Summer Aerosol Optical Depth Retrieval Based On Geostationary Satellites GF-4 And H8

The aerosol is an important factor that affects the environment and climate.It not only reduces the visibility of the atmosphere,leads to the decline of the environmental quality of the urban ecology,but also has a close relationship with people's health.Aerosol Optical Depth(AOD)is one of the important optical properties of atmospheric aerosols and an important parameter affecting atmospheric correction of optical remote sensing images.Therefore,it is of great significance to study aerosol optical depth.Based on previous studies,this paper develops an aerosol optical depth retrieval algorithm supported by an enhanced surface reflectivity library,using two geostationary orbit satellites(GF-4)and Himawari-8(H8).The algorithm improves the automatic threshold cloud detection,surface reflectance determination and dynamic aerosol type hypothesis.The specific contents and conclusions are as follows:(1)GF-4 data AOD retrieval.Considering the influence of imaging geometry on the accuracy of AOD inversion,the satellite angle of GF-4 pixel by pixel is calculated based on the principle of geostationary orbit satellite imaging,which reduces the angle effect error in retrieval;the improved GF-4 cloud detection scheme,the high contrast surface and The visible light reflectance threshold is combined,an automatic threshold method is adopted to avoid the subjectivity and uncertainty of artificially setting the threshold;the GF-4 surface reflectance determination scheme is improved,and the characteristics of the surface reflectance after atmospheric correction are analyzed.The percentage minimum value method is used to construct the surface reflectivity library.Based on the surface reflectivity database,the relationship between NDVI and blue band reflectivity is obtained.The high vegetation coverage area uses NDVI real-time data to determine the blue band surface reflectance.The vegetation coverage area is based on the quarterly surface reflectance pool to determine the blue band surface reflectance.(2)H8 data AOD retrieval.The H8 enhanced surface reflectivity library support method improves the surface reflectance determination scheme and the dynamic aerosol type hypothesis scheme based on the surface reflectivity pool in four different surface cover types(Dark surface vegetation area,dark-bright surface mixed arable land area,bright surface towns,bare soil area).The towns and bare soil areas of the bright surface are respectively fitted with the reflectance relationship between the short-wave infrared and the visible light,so as to realize the real-time determination of the surface reflectance of the visible light band.Two aerosol types were used to represent different absorption and scattering characteristics,and a dynamic aerosol type was constructed by coupling the historical s'ingle-scattering albedo data of the Beijing-Tianjin-Hebei region.(3)Using the Aeronet product data and multi-source satellite aerosol products,the GF-4 and H8 AOD inversion results were cross-validated and compared,and the correlation coefficients R,root mean square error RMSE and other verification indicators were used for evaluation.The verification results show that the correlation coefficient between GF-4 AOD inversion and Aeronet site is 0.962,the correlation coefficient with MODIS AOD product is greater than 0.8,the correlation coefficient between H8 AOD inversion result and Aeronet site is 0.92,and the correlation coefficient with MODIS AOD product is greater than 0.87.The H8 AOD inversion results are compared with the different time series data of Aeronet product data,and the average relative error is less than 30%and the accuracy is stable.