The influence of heat transport on Arctic amplification

Abstract

The Arctic surface air temperature has warmed nearly twice as much as the global mean since the mid-20th century. Arctic sea ice has also been declining rapidly in recent decades. There is still discussion about how much of this Arctic amplification is caused by local factors, such as changes in surface albedo, versus remote factors, such as changes in heat transport from the midlatitudes. This thesis focuses mainly on the role of poleward heat transport on Arctic amplification. Most of the previous studies on this topic have defined ocean heat transport as the zonally averaged ocean heat transport at 65∘N or 70∘N, which ignores the physical pathways of heat into the Arctic and may include recirculation of heat in the North Atlantic. In this thesis, we define the ocean heat transport as the heat transport across five sections surrounding the Arctic, to create a closed domain in the Arctic. Previous studies on Arctic amplification have used either a single model run or have compared results from a multi-model ensemble. While the multi-model ensemble approach may potentially average out biases in individual models, the ensemble spread confounds the model differences and the internal climate variability. In this thesis, we investigate the Arctic amplification in the Community Earth System Model version 1 (CESM1) Large Ensemble. The CESM1 Large Ensemble includes 40 members that use the same model and external forcing, but different initializations. This simulates different climate trajectories that can occur in a given atmosphere-ocean-land-cryosphere system. We find that CESM1 Large Ensemble projects a large increase towards the end of the 21st century in ocean heat transport into the Arctic, and that the increase in ocean heat transport is significantly correlated with Arctic amplification. The main contributor to the increase in ocean heat transport is the increase across the Barents Sea Opening. The increase in Barents Sea Opening ocean heat transport is highly correlated with the decrease in sea ice in the Barents-Kara Sea region. We propose that this is because the increase in ocean heat transport melts the ice at the sea ice margin, which results in increased surface heat flux from the ocean and further local feedback through decreased surface albedo and increased cloud coverage. We also find that while the changes in atmosphere heat transport into the Arctic circle at 66.5∘N are on the same order as the changes in ocean heat transport, they are not correlated with Arctic amplification.