Validation of TOA radiation and clouds simulated by different versions of LMD GCM in comparison to satellite and ground-based data amg

The validation of cloudiness and radiative fluxes simulated by three versions of the Laboratoire de Meteorologie Dynamique (LMD) general circulation model (GCM) is performed against satellite and ground- based observations. These GCM versions include quite different cloud schemes and surface albedo representa- tions. One model version implements the solar radiation diurnal cycle, while the others do not. Additionally, model versions are run with different spatial resolutions. Model validation is based on quite simple methodol- ogy to analyze the simulations for only three prechosen cross-sections in equatorial and middle latitudes, instead of studying the whole model field. It is shown that over middle latitudes simulated discrepancies exhibit a strong seasonal dependence. This is closely related to the representation of water vapor, cloudiness, and sur- face albedo in the model. In particular, errors in the simulated cloud radiative forcing (CRF) portray those for low-level cloudiness (for short-wave radiation) or high-level cloudiness (for long-wave radiation). As a whole, the errors in short-wave CRF are larger than those of the long-wave CRF, and are strongly amplified in summer. In the simulated total cloud amount, main errors are related to low-level cloudiness. It is also shown that the causes of the model deficiencies differ between versions. In some cases, errors coming from different origins could compensate each other. Although quite simple, the methodology followed in this intercomparison proved to be an effective approach for tracking systematic errors in a set of different model versions.