Multi-directional polarized optical sensors are increasingly vital in passive remote sensing, deepening our understanding of global cloud properties. Nevertheless, uncertainty lingers on how these observations can contribute to our knowledge of cloud diversity. The variability in cloud PSD (Particle Size Distribution) significantly influences a wide array of cloud characteristics, while unidentified factors in RT (Radiative Transfer) may introduce errors into the cloud PSD retrieval algorithm. Therefore, establishing unified evaluation criteria for both optical device configuration and inversion methods is crucial. Our study, based on Bayesian theory and RT, assesses the information content of both cloud effective radius and effective variance retrieval, along with the key factors affecting their retrieval in multi-directional polarized observations, using the calculation of DFS (Degree of Freedom for Signals).We consider the process of solar incidence, cloud scattering, and sensor reception, and discuss the impact of various sensor configurations, cloud characteristics, and other components on the retrieval of cloud PSD. Correspondingly, we observed a 48% improvement in the information content of cloud PSD with the incorporation of multi-directional polarized measurements in the rainbow region. Cloud droplet concentration significantly influences inversion, but its PSD does not cause monotonic linear interference on information content. The blending of particle mixtures with different PSD has a significant negative impact on DFS. In cases where the AOD (Aerosol Optical Depth) is less than 0.5 and the COT (Cloud Optical Thickness) exceeds 5, the influence of aerosol and surface contributions on inversion can be neglected. Our findings would serve as a foundation for future instrument design improvements and enhancements to retrieval algorithms.
Read full abstract