Different studies have used distinctive numerical approaches and coupled ocean-atmosphere models to address this problem. We put these studies on comparable footing using pattern scaling (Blackport and Kushner 2017) to separately estimate the part of the circulation response that scales with sea-ice loss in the absence of low-latitude warming from the part that scales with low-latitude warming in the absence of sea-ice loss. We consider well-sampled simulations from three different coupled ocean-atmosphere models (CESM1, CanESM2, CNRM-CM5), in which greenhouse warming and sea-ice loss are driven in different ways (sea ice albedo reduction/transient RCP8.5 forcing for CESM1, nudged sea ice/CO2 doubling for CanESM2, heat-flux forcing/constant RCP8.5-derived forcing for CNRM-CM5). Across these different simulations, surprisingly robust influences of Arctic sea-ice loss on atmospheric circulation can be diagnosed using pattern scaling. For boreal winter, the isolated sea-ice loss effect acts to increase warming in the North American Sub-Arctic, decrease warming of the Eurasian continent, enhance precipitation over the west coast of North America, and strengthen the Aleutian Low and the Siberian High. We will also discuss how Arctic free tropospheric warming might be enhanced via midlatitude ocean surface warming induced by sea-ice loss.
Less robust is the part of the response that scales with low-latitude warming, which, depending on the model, can reinforce or cancel the response to sea-ice loss. The extent to which a “tug of war” exists between tropical and high-latitude influences on the general circulation might thus be model dependent.