Changes in Extratropical Diffusivity with Uniform Surface Warming

Mixing-length theory is a useful framework for thinking about extratropical eddy transport. Here we use the framework to investigate diffusivity changes in response to uniform warming using 10 prescribed-SST aquaplanet general circulation models from the CMIP5 archive. In contrast to previous work that assumed fixed diffusivity, we examine changes in moist static energy diffusivity driven by MSE transport and low-level MSE gradient changes. We also investigate the extent to which diffusivity changes can be attributed to SST changes alone, as opposed to cloud and water vapor radiative feedbacks, using two of these models.

Defining diffusivity as D = -H/G, where H is the MSE transport and G is a low-level MSE gradient, we analyze changes in the latitude of (local) maximum diffusivity. The multi-model mean and most individual model diffusivity profiles have a midlatitude local maximum. In response to a uniform surface warming of 4K, this latitude of maximum multi-model mean diffusivity shifts poleward by 5.2 degrees, reminiscent of the well-known projection of a poleward shift of storm tracks with global warming. The shift is primarily driven by changes to G (84.6%), with H contributing another 11.8% and the remainder (3.5%) accounted for by the nonlinear term.

The extent to which SST changes and cloud and water vapor radiative feedbacks influence the diffusivity shift is assessed using simulations with the MPI and IPSL models where cloud and water vapor radiative fields are locked to their climatological values. In both models, all three processes are found to contribute to the poleward shift. They contribute roughly equally for the MPI model (31.7% SST, 30.7% clouds, 37.6% water vapor) but the cloud and SST effects dominate in the IPSL model (46.4% SST, 46.6% clouds, 7.0% water vapor).

Overall, we conclude that changes in the low-level MSE gradient play a key role in shifting the midlatitude local maximum of MSE diffusivity poleward with warming. This diffusivity shift cannot be understood purely in fluid dynamic or thermodynamic terms—instead, cloud and/or water vapor radiative feedbacks are important as well.