| Atmospheric aerosols,including those of natural and anthropogenic sources,play an important role in the climate system and the ambient atmospheric system.With the rapid development of economy and society,the sources of aerosol particles become complicated.The complexity of the sources,as well as the high spatiotemporal variability of aerosol physicochemical properties,leads to constant changes in the composition of aerosol models.Aerosol model is one of the key issues of aerosol inversion in satellite remote sensing.In the inversion process,improper selection of the aerosol model will leads to large errors in the inversion results.Therefore,typical aerosol modeling is helpful to select an appropriate aerosol model,thereby reducing the inversion error caused by the unsuitability of aerosol model in the aerosol inversion process.In order to meet application requirements of aerosol remote sensing for satellite single-angle multi-band polarization observation,we evaluated aerosol retrieval capacity quantitatively,based on the theory of optimal estimation and information analysis.The proposed aerosol optical thickness(AOD)optimization inversion algorithm(Optimal Estimate Single-angel Intensity and Polarization Joint Algorithm,OE-SIPJ)is verified by satellite measured data.The main research work are as follows:1.Modeling of typical aerosol in China.Eight aerosol models were obtained by cluster analysis based on remote sensing data of ground-based observation network with long time series,and the robustness of the clustering algorithm was verified.According to the physical and chemical properties of the aerosol and the spatiotemporal distribution characteristics of the clustering results,eight aerosol models were named as:Carbonaceous-black carbon(CBC),Carbonaceous-brown carbon(CBrC),Dust-desert(DD),Dust-city(DC),High-humidity and high-pollution(HHHP),High-humidity and low-pollution(HH-LP),low-humidity and high-pollution(LH-HP),low-humidity and low-pollution(LH-LP).Based on the established aerosol models,the seasonal distribution of typical aerosols in China was analyzed,and the seasonal variation of aerosol types were further analyzed by taking Beijing,Taihu Lake,Taiwan and Lanzhou as examples,and the main aerosol models of each station were given.2.Forward radiation simulation of aerosol properties.The vector radiation transfer software was used to simulate the forward radiation of the established aerosol models.First,the scattering phase function and polarization phase function of different aerosol models were analyzed,and then the sensitivity of reflectance and polarization reflectance of different bands at the top of the atmosphere to the geometry,aerosol optical depth(AOD),and surface reflectance was analyzed.1)The scattering phase function and polarization phase function of different aerosol models show that the scattering phase function of the fine mode and the scattering phase function of the coarse mode in the aerosol model have strong forward scattering characteristics,and the difference between the bands is not significant;However,the polarization phase function distributions of the fine mode and the coarse mode are quite different.2)The sensitivity analysis of the observation geometry shows that the apparent reflectance first decreases and then increases with the increase of the scattering angle.Except for DC and DD aerosols,under the same conditions,the longer the wavelength,the smaller the reflectance.When the scattering angle ranges from 80° to 140°,the apparent polarized reflectance decreases with the increase of the scattering angle,and the longer the wavelength,the smaller the polarized reflectance.3)The sensitivity analysis of AOD shows that the apparent reflectance increases with the increase of AOD.Generally,the longer the wavelength,the smaller the reflectance.However,in coarse-dominated aerosol,the reflectance of the short-wave infrared band is larger than that of the visible nearinfrared band under high AOD;the variation trend of the apparent polarized reflectance with AOD is largely related to the wavelength,and shorter wavelengths are easy to depolarization.4)Analyzing the sensitivity of the surface reflectance under different atmospheric conditions,the results show that the surface-atmosphere coupling of the apparent polarization reflectance is mainly related to the aerosol type and AOD,and has little to do with the surface reflectance;The earth-atmosphere coupling effect of the apparent reflectance is related to the surface reflectance,aerosol type,AOD and wavelength.3.Quantitative evaluation of satellite payload retrieval capability.Based on the optimal information theory,the methods of information analysis and posterior error analysis are introduced to quantitatively evaluate the aerosol retrieval capacity at single-angle measurement.The results show that:1)under different observation geometries,the total degree of freedom for signal(DFS)of aerosol parameters is very different,more aerosol parameters(column concentration,refractive index,etc.)can be retrieval in the range of scattering angle 140°～180°;2)In fine-dominated aerosol at arbitrary observation angle,the column concentration of fine mode can be retrieved,and the inversion error of column concentration of coarse mode is larger;In coarse-dominated aerosol,the column concentration of both coarse and fine modes can be retrieved;Some parameters of aerosol spectral distribution(effective radius)and refractive index parameters can be obtained by inversion at large scattering angles;3)For the DFS analysis of parameters under different AOD and surface reflectance,the increase of polarization information is helpful to the inversion of aerosol parameters under the bright surface,and the increase of the short-wave infrared band can improve the ability to obtain surface parameters under the condition of high AOD.4.Aerosol single-angle inversion algorithm based on optimization theory and verification of inversion results.Based on the optimal inversion theory,OE-SIPJ algorithm that makes full use of multi-band polarization scalar information is proposed.The OE-SIPJ algorithm uses visible to short wave infrared multi-band scalar and polarization information to improve the information content of instrument measurement data.In addition,the multi-band information set by the instrument itself is used for surface-atmosphere decoupling,which avoids the inversion error caused by the untimely updating of the surface reflectance database and the mismatch between time and space.The satellite measured data are used to verify the validity of the algorithm.Firstly,four regions(Beijing,Hefei,Baotou and Taiwan)with large geographical differences were selected for AOD inversion,and compared with the results of the AERONET(Aerosol Robotic Network)site,it was found that the correlation coefficients between them were 0.982,0.986,0.718 and 0.989 respectively,verifying the effectiveness of the algorithm in different areas.Then,the effectiveness of the inversion algorithm under different pollution conditions was analyzed,and the AOD product of MODIS(Moderate-resolution Imaging Spectroradiometer)was used to screen out the area of AOD>0.7 and AOD<0.7,and using satellite measured data to retrieve AOD of the corresponding area based on OE-SIPJ algorithm.It is found that it is consistent with the MODIS product in space as a whole.The correlation coefficient and root mean square error(RMSE)are 0.802 and 0.217 in the AOD high value region,and the correlation coefficient and RMSE in the AOD low value region are 0.944 and 0.022.respectively,which verifies the effectiveness of the algorithm under different pollution conditions. |
PDF Full Text Request