CloudNET: evaluating clouds in six operational forecast models using cloud radar and lidar observations

In the EU CloudNet project, near continuous observations from radar, lidar and radiometers at three sites in Europe are used to systematically evaluate the clouds in six forecast models. The representation of clouds in GCMs is crucial for both climate and weather forecast models, but the vertically resolved observation needed to evaluate them have up to now been lacking. Synergetic algorithms for retrieving profiles of cloud fraction, liquid water content, ice water content and other microphysical variables have been applied to many years of data. For each model and each variable being evaluated, level 3 data sets are produced containing monthly and yearly statistics of the comparison with observations, including comparisons of long-term mean values and PDFs. A number of different skill scores are also calculated, enabling the degradation of the model clouds with forecast lead time to be determined. In the talk, the strengths and weaknesses of each of the various models will be discussed. The Cloudnet system has recently been extended to include the Atmospheric Radiation measurement (ARM) sites in both the United States and the Tropical West Pacific. The techniques described here are appropriate for model evaluation using spaceborne obervations.