Robust increase in ocean heat transport into the Arctic under greenhouse gas forcing

The Arctic has warmed twice as rapidly as the rest of the world over recent decades - a phenomenon known as Arctic Amplification. Yet, climate models underestimate the current rate of Arctic warming and sea ice loss (Stroeve et al. 2007), and produce a wide range of 21st century predictions. While a variety of atmospheric and ocean processes control Arctic climate, recent research shows that the ocean plays a fundamental role in its accelerated warming. Historical climate model simulations suggest that Arctic Amplification over the coming century depends sensitively on future changes in ocean heat transport (OHT) into the Arctic (Hwang et al. 2011).

We calculate poleward OHT changes in CMIP5 simulations in which CO₂ is abruptly quadrupled. We find a robust increase in northward OHT across 70^o N in all CMIP5 models, even as a weakening Atlantic Meridional Overturning Circulation (AMOC) drives a decrease in northward OHT at 45^oN. We further find that while polar amplification and changes in OHT across 70^o N are positively correlated, the correlation is small and steadily decreases with time, becoming statistically insignificant after about a century. These results are in contrast to those of Hwang et al. 2011, who find a high correlation between the two quantities over the 21st century. We speculate that this may be due to differences in the OHT calculation, where Hwang et al. 2011 diagnose OHT from surface heat fluxes above sea ice, without accounting for the spatial structure of ocean heat storage. Our results suggest that a substantial portion of anomalous Arctic OHT occurs at depth rather than at the surface, and thus OHT explains less of the variance in polar amplification than do surface heat fluxes. We further examine the mechanisms relating Arctic OHT and AMOC under natural variability using CMIP5 pre-industrial control simulations. These indicate that increased poleward OHT across 70^o N is correlated with strong AMOC (Zhang et al. 2015), at odds with changes seen under greenhouse gas forcing. We hypothesize that different mechanisms link AMOC and Arctic OHT under natural variability than under anthropogenic forcing.