Find in Library

Ground-based sky imagers have become a valuable tool for cloud monitoring as they allow to observe the local sky in great detail due to their temporally and spatially high-resolution imagery. Set up in pairs, they can be used to estimate both cloud geometry and motion. While considerable effort is currently devoted to their use for cloud shadow forecasting and short-term prediction of solar irradiance for renewable energies, their potential for vertical wind profiling and cloud evolution monitoring has not yet received much attention. We evaluate the accuracy and reliability of such retrievals from two sky imagers in stereo mode by comparing derived 3D cloud motion vector fields with observations from a wind lidar. We find that the wind vectors derived from the detected motion of cumulus clouds and multiple cloud layers can be retrieved with a quality comparable to a wind lidar. Differences in wind speed and direction from both methods remain mostly below 1 to 3 ms−1 in wind speed and 6° to 20° in wind direction, depending on cloud type.