期刊名称：Journal of Advances in Modeling Earth Systems
电子版ISSN：1942-2466
出版年度：2017
卷号：9
期号：8
页码：2883-2905
DOI：10.1002/2017MS001111
出版社：John Wiley & Sons, Ltd.
摘要：Tropical cloud and circulation changes are large sources of uncertainty in future climate change. This problem owes partly to the scale separation between large‐scale tropical dynamics ( km) and convective dynamics ( km), which generally requires parameterizing convection in models that resolve large‐scale dynamics, or parameterizing (or omitting) large‐scale dynamics in models that permit convection. Here we discuss simulations of radiative‐convective equilibrium (RCE) across a wide range of surface temperatures in long‐channel geometry—where the domain size and resolution marginally resolve both large‐scale dynamics and convection. Self‐aggregation of convection in these simulations spontaneously produces realistic dynamical regimes of large‐scale vertical motion. The circulation weakens with surface warming but changes in the degree of self‐aggregation depend on the metric that is used; there is no obvious trend in aggregation with warming. Surface warming causes an upward shift and decrease in area of high clouds, and a sharp decline in midlevel clouds, but no systematic trend in low cloud cover. We introduce a method for approximate radiative kernel feedback analysis in RCE, and apply it to both simulations in long‐channel geometry and in a smaller square domain. The kernel‐corrected cloud feedback is positive but its magnitude varies across temperatures. Compared to simulations that do not have aggregation, there is a more negative net feedback due to the effects of aggregation on relative humidity and cloud cover. These results are consistent with the hypothesis that self‐aggregation moderately reduces climate sensitivity.
