Schematic diagram showing how the surface fluxes are applied in the standard 1%CO2 run (top) and the FixedIce run (bottom) over Arctic sea-ice covered areas. In the FixedIce run (with the same 1%-per-year increase in atmospheric CO2), sea-ice loss (outlined by the dashed lines in the lower-right panel) is small, and the fluxes from the ice model are applied to the same ice fraction as in year 1 (that is, they are extended to the volume outlined by the dashed lines in the lower-right panel), and the atmosphere and ocean components only see a fixed ice cover (with seasonal cycle). However, the ice model still dynamically calculates the ice fraction and the fluxes over sea ice. The ice model does not see this artificial ice fraction change but it feels the changed surface fluxes and near-surface states resulting from this change, and this leads to much slower ice melting and greatly reduced Arctic amplification in the FixedIce run than in the standard 1%CO2 run. The main ice-atmosphere and water-atmosphere fluxes include sensible (SH) and latent (LH) heat fluxes, longwave (LW) and shortwave (not shown) radiative fluxes, and wind stress fluxes (not shown). The ice-ocean fluxes include heat (H), salt (S), freshwater (W), and wind stress (not shown) fluxes.