Climate Dynamics

, Volume 24, Issue 7–8, pp 655–666

Polar amplification of surface warming on an aquaplanet in “ghost forcing” experiments without sea ice feedbacks

  • V. A. Alexeev
  • P. L. Langen
  • J. R. Bates


Polar amplification of surface warming has previously been displayed by one of the authors in a simplified climate system model with no ice-albedo feedbacks. A physical mechanism responsible for this pattern is presented and tested in an energy balance model and two different GCMs through a series of fixed-SST and “ghost forcing” experiments. In the first ghost forcing experiment, 4 W/m2 is added uniformly to the mixed layer heat budget and in the second and third, the same forcing is confined to the tropics and extra-tropics, respectively. The result of the uniform forcing is a polar amplified response much like that resulting from a doubling of CO2. Due to an observed linearity this response can be interpreted as the sum of the essentially uniform response to the tropical-only forcing and a more localized response to the extra-tropical-only forcing. The flat response to the tropical forcing comes about due to increased meridional heat transports leading to a warming and moistening of the high-latitude atmosphere. This produces a longwave forcing on the high-latitude surface budget which also has been observed by other investigators. Moreover, the tropical surface budget is found to be more sensitive to SST changes than the extra-tropical surface budget. This strengthens the tendency for the above mechanism to produce polar amplification, since the tropics need to warm less to counter an imposed forcing.

Copyright information

© Springer-Verlag 2005

Authors and Affiliations

  • V. A. Alexeev
    • 1
  • P. L. Langen
    • 2
  • J. R. Bates
    • 3
  1. 1.International Arctic Research CenterUniversity of Alaska FairbanksFairbanksUSA
  2. 2.Niels Bohr InstituteUniversity of CopenhagenCopenhagen ODenmark
  3. 3.Mathematical Physics DepartmentUniversity College DublinBelfield, Dublin 4Ireland

Personalised recommendations