Estimation of cirrus cloud effective ice crystal shapes using visible reflectances from dual-satellite measurements

[1] This study develops and examines a multiangle, multisatellite method for determining effective cloud particle shapes from reflectances observed at visible wavelengths. The technique exploits the significant differences in the various cloud particle shape phase functions near the backscatter direction to infer particle shape from a combination of views from a near-backscatter angle and a side scattering angle. Adding-doubling calculations confirm that the optimal viewing combinations include one near-backscatter angle and another between 60 and 150. Sensitivity to shape increases with solar zenith angle. A total of 28 collocated, visible images from pairs of currently operating meteorological satellites with the desired viewing combinations were analyzed for particle shape. Matching reflectances from images with optimal viewing angles clearly separates water droplet from ice crystal clouds. Reflectance pairs from matched pixels containing ice crystals can be explained by the range of selected microphysical models. The most common retrieved shapes correspond to combinations of hexagonal compacts (aspect ratio of unity), hexagonal columns, and bullet rosettes. Although no single microphysical model can account for the observed variability, taken together, the models used for retrievin