Evaluation of potential biases in CloudSat hydrometeor frequency of occurrence using ARM and Lindenberg cloud radar observations

Description

The CloudSat-CALIPSO radar-lidar combination is currently the best way to characterize the vertical distribution of hydrometeor frequency of occurrence on a global scale. Nevertheless, some characteristics of the instruments onboard those satellites imply that probably not all clouds are sampled by this combination. One particular difficult situation is that of low-level clouds near the oceanic or continental surface, because the lidar signal generally will be extinguished rapidly inside these water clouds, and the CloudSat radar signal may not have the required sensitivity to detect the often small-sized droplets they contain. Also the CloudSat cloud signal could be mixed with ground clutter, making the detection of these clouds difficult. The objective of this study is to characterize the errors associated with these spaceborne instrumental limitations on the hydrometeor frequency of occurrence. The underlying motivation is also to provide error bars for crucial quantitative studies ongoing with the CloudSat radar (such as the radiative budget calculations, the derivation of cloud climatologies, and the evaluation of large-scale models). In order to quantify these errors, we compare the mean vertical profiles of hydrometeor frequency of occurrence as derived over cloud radar sites at three different latitude bands (Barrow, Alaska; Lindenberg, Germany; Darwin, Australia) with the same profiles derived from extractions of CloudSat-CALIPSO observations within a 200-kilometer range from these sites.