Review of Global Circulation Models Suggests Higher Temperatures in Store

January 27, 2014

A headline-making article by Sherwood, et al. in the January issue of Nature ascribes the wide spread in climate sensitivities reported by global circulation models (GCMs) to how they account for atmospheric convective mixing, a process that controls cloud formation in the lower troposphere.  The authors suggest that GCMs that properly account for the influence of convective mixing predict fewer low elevation clouds and correspondingly higher temperatures in response to the global buildup of CO2.   

For the past 40 years, GCMs developed by researchers have differed in their estimates of climate sensitivity – the Earth’s temperature response to a doubling of carbon dioxide – from as little as 1.5 to as much as 4.5 degrees Kelvin.  Climate change deniers have used the wide range in estimates to attack the science.  In that context, Sherwood et al., made news by concluding that models predicting less than a 3-degree response relied on “unrealistically weak” estimates of convective mixing.  As a consequence they over-estimated the formation of low clouds that exert a net cooling effect, producing unreliably low climate sensitivity estimates.

The low clouds in question include stratus, stratocumulus and nimbostratus, whose cloud bases are found below 2000 m in elevation.  They occur primarily in descending regions of the atmosphere, especially in the subtropics (between 20° and 40° latitude in both hemispheres). With their high albedo, low clouds reflect sunlight and exert a strong cooling effect.  Their formation and their lifecycle have not been well understood, however, and their representations in GCMs rely on poorly constrained assumptions. 

Convective mixing responsible for the variations in GCM simulations refers to heat and moisture transport from the near-surface boundary layer to the troposphere by small-scale vertical transport too small to be resolved by a GCM.  Sherwood et al. (2014) cites how boundary layer moisture content rises nearly 6% per degree Kelvin of surface temperature increase.  The contrast with the troposphere leads to convective mixing and drying of the boundary layer resulting in fewer low clouds and amplified surface warming.

While most reviewers have focused on Sherwood’s conclusion that GCM’s predicting a 1.5 to 3 degree climate sensitivity are incorrect, the same models may still have utility when looking at other outputs on various time and spatial-scales.



Sherwood, S.C., S. Bony and J.L. Dufresne (2014): Spread in model climate sensitivity traced to atmospheric convective mixing. Nature, 505, 37-42.

Shiogama, Hideo and T. Ogura. News and Views. Nature, 505, 34-35