Satellite Versus Gcm-Simulated Radiation Balance: Comparisons and Implications for Climate Modeling.

Measurements of the earth radiation budget (ERB) from space-borne instruments and long-term simulations of climate with general circulation models (GCMs) make complementary contributions to improve our understanding of cloud-radiation-climate interactions and the role of clouds in the atmospheric general circulation. The first step of an increased use of three-dimensional climate models to study the forcing of clouds on climate is the validation against observations of simulated climates, including their temporal and spatial variability. The global distributions of the planetary radiation balance components and cloudiness computed from long-term climate simulations with the NCAR Community Climate Model (NCAR CCM) were compared against those measured during the Nimbus-7 satellite mission. The latest version of the NCAR CCM, also known as version CCM1, was used. The model-generated fields were obtained from a 15-year run including a seasonal cycle. Daily broad -spectral-band narrow angle field-of-view ERB measurements and estimates of cloud amounts derived from satellite radiances were taken during the time period between May 1979 and June 1980. Our comparison between model-generated and satellite -observed radiation fields helped identify a major discrepancy in the simulation of climate with the NCAR CCM. The model reproduced successfully the mean steady state of the radiation budget components, especially at long wavelengths, but failed to simulate their temporal variability. Analyses of the model-generated cloudiness and its forcing on radiation demonstrated that the high frequency of occurrence of clouds explained most of the difference in the temporal variability between model and observations. It was showed that the prediction of clouds in term of the large-scale relative humidity field and the atmospheric stability, and the treatment of condensation as a complete rainout process, maintained a decoupling between clouds, radiation, and the hydrologic cycle. The impacts of the cloud prediction scheme and the parameterization of the interactions between clouds, and the radiative, dynamic, and thermodynamic processes were separately analyzed from two different climate experiments with CCM1.