Baroclinic Eddies and the Extent of the Hadley Circulation: An Idealized GCM Study
The Hadley circulation has widened over the past 30 years. This widening has been qualitatively reproduced in general circulation model (GCM) simulations of a warming climate. Comprehensive GCM studies suggest this widening may be caused by a poleward shift in baroclinic eddy activity. Yet the limited amplitude of the climate change signals analyzed so far precludes a quantitative comparison with theories. This study uses two idealized GCMs, one with and one without an active hydrologic cycle, to investigate changes in the extent of the Hadley circulation over a wide range of climates. The climates span global-mean temperatures from 243 to 385 K and equator-to-pole temperature contrasts from 12 to 100 K. Baroclinic eddies control the extent of the Hadley circulation across most of these climates. A supercriticality criterion that quantifies the depth of baroclinic eddies relative to that of the troposphere turns out to be a good indicator of where baroclinic eddies become deep enough to terminate the Hadley circulation. The supercriticality depends on meridional temperature gradients and an effective stability that accounts for the effect of convective heating on baroclinic eddies. As the equator-to-pole temperature contrast weakens or the convective static stability increases, convective heating increasingly influences the thermal stratification of the troposphere and the supercriticality. Consistent with the supercriticality criterion, the Hadley circulation contracts as meridional temperature gradients increase, and it widens as the effective static stability increases. The former occurs during El Niño and may account for the observed Hadley circulation contraction then; the latter occurs during global warming.
© 2015 American Meteorological Society. Manuscript received 27 May 2014, in final form 31 January 2015. We thank Paul O'Gorman for helpful clarifications on the effective static stability, Timothy M. Merlis for discussions on both linear baroclinic wave theories and ocean–atmosphere interactions, and Tobias Bischoff for his helpful comments on this work and its relation to ENSO. We are grateful for support by the National Science Foundation (Grants AGS-1019211 and AGS-1049201) and a Yale Climate and Energy Institute Fellowship. The simulations were performed on the Division of Geological and Planetary Sciences' Dell cluster at the California Institute of Technology (the program code for the simulations described in this paper is available at www.clidyn.ethz.ch/gcms/).
Published - jas-d-14-0152.1.pdf