An Aerosol Retrieval Algorithm Over Land From Multispectral Single-Viewing Measurements of Intensity and Polarization

Currently, there is no generalized optimization algorithm specifically designed for single-viewing-angle polarimeters. To better explore the potential retrieval capabilities of such sensors, we have developed a generalized optimized aerosol retrieval framework. To further validate its effectiveness, we applied this framework to Chinese particulate observing scanning polarimeter (POSP), launched in September 2021. POSP is the first high-precision spaceborne cross-track scanning polarimeter, featuring a broader spectral range (UV-VIS-NIR-SWIR) and full-spectrum polarization capabilities. POSP measurements between January 2022 and December 2022 have been used to generate an aerosol product. Validation with ground-based AErosol RObotic NETwork (AERONET) aerosol products shows good consistency, for AOD with an root mean square error (RMSE) of 0.086, bias of 0.003, and the fraction within the Global Climate Observing System (GCOS) requirements is 58.73%; for AE with an RMSE of 0.404, and the fraction within the expected error (EE) is 71.61%; for SSA with RMSE of 0.067, and the fraction within the EE is 76.06%. Additionally, we compared POSP and MODIS/Terra AOD products. The spatial distributions of POSP and MODIS-DB AOD were similar, exhibiting good consistency in highly polluted areas, while the POSP AOD revealed more detailed information. The information content analysis shows that adding polarimetric measurements can effectively improve the degrees of freedom for signal (DFSs) available for aerosol retrieval, especially for coarse-mode dominated aerosols. This is confirmed by the actual retrieval results showing that adding polarimetric measurements effectively reduces underestimation. Error analysis shows that adding polarization results in a reduction in the uncertainty of POSP AOD at AOD <0.2 and over sparsely vegetated areas (NDVI <0.1). The validation and analysis results both demonstrate that the retrieval framework used in this study effectively utilizes both intensity and polarization, and that adding polarization significantly enhances the retrieval accuracy.