One of the puzzling features of sea level on the east coast of the United States is the decedal-scale variability;the fluctuations are 10–15 cm, peak to peak, at periods longer than a few years. The authors find that this variability, in the frequency band treated with the model, is largely caused by a deep-sea signal generated by the wind stress curl over the North Atlantic. A simple forced long Rossby wave model of the response of the thermocline to wind forcing is used, computing long-wave speeds from observed hydrographic data. The authors model the response of the ocean at periods longer than 3 years for the full width of the Atlantic and for the north–south extent of the main anticyclonic gyre, 18°–38°N. The model output in deep water shows remarkably good agreement with tide gauges, both at Bermuda (32°N) and Puerto Rico (18°N), as well as with dynamic height fluctuations of ∼20 cm peak to peak. Once these fluctuations reach the western side of the ocean, the authors estimate coastal sea l... Abstract One of the puzzling features of sea level on the east coast of the United States is the decedal-scale variability;the fluctuations are 10–15 cm, peak to peak, at periods longer than a few years. The authors find that this variability, in the frequency band treated with the model, is largely caused by a deep-sea signal generated by the wind stress curl over the North Atlantic. A simple forced long Rossby wave model of the response of the thermocline to wind forcing is used, computing long-wave speeds from observed hydrographic data. The authors model the response of the ocean at periods longer than 3 years for the full width of the Atlantic and for the north–south extent of the main anticyclonic gyre, 18°–38°N. The model output in deep water shows remarkably good agreement with tide gauges, both at Bermuda (32°N) and Puerto Rico (18°N), as well as with dynamic height fluctuations of ∼20 cm peak to peak. Once these fluctuations reach the western side of the ocean, the authors estimate coastal sea l...