| In recent years,with the occurrence of air pollution problems caused by urban development,such as smog,atmospheric environment has attracted increasingly attention.Atmospheric aerosol is one of the main factors causing air pollution.The aerosol optical depth(AOD)is a physical parameter which is used to study aerosol properties.At present,using satellite remote sensing technology to study the characteristics of atmospheric aerosols is the main means to monitor and manage the atmospheric environment problems.Gaofen-4 satellite(GF-4)is a new generation of geostationary satellite launched by China.It has visible light-multiple bands in the near-infrared band and has high time and high spatial resolution.It can continuously scan an area at a fixed angle,which can provide reliable data for continuous inversion of AOD.GF-4 satellite has advantages of wide monitoring range,short period,fast response speed,etc.,and it has a strong application value in the inversion of aerosol optical depth.However,GF4-PMS lacks the short-wave infrared band of 2.12μm,so the traditional dark target algorithm of MODIS cannot be used to calculate the surface reflectivity directly.Gaofen-4 satellite can get more than one image from different times in one day.The change in surface reflectance between images is affected by the geometric position of the sun,target,and sensor.Therefore,this paper took the concept of bidirectional reflectance of consideration.The basic thoughts is using the MODIS BRDF model and MODIS V5.2 algorithm to take the angle conversion between MODIS and GF-4 images at different times,and then the surface reflectance of GF-4 can be determined.Finally,we used the 6S model to simulate atmospheric parameters and built a lookup table to complete the aerosol inversion of GF-4.This paper selected part of the Yangtze River Delta as the research area for studying the GF-4 aerosol inversion algorithm.The accuracy of its spatial distribution is verified by the MODIS 04 aerosol product,and the accuracy of the results is verified by the observed data from the AERNOT.Results are showed as follow:(1)In this paper,the MODIS surface reflectance is calculated by the MODIS V5.2algorithm,and then the MODIS BRDF product MCD43A1 is used to convert the MODIS surface reflectance to the GF-4 observation geometric angle.Then by establishing the relationship between the spectral response functions of the two sensors,the MODIS surface reflectance is converted to the GF-44 surface reflectance.(2)In this paper,the 6S model is used to establish a look-up table.Combined with the determined GF-4 surface reflectance,the entire AOD inversion process is programmed by using IDL to realize the GF-4 aerosol optical depth inversion.(3)It is showed that the spatial distribution characteristics of the aerosol optical depth retrieved by the algorithm are generally consistent with the MODIS aerosol product in the study area.It shows the characteristics of low values in vegetation areas and high values in urban areas.(4)Additionally,the results value also had a significant correlation with the observation data.R~2of regression analysis is higher than 0.90.Overall,the value of inversion results of GF-4 is bigger than the observed value.In summary,the inversion results in this paper are credible. |
PDF Full Text Request