The performance of several numerical ocean models is assessed with respect to their simulation of sea surface height (SSH) in the Arctic Ocean, and the main patterns of SSH variability and their causes over the past 40 years (1970-2009) are analyzed. In comparison to observations, all tested models broadly reproduce the mean SSH in the Arctic and reveal a good correlation with both tide gauge data and SSH anomalies derived from satellite observations. Although the models do not represent the positive Arctic SSH trend observed over the last two decades, their interannual-to-decadal SSH variability is in reasonable agreement with available measurements. Focusing on results from one of the models for a detailed analysis, it is shown that the decadal-scale SSH variability over shelf areas and deep parts of the Arctic Ocean have pronounced differences that are determined mostly by salinity variations. A further analysis of the three time periods 1987-1992, 1993-2002, and 2003-2009, corresponding to the transition times between cyclonic and anticyclonic regimes of the atmospheric circulation over the Arctic, revealed an unusual increase of SSH in the Amerasian basin during 2003-2009. Results from this model support the recent finding that the increase is caused mainly by changes in freshwater content brought about by the freshwater export through the Canadian Arctic Archiplago and increased Ekman pumping in the Amerasian basin and partly by lateral freshwater transport changes, leading to a redistribution of low-salinity shelf water. Overall, we show that present-day models can be used for investigating the reasons for low-frequency SSH variability in the region.
Journal of Geophysical Research-oceans, 2014, Vol 119, Issue 12, p. 8936-8954