Chiung-Yin Chang, Timothy M Merlis 

Atmospheric macroturbulence transports energy down the equator-to-pole gradient. This is represented by diffusion in energy balance models (EBMs), and EBMs have proven valuable to understanding and quantifying the pattern of surface temperature change. They typically assume climate-state independent diffusivity, chosen to well represent the current climate, and find that this is sufficient to emulate warming response in general circulation models (GCMs). Here, we examine the role that changes in diffusivity play in the large-scale equator-to-pole contrast in surface warming in EBMs, motivated by theories for polar amplified warming. New analytic theories for two formulations of climate-state dependent diffusivity successfully capture the results of numerical EBM solutions. While existing GCM studies do not agree on the sign of simulated diffusivity changes with warming, they have never found enough diffusivity reduction to eliminate polar amplified warming. For reasonable choices of parameter values, the success of the new analytic theories reveals why.