In polar regions, clouds cover large areas of snow and sea ice. In that case, the retrieval of snow and cloud properties is difficult as the optical properties of clouds and snow are similar. Based on radiative transfer simulations, the uncertainties of cloud retrieval with respect to the assumed snow albedo (snow grain size, rsnow) are quantified. Significant biases in retrieved cloud optical thickness (Δτ ≤ 3) and cloud droplet effective radius (Δreff ≤ 3 μm) are found for typical snow grain sizes of 200 μm and 50 μm. Based on the spectral absorption characteristics of snow and liquid clouds a combination of wavelengths was found which allowed to separate the impact of clouds and snow on the reflected radiation measured above the clouds. While snow grain size dominates the absorption at a wavelength of 1.04 μm, information on cloud optical thickness and cloud particle effective radius was extracted at wavelengths of 1.65 μm and 2.1 μm, respectively. Based on these wavelengths an algorithm for a simultaneous retrieval of τ, reff or rsnow was developed. Using ratios instead of absolute radiances reduces the uncertainties significantly. The new algorithm was applied to a specific case observed during the VERDI campaign where stratocumulus clouds were located above an ice edge. It could be shown that the method works also over water surfaces and provides similar cloud optical properties above ice covered and ice free surfaces. In addition the snow grain size could be derived in cloud covered areas.
© 2016 Optical Society of AmericaPDF Article