Characterizing Ice Cloud Particle Shape and Surface Roughness from Polarimetric Satellite Observations

The single scattering properties of ice cloud particles are inferred from spaceborne multi-angle satellite sensors with two newly developed noise-resilient retrieval techniques. The first presented method parameterizes the phase function and phase matrix elements by a few parameters to implement the maximum likelihood estimation in the retrieval system. The second method retrieves the renormalized phase function as a difference from a known phase function. The effect of noise is more predictable for both methods than the conventional “best-fit” method, which selects the best-fitting shape and surface roughness from a predetermined particle set. The first method is applied to the data from the Polarization and Directionality of the Earth’s Reflectance (POLDER) sensor. The retrieval results indicate that long column shape (ratio of basal face diameter to prism height greater than 9) with surface roughness parameter between 0.1 and 0.5 represents the extratropical observations well. Weak temperature dependence of the surface roughness is found in the extratropical data stratified by the cloud top temperature. The tropical retrieval was not successful, and the second method is applied to the Multi-angle Imaging Spectroradiometer (MISR) data. Short hexagonal column particles or their aggregates are found to match with estimated renormalized phase function. In addition to these results, the surface roughness simulation is summarized and the derivation of the δ-fit truncation technique for polarimetric radiative transfer is included