Fast and slow components of the extratropical atmospheric circulation response to CO2 forcing

• Published in: Journal of Climate
• Main Authors: Ceppi, Paulo, Zappa, Giuseppe, Shepherd, Theodore G., Gregory, Jonathan M.
• Format: Article in Journal/Newspaper
• Language: English
• Published: American Meteorological Society 2018
• Subjects: Antarctic; Pacific; Southern Ocean; The Antarctic; Antarc*
• Online Access: https://centaur.reading.ac.uk/72603/; https://centaur.reading.ac.uk/72603/7/jcli-d-17-0323.1.pdf; https://centaur.reading.ac.uk/72603/1/fast_slow_circulation_response_accepted.pdf

Poleward shifts of the extratropical atmospheric circulation are a common response to CO2 forcing in global climate models (GCMs), but little is known about the time dependence of this response. Here it is shown that in coupled climate models, the long-term evolution of sea surface temperatures (SSTs) induces two distinct time scales of circulation response to step-like CO2 forcing. In most Coupled Model Intercomparison Project phase 5 GCMs as well as in the multi-model mean, all of the poleward shift of the midlatitude jets and Hadley cell edge occurs in a fast response within 5 to 10 years of the forcing, during which less than half of the expected equilibrium warming is realized. Compared with this fast response, the slow response over subsequent decades to centuries features stronger polar amplification (especially in the Antarctic), enhanced warming in the Southern Ocean, an El Nino-like pattern of tropical Pacific warming, and weaker land-sea contrast. Atmosphere-only GCM experiments demonstrate that the SST evolution drives the difference between the fast and slow circulation responses, although the direct radiative effect of CO2 also contributes to the fast response. It is further shown that the fast and slow responses determine the long-term evolution of the circulation response to warming in the RCP4.5 scenario. The results imply that shifts in midlatitude circulation generally scale with the radiative forcing, rather than with global-mean temperature change. A corollary is that time slices taken from a transient simulation at a given level of warming will considerably overestimate the extratropical circulation response in a stabilized climate.