北京市气溶胶光学厚度影响因子分析

Analysis of Influencing Factors of Aerosol Optical Depth in Beijing

大气污染成分、道路分布情况、地表覆盖类型等是影响气溶胶光学厚度(aerosol optical depth,AOD)的主要因素,针对区域性的顾及多种影响因素的AOD研究意义重大.研究采用北京市2020年2月至2021年2月的高分辨率气溶胶光学厚度产品数据、大气污染数据、道路线数据、地表分类数据、NPP夜光遥感数据,提出了可拓展简化气溶胶反演算法(extensible simplified aerosol retrieval algorithm,XSARA),利用Pearson指数、主成分分析、多元回归、地理加权回归模型等方法,系统分析了北京市2020—2021年气溶胶光学厚度及其驱动因素.结果发现:①湿度与AOD呈正相关,而风力风向与AOD呈负相关,通过进行多元线性回归获得AOD和各组合因子的关系方程,其中R2为0.97068,残差平方和为0.0291,表明回归效果较好;②道路沿线范围的AOD均高于周边区域,并且随着道路愈加密集,AOD也随之升高;③以人造地表为代表的地表覆盖能够增强AOD浓度,而以植被为代表的地表覆盖则减弱AOD浓度;④通过采用夜间灯光数据,发现该数据与AOD的空间分布模式相关性很高,并且采用地理加权回归模型得到的R2大于0.9,进一步表明两者相关程度较高.

Aerosol optical depth(AOD)is influenced by atmospheric pollution composition,road distribution,and surface cover type,etc.It is significant to study AOD for a regional scale considering multiple influencing factors.The study uses the highresolution aerosol optical depth product data,air pollution data,road route data,surface classification data,and NPP night-light remote sensing data from February 2020 to February 2021 in Beijing,and proposes the extensible simplified aerosol retrieval algorithm(XSARA),using Pearson index,principal component analysis,multiple regression,and geographically weighted regression model to systematically analyze the aerosol optical thickness and its driving factors in Beijing from 2020 to 2021.The results show follows(1)Humidity is positively correlated with AOD,while wind direction is negatively correlated with AOD.The relationship equation between AOD and each combination factor is obtained by multiple linear regression,in which R2 is 0.97068 and the sum of squares of residuals is 0.02910,indicating that the regression effect is good.(2)The AOD along the road is higher than the surrounding areas,and the AOD increases with the increasing density of the road.(3)The surface cover represented by man-made surface can enhance the AOD concentration,while the surface cover represented by vegetation can weaken the AOD concentration.(4)By using the night light data,it is found that the data has a high correlation with the spatial distribution pattern of AOD,and the R2 obtained by geographic weighted regression model is greater than 0.9,which further indicates that the correlation between the two is high.