Can a Convective Cloud Feedback Help to Eliminate Winter Sea Ice at High CO2 Concentrations?

Abstract

Abstract Winter sea ice dramatically cools the Arctic climate during the coldest months of the year and may have remote effects on global climate as well. Accurate forecasting of winter sea ice has significant social and economic benefits. Such forecasting requires the identification and understanding of all of the feedbacks that can affect sea ice. A convective cloud feedback has recently been proposed in the context of explaining equable climates, for example, the climate of the Eocene, which might be important for determining future winter sea ice. In this feedback, CO2-initiated warming leads to sea ice reduction, which allows increased heat and moisture fluxes from the ocean surface, which in turn destabilizes the atmosphere and leads to atmospheric convection. This atmospheric convection produces optically thick convective clouds and increases high-altitude moisture levels, both of which trap outgoing longwave radiation and therefore result in further warming and sea ice loss. Here it is shown that this convective cloud feedback is active at high CO2 during polar night in the coupled ocean–sea ice–land–atmosphere global climate models used for the 1% yr−1 CO2 increase to the quadrupling (1120 ppm) scenario of the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report. At quadrupled CO2, model forecasts of maximum seasonal (March) sea ice volume are found to be correlated with polar winter cloud radiative forcing, which the convective cloud feedback increases. In contrast, sea ice volume is entirely uncorrelated with model global climate sensitivity. It is then shown that the convective cloud feedback plays an essential role in the elimination of March sea ice at quadrupled CO2 in NCAR’s Community Climate System Model (CCSM), one of the IPCC models that loses sea ice year-round at this CO2 concentration. A new method is developed to disable the convective cloud feedback in the Community Atmosphere Model (CAM), the atmospheric component of CCSM, and to show that March sea ice cannot be eliminated in CCSM at CO2 = 1120 ppm without the aide of the convective cloud feedback.