Can multi-angular polarimetric measurements in the oxygen-A and B bands improve the retrieval of aerosol vertical distribution?

Based on the multi-angle measurements at the top of the atmosphere including radiance and polarization in O2 A and B absorption bands simulated by forward model UNL-VRTM, the information content and posterior uncertainties about the altitude of peak aerosol extinction (Hpeak ) and half-width (w) characterizing quasi-Gaussian aerosol vertical profiles are analyzed for different observation combinations. Although the information content described as degree of freedom for signal (DFS) for Hpeak and w increases with the number of angles, this increase reaches a limit when more than around six angles are used, regardless of surface type (ocean, vegetation or soil). Due to high surface reflectance of vegetation in O2 A band, the DFS of Hpeak from multi-angle measurements could be lower than 0.8, whereas the DFS > 0.9 for soil and ocean. At a single angle, polarization measurement is more sensitive to Hpeak than radiance when Hpeak is near the surface. Compared with single-angle radiance in O2 A band, adding multi-angle radiances reduces retrieval uncertainty for Hpeak by 10–20%, larger than that from only adding polarization at a single angle (< 10%), especially over vegetation. Over vegetated surface, the multi-angle radiances in O2 A and B band have comparable information, whereas for single angle, O2 B band has richer information than O2 A, especially for lower Hpeak . Multi-angle radiances and polarization in the O2 A band as well as radiances in O2 B band reduce the posterior uncertainty of w more than Hpeak , especially at large AOD, while the polarized measurements in O2 B band are more useful for Hpeak retrieval. The most significant impact of O2 B band radiances is to mitigate aerosol height retrieval uncertainties due to uncertainties in the surface parameters, whereas polarized measurements reduce errors from all error sources including measurements and forward model parameters.