AbstractPatterns of correlations between the ocean and the atmosphere are examined using a high-resolution (1/12° ocean and ice, 1/3° atmosphere) ensemble of data assimilative, coupled, global, ocean–atmosphere forecasts. This provides a unique perspective into atmosphere–ocean interactions constrained by assimilated observations, allowing for the contrast of patterns of coupled processes across regions and the examination of processes affected by ocean mesoscale eddies. Correlations during the first 24 h of the coupled forecast between the ocean surface temperature and atmospheric variables, and between the ocean mixed layer depth and surface winds are examined as a function of region and season. Three distinct coupling regimes emerge: 1) regions characterized by strong sea surface temperature fronts, where uncertainty in the ocean mesoscale influences ocean–atmosphere exchanges; 2) regions with intense atmospheric convection over the tropical oceans, where uncertainty in the modeled atmospheric convection impacts the upper ocean; and 3) regions where the depth of the seasonal mixed layer (MLD) determines the magnitude of the coupling, which is stronger when the MLD is shallow and weaker when the MLD is deep. A comparison with models at lower horizontal (1/12° vs 1° and 1/4°) and vertical (1- vs 10-m depth of the first layer) ocean resolution reveals that coupling in the boundary currents, the tropical Indian Ocean, and the warm pool regions requires high levels of horizontal and vertical resolution. Implications for coupled data assimilation and short-term forecasting are discussed.