Arctic climate variability as derived from coupled AOGCM REMO/MPIOM

A regional coupled high resolution ocean – sea ice – atmosphere model was implemented to investigate inter-annual variability in the Nordic Seas and Arctic ocean. Its ocean component (MPIOM) is formally global but has a high resolution in the region of interest and contains a state-of-the-art sea ice model. The regional atmospheric component (REMO) covers the North Atlantic, Europe and the entire Arctic. Several ensemble experiments were carried out for the years 1958-2007 using NCEP reanalysis data as lateral boundary values for REMO and as atmospheric forcing outside the coupled domain for the ocean. To investigate the impact of large atmospheric processes as well as internal variability of the regionally coupled ocean-atmosphere system, two different REMO model configurations were chosen. The smaller domain includes Arctic Ocean and the NAO pattern, while the larger one was extended into the Eurasian continent and includes the Siberian High. Although Arctic sea ice dynamics differ among ensemble members - mean correlation in ice export through the Fram strait is about 0.3 -, it is shown that strong ice export events are preconditioned by the large scale atmospheric circulation. On the other hand sea ice dynamics depend on the state of the coupled ocean-atmosphere system which can accelerate or delay the maximum ice export through the Fram Strait for 1-2 years. The use of the larger REMO domain in the coupled system provides higher ‘internal' ensemble variability and the possibility to investigate the role of Siberian High in the Arctic climate system.