AbstractsAstronomy & Space Science

In this thesis spectral solar zenith radiances are analyzed which were obtained from ship-based measurements over the Atlantic ocean. In combination with high-resolution lidar and microwave remote sensing optical and microphysical cloud properties were retrieved using spectral radiation data. To overcome problems of existing transmissivity-based cloud retrievals, a new retrieval algorithm is introduced which circumvents retrieval ambiguities and reduces the influence of measurement uncertainties. The method matches radiation measurements of ratios of spectral transmissivity at six wavelengths with modeled transmissivities. The new retrieval method is fast and accurate, and thus suitable for operational purposes. It is applied to homogeneous and inhomogeneous liquid water and cirrus clouds. The results from the new algorithm are compared to observations of liquid water path obtained from a microwave radiometer, yielding an overestimation for thick liquid water clouds but a slight underestimation for thin clouds. A statistical analysis of retrieved cloud properties during three Atlantic transects is introduced. Similar characteristics of cloud properties are found in the mid latitudes and northern subtropics but the large variability of meridional distribution in the remaining regions imply the prevailing influence of weather systems compared to typical cloud distributions. With about 63% homogeneous stratocumulus clouds are found to be the prevailing cloud type over ocean, while scattered and inhomogeneous liquid water clouds amount to 16% and 21%, respectively. All analyzed distributions are affected by an increased frequency of small values of cloud properties caused by 3D radiative effects. The comparison with satellite-based and ship-based cloud retrievals along the cruise track show comparable results for the cloud optical thickness with limitations for thick liquid water clouds. The meridional distribution of effective radius agreed within the uncertainties of both methods, however, the satellite-derived values are biased toward larger mean values.