| At present,satellite aerosol retrieval has become an important means of monitoring global aerosol distribution.The aerosol product developed based on the MODIS(moderate-resolution imaging spectroradiometer)sensor is one of the most widely used satellite aerosol products.Since both Terra and Aqua have exceeded their design lifespans,the development of new satellites for aerosol retrieval studies is of great importance.The MERSI-II(Medium-Resolution Spectral Image II)sensor carried on China Fengyun-3 D satellite(FY-3D satellite for short)and MODIS sensor have similar band settings,and have the main band for aerosol retrieval,which can be used as a useful supplement to the global aerosol monitoring satellite sensor.The author’s team introduced the idea of the Dark Target(DT)algorithm and developed the first version of the aerosol algorithm for MERSI-II,based on the verification of the aerosol retrieval accuracy of the algorithm,this paper analyzes the shortcomings and deficiencies of the algorithm.Surface type factors such as Normalized Difference Vegetation Index(NDVI)and Urban Percentage(UP)were introduced into the estimation model to further improve the aerosol retrieval algorithm of the first version of MERSI-II.The main research work and conclusions are as follows:(1)This paper firstly verifies the accuracy of the aerosol retrieval results of the MERSI-II DT algorithm version 1,and compares it with the MODIS C6.1(Collection 6.1)DT algorithm aerosol product verification results.The results show that the correlation coefficient between the aerosol retrieval results of the first version of the MERSI-II DT algorithm and the ground-based measurement results is 0.859,the root mean square error is0.162,and the proportion of sample points that fall within the expected error is 61.56%,which basically meets the accuracy requirements of similar international aerosol products.However,compared with MODIS aerosol products,there is still a certain gap.Comparing the verification results of these two products,it is found that the difference is most obvious on the urban surface.The retrieval verification results of the MODIS product on the urban surface have 17.9% more sample points falling within the EE range than MERSI-II,and the retrieval accuracy is better,mainly because the MODIS algorithm takes special consideration in urban surface estimation,which has not been done by the first version of the MERSI algorithm.(2)Based on the verification of the aerosol retrieval results of the first version of the MERSI-II DT algorithm and the comparison with MODIS,this study improves the relationship between the surface bands in urban areas,mainly by adding influencing factors such as NDVI and UP to the surface estimation model of the algorithm.A large number of statistical experiments have found that the reflectance ratio of the 0.65/2.13 band changes linearly with NDVI,but does not change significantly with UP;while the reflectance ratio of the 0.47/0.65 band not only changes quadratic curve with NDVI,but also linearly changes with UP.To this end,this paper uses the above band relationship to reconstruct the quadratic nonlinear surface estimation model of the 0.47μm,0.65μm and 2.13μm band ratios for UP and NDVI,and replaces the surface estimation model in the first version of the MERSI-II DT algorithm with this model,thereby establishing an improved MERSI-II DT aerosol retrieval algorithm.(3)In this paper,the first version of the MERSI-II DT algorithm and the improved algorithm are tested globally for three years,and the results are ground-based verification and comparative analysis.The results show that the correlation coefficient between the aerosol retrieval results of the improved MERSI-II DT algorithm and the ground-based observation results is 0.869,which is 0.01 higher than that of the first version of the algorithm;the root mean square error is 0.151,which is 0.011 lower than that of the first version of the algorithm;the proportion of sample points falling into EE is 65.40%,which is 3.84% higher than that of the first version of MERSI-II DT algorithm.In addition,on the urban surface,with the increase of UP,the correlation of the ground-based verification results of the improved algorithm is higher than that of the first version of MERSI-II DT algorithm,and the proportion of sample points falling into EE in each UP interval is also higher than that of the algorithm before the improvement,which is increased by 1.4%～42.17% respectively.From the comparison of the monthly average results of the global-scale aerosol optical depth,the monthly average of MERSI-II aerosol optical depth has a similar spatial distribution to MODIS,and the monthly average results are all above 0.9;compared with MODIS aerosol retrieval results,the aerosol retrieval results of the first version of MERSI-II algorithm have a certain overestimation,and the improved MERSI-II DT algorithm reduces this overestimation phenomenon to a certain extent.To sum up,the improved MERSI-II DT algorithm improves the accuracy of aerosol retrieval,especially in urban areas,which shows that the improvement of the DT algorithm in this paper is effective.By establishing a more accurate surface estimation model,the accuracy of aerosol retrieval can be better improved,which can provide a basis for further improvement of the MERSI-II terrestrial aerosol retrieval algorithm.46 pictures,5 tables,74 references... |
PDF Full Text Request