Aerosol Inversion In Urban Areas Supported By Surface Reflectance Data

Aerosols play a very important role in monitoring climate change and environmental pollution,and are of great significance to densely populated urban areas.Aerosol Optical Depth(AOD)can be obtained by satellite remote sensing inversion.In order to obtain the real aerosol optical thickness,it is necessary to estimate the Land Surface Reflectance(LSR)accurately.Dense Dark Vegetation(Dense Dark Vegetation,also known as Dark Target or Dark Pixel Method)can accurately retrieve aerosol optical thickness in dense vegetation covered areas with low surface reflectance.However,for urban areas,the surface reflectance is higher,and the accuracy of the determination of surface reflectance has higher requirements.Due to the complexity and diversity of urban surface types and the strong spatial heterogeneity,it is difficult to accurately determine the surface reflectance due to the influence of bidirectional reflectance characteristics.In view of the difficulties in aerosol inversion in urban areas,this paper presents a method for aerosol inversion based on MODIS BRDF/Albedo(Moderate-resolution imaging Spectroradiometer,MODIS;Bidirectional Reflectance Distribution Function,BRDF)products.The main work includes the following aspects:(1)Calculating the surface reflectance in blue band by using the parameters of nuclear driving model provided by MODIS BRDF/Albedo products.In order to improve the accuracy of surface reflectance determination and expand the space range of aerosol inversion,it is proposed to use the mean of MODIS BRDF/Albedo products for many years to determine the final surface reflectance,so as to effectively reduce the impact of cloud pixels and reduce the accidental error in the process of product generation.(2)Based on the atmospheric radiation transfer equation,according to the observed geometric parameters of MODIS sensor,the spectral response function of the band to be inverted,the possible range of surface reflectance distribution in urban area,the atmospheric model and aerosol types in the area to be inverted,a look-up table for aerosol inversion in urban area is constructed by using 6S radiation transfer model.It is used to determine the aerosol optical thickness of the image to be inverted.(3)Aerosol inversion experiments were conducted in three major cities,Beijing,Baltimore and Paris.MODIS BRDF/Albedo products were obtained from three urban areas for six consecutive years from 2012 to 2017 to support aerosol inversion experiments.The inversion results were verified by using AERONET(Aerosol Robotic Network)ground observation stations distributed in three urban areas.At the same time,the inversion results are compared with those of MOD04 C6 and MOD04 C6.1 aerosol products.The results show that the correlation coefficient between the inversion results of the new algorithm and AERONET ground observation data is 0.944,the root mean square error is 0.111,the average absolute error is 0.076,and the relative average deviation is 1.102.The proportion of observation points satisfying the expected error is as high as 70.08%,which is better than 54.27%,64.39% of MOD04 C6 and MOD04 C6.1 products.Moreover,the aerosol results inverted by this method have higher spatial resolution and continuity.(4)To explore the correlation between aerosol inversion results and urban air pollution in urban areas,the inversion results are compared with PM2.5 and PM10 data monitored by ground stations.The results show that the aerosol inversion results are in good spatial consistency with the spatial distribution of PM2.5 and PM10 data.It can provide a reference for the study of atmospheric particulate matter pollution in urban areas.The existence of bidirectional reflectance is the main reason for the difficulty in extracting various types of urban parameters.The application of remote sensing bidirectional reflectance products in aerosol inversion in urban areas can provide a reference for extracting various types of urban surface parameters.