EVALUATION OF RADIATIVE PARAMETERIZATIONS USING AN EXPLICIT CLOUD MICROPHYSICAL MODEL

Based on the simulations with a 3-D large-eddy simulation model of mar ine cloud-topped boundary layer that includes explicit cloud physics f ormulation, we have evaluated the effect of spatial inhomogeneities in cloud macro- and microstructure on the performance of parameterizatio ns of optical depth commonly used in large-scale models. We have shown that an accurate parameterization of the grid average optical depth a lone is not sufficient for correct determination of cloud transmittanc e to solar radiation due to the non-linear dependence between these tw o variables. The problem can be solved by introducing the ''equivalent '' value of optical depth that differs from the ordinarily defined mea n optical depth by a factor alpha(t) that depends on the degree of clo ud inhomogeneity and ranges from about 2 in the cumulus case to about 1.3 in the stratiform case. The accuracy of cloud optical depth parame terizations commonly employed in large-scale models has been evaluated using the data from the explicit microphysical model as a benchmark f or comparison. It has been shown that in the cumulus cloud case the pa rameterized expressions can err by as much as 100%. The error is small er for more uniform stratiform clouds, where the error for some parame terizations varied in the 10-40% range. The best results are given by parameterizations that account for vertical stratification of paramete rs on which they are based. However, the error given by a particular p arameterization varies and is different at cloud and surface levels. T he results show the limitations of the existing simplified parameteriz ations and illustrate the scope and complexity of the cloud radiation parameterization problem.