model.

关 键 词：淮河流域 大气污染 PM2.5 气象要素 气溶胶光学厚度(AOD) 

分 类 号：X513[环境科学与工程—环境工程]