Surface polarized reflectance is related to its optical properties and observation geometry. It disturbs remote sensing of aerosol. In this study, we differentiated the polarized reflectance of a natural surface based on the normalized difference vegetation index and scattering angle, and the interband correlation coefficients were generally higher than that without differentiation. The spectral relationship was determined based on normalized difference vegetation index and scattering angle and then employed for aerosol retrieval from polarization and anisotropy of reflectances for atmospheric sciences coupled with observations from a lidar measurements. The retrieved aerosol optical depths (AODs) were contrasted with Aerosol Robotic Network results. The correlation coefficient between the retrieved and Aerosol Robotic Network AODs increased from 0.723 to 0.836, and the root‐mean‐square error reduced from 0.348 to 0.245 after considering the differentiated relationship. Thus, the accuracy of the retrieved AOD values was improved.

No abstract

North China has one of the world's largest densities of atmospheric aerosol particles. The surface and aerosol properties in this area are complex. We evaluated the spectral relationship between surface polarized reflectances and introduced specific types of aerosol to improve the remote sensing of aerosols in this area. First, we searched for low-aerosol-loading areas (clean areas) based on Moderate resolution Imaging Spectroradiometer (MODIS) measurements between 2009 and 2011. Then, we recalculated the spectral relationships of surface polarized reflectance from corrected Polarization and Anisotropy of Reflectances for Atmospheric Sciences coupled with Observations from a Lidar (PARASOL) measurements. Finally, the re-derived relationships and aerosol clustering modes for East Asia were applied using the adaptive land-atmospheric decoupling (ALAD) algorithm to retrieve aerosol optical depth (AOD). We collected 914 PARASOL measured samples matched the site data of the aerosol robotic network (AERONET), within a wide AOD range between 2005 and 2013. The AOD trends from AERONET and PARASOL were similar. The slope of the regression line was 0.836, with a low intercept (0.032) and high correlation coefficient (0.854), and 53.6% of the retrieved AODs were within the range of the expected error. Compared with the MODIS daily AOD, we found that the variation in PARASOL results displayed a smooth tendency with the increase of AODs. The 914 sampling points were classified according to location and season to identify any discrepancies in the retrieved results. It was found that vegetation surfaces were responsible for most of the uncertainty due to their seasonal characteristics. Key Points:• We re-derived the spectral relationships of surface polarized reflectance over north China • Aerosol optical depths over north China were retrieved from polarized satellite observations using the new relationships • The seasonal variation of vegetable surface is the main disturbance in aerosol retrieval

No abstract