资源环境科技发展态势分析平台

In this study, the dependence of cloud radiation on several cloud subgrid-scale structures that are unresolved by conventional climate models is explored, using datasets produced by a cloud-resolving model for both stratiform and convective clouds. These subgrid-scale structural parameters to be investigated are decorrelation length L_cf for overlapping of cloud fraction, L_cw for overlapping cloud condensate, and shape parameter v for measuring cloud inhomogeneity. It is found the decorrelation length L_cw has a similar role to the inhomogeneity parameter v in modulating cloud radiative effects (CREs), with increasing (decreasing) L_cw generally leading to a weakening (enhancing) of CREs, which is equivalent to increasing (decreasing) v in individual layers. However, the uncertainty of CREs caused by changes in L_cw is relatively smaller than that induced by changes in v, with the former about twice smaller than the latter. For fractional clouds with multiple layers, it is revealed that cloud overlap parameter L_cf exerts impacts on radiative transfer process first and foremost, and the inhomogeneity parameter v is of secondary importance, followed by the alignment parameter L_cw, which renders moderate modulation that can not be neglected. This study also highlights the importance of the vertically varying structure of v in cloud radiation. The simulated CREs and radiative heating fields can be significantly biased if these varying characteristics are missing, even if the obtained variability of cloud water path is close to reality. This calls for more attention to the vertical structure of v in parameterization development, which is however neglected or poorly represented in the literature.