Arctic amplification under global warming of 1.5 and 2 °C in NorESM1-Happi
Differences between a 1.5 and 2.0 ∘C warmer climate than 1850 pre-industrial conditions are investigated using a suite of uncoupled (Atmospheric Model Intercomparison Project; AMIP), fully coupled, and slab-ocean experiments performed with Norwegian Earth System Model (NorESM1)-Happi, an upgraded version of NorESM1-M. The data from the AMIP-type runs with prescribed sea-surface temperatures (SSTs) and sea ice were provided to a model intercomparison project (HAPPI – Half a degree Additional warming, Prognosis and Projected Impacts; http://www.happimip.org/, last access date: 14 September 2019). This paper compares the AMIP results to those from the fully coupled version and the slab-ocean version of the model (NorESM1-HappiSO) in which SST and sea ice are allowed to respond to the warming, focusing on Arctic amplification of the global change signal. The fully coupled and the slab-ocean runs generally show stronger responses than the AMIP runs in the warmer worlds. The Arctic polar amplification factor is stronger in the fully coupled and slab-ocean runs than in the AMIP runs, both in the 1.5 ∘C warming run and with the additional 0.5 ∘C warming. The low-level Equator-to-pole temperature gradient consistently weakens more between the present-day climate and the 1.5 ∘C warmer climate in the experiments with an active ocean component. The magnitude of the upper-level Equator-to-pole temperature gradient increases in a warmer climate but is not systematically larger in the experiments with an active ocean component. Implications for storm tracks and blocking are investigated. We find considerable reductions in the Arctic sea-ice cover in the slab-ocean model runs; while ice-free summers are rare under 1.5 ∘C warming, they occur 18 % of the time in the 2.0 ∘C warming simulation. The fully coupled model does not, however, reach ice-free conditions as it is too cold and has too much ice in the present-day climate. Differences between the experiments with active ocean and sea-ice models and those with prescribed SSTs and sea ice can be partially due to ocean and sea-ice feedbacks that are neglected in the latter case but can also in part be due to differences in the experimental setup.