期刊名称：Journal of Advances in Modeling Earth Systems
电子版ISSN：1942-2466
出版年度：2019
卷号：11
期号：1
页码：127-148
DOI：10.1029/2018MS001461
出版社：John Wiley & Sons, Ltd.
摘要：The representation of stratocumulus by the atmospheric component of the Centre National de Recherches Météorologiques model version 6 (CNRM‐CM6‐1) is assessed. An Atmospheric Model Intercomparison Project‐type simulation is first used to document the main model errors, namely, a large lack of stratocumulus over the eastern part of tropical ocean basins. Short‐term hindcasts, following the Transpose‐Atmospheric Model Intercomparison Project framework, are then used to better assess the timescales associated with the cloud bias growth and to highlight the processes leading to them. These biases are shown to appear within only a few hours, independently of errors in the large‐scale circulation that set up within a few days. Key processes underlying the low‐cloud formation are thus mainly local and, to the first order, do not imply any feedback between the model physics and the large‐scale dynamics. As a consequence, short‐term hindcasts provide a relevant framework to investigate whether the low‐cloud underestimate is related to errors in the large‐scale state variables or to errors in the model parameterizations. Sensitivity tests highlight that the involved processes arise (1) mostly from misrepresentation of subgrid effects on cloud formation and (2) partly from biases in drying induced by cloud‐top entrainment mixing. Improvements in the representation of stratocumulus in the CNRM‐CM6‐1 model might thus be expected by including a more realistic subgrid‐scale temperature and moisture distribution, that would link convective and turbulence processes. Finally, this study confirms the potential of short‐term hindcasts, which provide a trustworthy framework to evaluate and develop climate model parameterizations.
