Cloud simulations with the Max Planck Institute for Meteorology general circulation model ECHAM4 and comparison with observations

In this study, cloud parameters as simulated by the latest version of the Max Planck Institute for Meteorology general circulation model are documented and compared with observations. The model simulations generally agree with the observed spatial distribution and temporal variation of the total cloud amount. There are, however, biases in the details. Underestimation in the total cloud amount found over the midlatitude oceans in summer leads to significant biases in the simulated radiation budget. Considerable uncertainties of the observed total cloud amount in the polar region call for improved measurement techniques for further model validation. With a similar mean state of total cloud amount in the tropics between model and observation, fewer daily and interannual variabilities are found in the model. Despite large uncertainties in the current cloud liquid water path retrievals, the main pattern and magnitude of the space-time distribution of cloud liquid water path is reseasonably well reproduced by the model. Lack of contrast between simulated cloud liquid water path in the subtropics and midlatitudes and failure to capture the observed summer local maxima of cloud liquid water path in the subtropical eastern ocean basins are the major discrepancies found in the model in comparison with the observations. The systematic differences in the magnitude of cloud liquid water path retrievals as shown in the comparison reveal a need for careful calibration of satellite retrieval algorithms. With a simple approach to prescribing land-sea contrast in cloud droplet number concentrations the simulated distribution of effective radii of cloud water droplets in the tropics is in good agreement with observations. However, some physical processes (e.g., precipitation efficiency and air mass advection) in regulating cloud droplet number concentrations cannot be resolved by the simple prescription. The model simulation of cloud effective drop radii over the midlatitude oceans reveals errors larger than those in other regions of the globe.