Retrieval of water cloud droplet size from multi-angle satellite imagery

The distribution of water droplets in clouds determines solar radiation and microphysical properties, thereby influencing the global energy balance and water cycle processes. Accurate retrieval of cloud water droplet distribution is of significant scientific value for improving the accuracy of weather forecasts, understanding feedback mechanisms in the climate system, and optimizing climate model predictions. This study employs multi-angle data from the second-generation Directional Polarization Camera on board the Chinese Gaofen-5 02 satellite (GF-5B DPC). By performing a detailed analysis of the polarization characteristics of water clouds and integrating the UNL-VRTM vector radiative transfer model, a lookup table for retrieving the water cloud droplet distribution. Using this approach, the effective particle radius and effective particle variance of water clouds over three surface types: forest, desert/semi-desert, and ocean, were successfully retrieved from the lookup table and polarization reflectance data. The retrieval results under various surface conditions were compared with the MOD06 cloud products. The correlation coefficient between the two sets of water cloud droplet distribution products ranged from 0.74 to 0.82, indicating strong consistency and validating the effectiveness and accuracy of the proposed retrieval algorithm. By utilizing GF-5B-DPC multi-angle polarization data in conjunction with the UNL-VRTM model, we analyzed the spatial distribution of water cloud droplets. The results demonstrate that the polarization measurements, particularly from GF-5B-DPC, exhibit high sensitivity to small particles, providing valuable insights into cloud microphysical properties.