Tellus: Series A, Dynamic Meteorology and Oceanography (May 2016)

Surface air temperature variability and trends in the Arctic: new amplification assessment and regionalisation

  • Ola M. Johannessen,
  • Svetlana I. Kuzmina,
  • Leonid P. Bobylev,
  • Martin W. Miles

DOI
https://doi.org/10.3402/tellusa.v68.28234
Journal volume & issue
Vol. 68, no. 0
pp. 1 – 12

Abstract

Read online

Arctic amplification of temperature change is theorised to be an important feature of the Earth's climate system. For observational assessment and understanding of mechanisms of this amplification, which remain uncertain, thorough and detailed analyses of surface air temperature (SAT) variability and trends in the Arctic are needed. Here we present an analysis of Arctic SAT variability in comparison with mid-latitudes and the Northern Hemisphere (NH), based on an advanced SAT dataset – NansenSAT. We define an index for the Arctic amplification as the ratio between absolute values of the Arctic (65–90°N) and NH 30-yr running linear SAT trends. It is demonstrated that the temperature amplification in the Arctic is characteristic not only for the recent warming but also the early 20th century warming (ETCW) and subsequent cooling. The amplification appears to be weaker during the recent warming than in the ETCW, simply because the index values reflect the more pervasive nature of the recent warming that reflects the background of anthropogenic global warming. We also produced a new Arctic regionalisation created from hierarchical cluster analysis, which identifies six major natural regions in the Arctic that reflect SAT variability. Statistical comparison with several climate indices shows that the Atlantic Multidecadal Oscillation (AMO) is the mode of variability that is most significantly associated with the amplified warming–cooling in the Arctic, with a stronger correlation during the ETCW and recent warming than during the intermediate period. Regionally, differences are identified in terms of annual and seasonal rates of change and in their correlations with modes of variability.

Keywords