In the central North Atlantic Ocean there are large decadal-scale fluctuations of sea level and of the depth of the thermocline. This variability can be explained by low-frequency Rossby waves forced by wind. The authors have used a simple model of ocean fluctuations driven by the COADS wind-stress curl and find that their model results and hydrographic data, to the limited extent they can be compared meaningfully, agree rather well. The variation in wind curl over the ocean leads to surprisingly large north–south variability in the computed oceanic response over the subtropical gyre. The peak-to-peak sea level differences are as great as 20 cm and persist for many years. These large variations could induce major errors in calculations of the mean ocean flow when hydrographic sections from many years are combined. It appears possible, however, to correct for these wind-induced effects to allow the lower-frequency signals to be determined. The westernmost edge of our oceanic calculations is at the... Abstract In the central North Atlantic Ocean there are large decadal-scale fluctuations of sea level and of the depth of the thermocline. This variability can be explained by low-frequency Rossby waves forced by wind. The authors have used a simple model of ocean fluctuations driven by the COADS wind-stress curl and find that their model results and hydrographic data, to the limited extent they can be compared meaningfully, agree rather well. The variation in wind curl over the ocean leads to surprisingly large north–south variability in the computed oceanic response over the subtropical gyre. The peak-to-peak sea level differences are as great as 20 cm and persist for many years. These large variations could induce major errors in calculations of the mean ocean flow when hydrographic sections from many years are combined. It appears possible, however, to correct for these wind-induced effects to allow the lower-frequency signals to be determined. The westernmost edge of our oceanic calculations is at the...