Variability at all timescales, including low-frequency variability, is found in the North Atlantic sector in a 300-yr control integration of the coupled ocean–atmosphere general circulation model (CGCM) ECHAM4/OPYC3. The atmospheric variability is dominated by the North Atlantic Oscillation (NAO). Only rather weak spectral peaks are superimposed on the “white noise” power spectrum of sea level pressure and on the essentially “red noise” spectrum of SST in highly active regions. Replacing the full ocean model with a 50-m fixed-depth mixed layer ocean (MLO) and coupling it to the atmosphere yields qualitatively and quantitatively very similar power spectra of the NAO index. Net surface heat fluxes, describing the coupling between the ocean and the atmosphere for the long-term variations (>10 yr) of the NAO are much weaker in the MLO model, but show general agreement in both simulations regarding spatial distributions. This spatial agreement with respect to NAO variability occurs even though the ass... Abstract Variability at all timescales, including low-frequency variability, is found in the North Atlantic sector in a 300-yr control integration of the coupled ocean–atmosphere general circulation model (CGCM) ECHAM4/OPYC3. The atmospheric variability is dominated by the North Atlantic Oscillation (NAO). Only rather weak spectral peaks are superimposed on the “white noise” power spectrum of sea level pressure and on the essentially “red noise” spectrum of SST in highly active regions. Replacing the full ocean model with a 50-m fixed-depth mixed layer ocean (MLO) and coupling it to the atmosphere yields qualitatively and quantitatively very similar power spectra of the NAO index. Net surface heat fluxes, describing the coupling between the ocean and the atmosphere for the long-term variations (>10 yr) of the NAO are much weaker in the MLO model, but show general agreement in both simulations regarding spatial distributions. This spatial agreement with respect to NAO variability occurs even though the ass...