The correct representation of the 18.61 year nodal tide is essential for an interpretation of the evolution of mean sea level, as errors cause misleading bias. The nodal tide is currently estimated by applying correction factors in harmonic analysis, which are derived from the equilibrium tide. From the equilibrium tide, correction values f for amplitude and u for phase are determined, which alter lunar tidal constituents, depending on the nodal cycle. This approach has proven to be valid for many tide gauges, even though the impact of the nodal tide in shelf seas has been shown to differ from their theoretical correction value. Hence, tidal constituents from tide records in the North Atlantic shelf region were analyzed for their nodal amplitude and phase lag with a new multiple, non-linear regression approach, which is able to approximate the nodal modulation quantitatively and its agreement to the theoretical equilibrium tide. Results show an overestimation of the lunar M and N constituents by the equilibrium of more than 2.7% in the Wadden Sea, while O and K are underestimated by 1 to 4.6%, which would produce an error of 2 to 5 cm e.g. in the German Wadden Sea. Additionally, a process-based model of the North Sea was applied at the diurnal minimum and maximum of the nodal cycle to calculate the spatial distribution of f and u. Model results reproduce a regionally varying pattern of f and u, indicating how the amplitude modulation of nodal constituents in shallow areas is distributed.